The ultimate guide to A/B testing | Ronny Kohavi (Airbnb, Microsoft, Amazon)

Ronny Kohavi, PhD, is a consultant, teacher, and leading expert on the art and science of A/B testing. Previously, Ronny was Vice President and Technical Fellow at Airbnb, Technical Fellow and corporate VP at Microsoft (where he led the Experimentation Platform team), and Director of Data Mining and Personalization at Amazon. He was also honored with a lifetime achievement award by the Experimentation Culture Awards in September 2020 and teaches a popular course on experimentation on Maven. In today’s podcast, we discuss: • How to foster a culture of experimentation • How to avoid common pitfalls and misconceptions when running experiments • His most surprising experiment results • The critical role of trust in running successful experiments • When not to A/B test something • Best practices for helping your tests run faster • The future of experimentation Enroll in Ronny’s Maven class, Accelerating Innovation with A/B Testing, at https://bit.ly/ABClassLenny. Promo code “LENNYAB” will give $500 off the class for the first 10 people to use it. — Brought to you by Mixpanel—Event analytics that everyone can trust, use, and afford: https://mixpanel.com/startups | Round—The private network built by tech leaders for tech leaders: https://www.round.tech/apply?utm_campaign=lennys-letter&utm_medium=email-ad&utm_source=email-marketing&utm_content=send-2-2023-07-27 | Eppo—Run reliable, impactful experiments: https://www.geteppo.com/ Find the full transcript at: https://www.lennysnewsletter.com/p/the-ultimate-guide-to-ab-testing Where to find Ronny Kohavi: • Twitter: https://twitter.com/ronnyk • LinkedIn: https://www.linkedin.com/in/ronnyk/ • Website: http://ai.stanford.edu/~ronnyk/ Where to find Lenny: • Newsletter: https://www.lennysnewsletter.com • Twitter: https://twitter.com/lennysan • LinkedIn: https://www.linkedin.com/in/lennyrachitsky/ In this episode, we cover: (00:00) Ronny’s background (04:29) How one A/B test helped Bing increase revenue by 12% (09:00) What data says about opening new tabs (10:34) Small effort, huge gains vs. incremental improvements (13:16) Typical fail rates (15:28) UI resources (16:53) Institutional learning and the importance of documentation and sharing results (20:44) Testing incrementally and acting on high-risk, high-reward ideas (22:38) A failed experiment at Bing on integration with social apps (24:47) When not to A/B test something (27:59) Overall evaluation criterion (OEC) (32:41) Long-term experimentation vs. models (36:29) The problem with redesigns (39:31) How Ronny implemented testing at Microsoft (42:54) The stats on redesigns (45:38) Testing at Airbnb (48:06) Covid’s impact and why testing is more important during times of upheaval (50:06) Ronny’s book, Trustworthy Online Controlled Experiments: A Practical Guide to A/B Testing (51:45) The importance of trust (55:25) Sample ratio mismatch and other signs your experiment is flawed (1:00:44) Twyman’s law (1:02:14) P-value (1:06:27) Getting started running experiments (1:07:43) How to shift the culture in an org to push for more testing (1:10:18) Building platforms (1:12:25) How to improve speed when running experiments (1:14:09) Lightning round Referenced: • Trustworthy Online Controlled Experiments: A Practical Guide to A/B Testing: https://experimentguide.com/ • Seven rules of thumb for website experimenters: https://exp-platform.com/rules-of-thumb/ • GoodUI: https://goodui.org • Defaults for A/B testing: http://bit.ly/CH2022Kohavi • Ronny’s LinkedIn post about A/B testing for startups: https://www.linkedin.com/posts/ronnyk_abtesting-experimentguide-statisticalpower-activity-6982142843297423360-Bc2U • Sanchan Saxena on Lenny’s Podcast: https://www.lennyspodcast.com/sanchan-saxena-vp-of-product-at-coinbase-on-the-inside-story-of-how-airbnb-made-it-through-covid-what-he8217s-learned-from-brian-chesky-brian-armstrong-and-kevin-systrom-much-more/ • Optimizely: https://www.optimizely.com/ • Optimizely was statistically naive: https://analythical.com/blog/optimizely-got-me-fired • SRM: https://www.linkedin.com/posts/ronnyk_seat-belt-wikipedia-activity-6917959519310401536-jV97 • SRM checker: http://bit.ly/srmCheck • Twyman’s law: http://bit.ly/twymanLaw • “What’s a p-value” question: http://bit.ly/ABTestingIntuitionBusters • Fisher’s method: https://en.wikipedia.org/wiki/Fisher%27s_method • Evolving experimentation: https://exp-platform.com/Documents/2017-05%20ICSE2017_EvolutionOfExP.pdf • CUPED for variance reduction/increased sensitivity: http://bit.ly/expCUPED • Ronny’s recommended books: https://bit.ly/BestBooksRonnyk • Chernobyl on HBO: https://www.hbo.com/chernobyl • Blink cameras: https://blinkforhome.com/ • Narrative not PowerPoint: https://exp-platform.com/narrative-not-powerpoint/ Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email [email protected]. Lenny may be an investor in the companies discussed.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  I'm very clear that I'm a big fan of test everything, which is any code change that you make, any feature that you introduce has to be in some experiment because again, I've observed this sort of surprising result that even small bug fixes, even small changes can sometimes have surprising, unexpected impact. And so I don't think it's possible to experiment too much. You have to allocate sometimes to these high-risk, high-reward ideas. We're going to try something that's most likely to fail. But if it does win, it's going to be a home run. And you have to be ready to understand and agree that most will fail.

• 0:43Speaker 00

  And it's amazing how many times I've seen people come up with new designs or a radical new idea and they believe in it and that's okay. I'm just cautioning them all the time to say, if you go for something big, try it out, but be ready to fail 80% of the time.

• 1:05Lenny Rachitsky?

  Welcome to Lenny's Podcast where I interview world-class product leaders and growth experts to learn from their hard-won experiences building and growing today's most successful products. Today, my guest is Ronny Kohavi. Ronny is seen by many as the world expert on A/B testing and experimentation. Most recently, he was VP and Technical Fellow of Relevance at Airbnb, where he led their search experience team. Prior to that, he was Corporate Vice President at Microsoft, where he led Microsoft's experimentation platform team. Before that, he was Director of Data Mining and Personalization at Amazon. He's currently a full-time advisor and instructor. He's also the author of the go-to book on experimentation called Trustworthy Online Controlled Experiments. And in our show notes, you'll find a code to get a discount

• 1:49Lenny Rachitsky?

  on taking his live cohort-based course on Maven. In our conversation, we get super tactical about A/B testing. Ronnie shares his advice for when you should start considering running experiments at your company, how to change your company's culture to be more experiment-driven, what are signs your experiments are potentially invalid, why trust is the most important element. Of a successful experiment, culture and platform. How to get started if you want to start running experiments at your company. He also explains what actually is a P value and something called Twieman's Law, plus some hot takes about Airbnb and experiments in general. This episode is for anyone who's interested in either creating an experiment-driven culture at their company, or just fine-tuning one that already exists.

• 2:33Lenny Rachitsky?

  Enjoy this episode with Ronnie Kohavi after a short word from our sponsors. This episode is brought to you by Mixpanel. Get deep insights into what your users are doing at every stage of the funnel at a fair price that scales as you grow. Mixpanel gives you quick answers about your users from awareness to acquisition through retention. And by capturing website activity, ad data, and multi-touch attribution right in Mixpanel, you can improve every aspect of the full user funnel. Powered by first-party behavioral data instead of third-party cookies, Mixpanel is built to be more powerful and easier to use than Google Analytics. Explore plans for teams of every size and see what Mixpanel can do for you at mixpanel.com/friends/lenny.

• 3:18Lenny Rachitsky?

  And while you're at it, they're also hiring. So check it out at mixpanel.com/friends/lenny.

• 3:27Lenny Rachitsky?

  This episode is brought to you by Round. Round is the private network built by tech leaders for tech leaders. Round combines the best of coaching, learning, and authentic relationships to help you identify where you want to go and accelerate your path to get there, which is why their waitlist tops thousands of tech execs. Round is on a mission to shape the future of technology and its impact on society. Leading in tech is uniquely challenging, and doing it well is easiest when surrounded by leaders who understand your day-to-day experiences. When we're meeting and building relationships with the right people, we're more likely to learn, find new opportunities, be dynamic in our thinking, and achieve our goals. Building and managing your network doesn't have to feel like networking.

• 4:11Lenny Rachitsky?

  Join Round to surround yourself with leaders from tech's most innovative companies. Build relationships, be inspired, take action. Visit round.tech/apply and use promo code Lenny to skip the waitlist. That's round.tech/apply.

• 4:30Lenny Rachitsky?

  Ronnie, welcome to the podcast.

• 4:33Speaker 00

  Thank you for having me.

• 4:34Lenny Rachitsky?

  So you're known by many as maybe the leading expert on A/B testing and experimentation, which I think is something every product company eventually ends up trying to do, often badly. And so I'm excited to dig quite deep into the world of experimentation and A/B testing to help people run better experiments. So thank you again for being here.

• 4:54Speaker 00

  Great goal. Thank you.

• 4:56Lenny Rachitsky?

  Let me start with kind of a fun question. What is maybe the most unexpected A/B test you've run? Or maybe the most surprising result from an A/B test that you've run?

• 5:05Speaker 00

  Yeah, so I think the opening example that I use in my book and in my class is the most surprising public example we can talk about. And this is, this was kind of an interesting experiment. Somebody proposed to change the way that ads were displayed on Bing, the search engine. And he basically said, let's take the second line and move it, promote it to the first line so that the title line becomes larger.

• 5:37Speaker 00

  And when you think about that, and there's, you know, if you're gonna look in my book or in the class, there's an actual diagram of what happened, the screenshots. But if you think about it, just realistically, it looks like a meh idea. Like, why would this be such a reasonable, interesting thing to do? And indeed, when we went back to the backlog, it was on the backlog for months and languished there, and many things were rated higher. But the point about this is it's trivial to implement. So if you think about return on investment, we could get the data by having some engineer spend a couple of hours implementing it. And that's exactly what happened.

• 6:20Speaker 00

  Somebody, Ed Bing, who kept seeing this in the backlog and said, my God, we're spending too much time discussing it, I could just implement it. And he did. He spent a couple of days implementing it. And as is, you know, the common thing at Bing, he launched the experiment. And a funny thing happened. We had an alarm, big escalation, something is wrong with the revenue metric. Now this alarm fired several times in the past when there were real mistakes where somebody would log revenue twice, or you know, there's some data problem. But in this case, there was no bug. We, that simple idea increased revenue by about 12%. And this is something that just doesn't happen.

• 7:04Speaker 00

  We can talk later about Twyman's Law, but that was the first reaction, which is this is too good to be true. Let's find the bug. And we did, and we looked for several times and we replicated the experiment several times and there was nothing wrong with it. This thing was worth a hundred million dollars at the time when Bing was a lot smaller. And the key thing is it didn't hurt the user metrics. So it's very easy to increase revenue by doing theatrics that, you know, displaying more ads is a trivial way to raise revenue. But it hurts the user experience and we've done the experiments to show that. In this case, this was just a home run that improved revenue didn't significantly hurt the, the guardrail metrics.

• 7:45Speaker 00

  And so I was, we were like in awe of, you know, what a trivial change that was the biggest revenue impact of Bing in all its history.

• 7:57Lenny Rachitsky?

  And that was basically shifting in two lines, right? Switching two lines in the search results, right?

• 8:02Speaker 00

  And this was just moving the second line to the first line. Now you then go and run a lot of experiments to understand what happened here. Is it the fact that the title line has a bigger font, sometimes different color? So we ran a whole bunch of experiments and this is what usually happens. We have a breakthrough. You start to understand more about what can we do? And there was suddenly a shift towards, okay, what are other things we could do that would allow us to improve revenue? We came up with a lot of follow on ideas that helped a lot. But to me, this was an example of a tiny change. That was the best revenue generating idea in Bing's history. And we didn't rate it properly, right?

• 8:46Speaker 00

  Nobody gave this the priority that in hindsight it deserves. And that's something that happens often. I mean, we are often humbled by how bad we are at predicting the outcome of experiments.

• 9:01Lenny Rachitsky?

  This reminds me of a classic experiment at Airbnb while I was there. And we'll talk about Airbnb in a bit. The search team just ran a small experiment of what if we were to open a new tab? Every time someone clicked on a search result instead of just going straight to that listing? And that was one of the biggest wins in search.

• 9:18Speaker 00

  Yeah, and by the way, I don't know if you know the history of this, but I tell about this in class. We did this experiment way back around 2008, I think. And so this predates Airbnb. And I remember it was heavily debated, like, why would you open something in a new tab? The users didn't ask for it. Uh, there's a lot of pushback from the designers and we ran that experiment. And again, it was one of these highly surprising results that made it. That we learned so much from it. So we first did this, it was done in the UK for opening Hotmail. And then we moved it to MSN so it would open search in new tab. And all the set of experiments were highly, highly beneficial.

• 10:01Speaker 00

  We published this and I have to tell you, when I came to Airbnb, I talked to our joint friend Ricardo about this. And it was sort of done, it was very beneficial and then it was semi-forgotten. Which is one of the things you learn about institutional memories. When you have winners, make sure to address them and remember them. So it was at Airbnb done for a long time before I joined that listings open in a new tab. But other things that were designed in the future were not done. And I reintroduced this to the team and we saw big improvements.

• 10:35Lenny Rachitsky?

  Shout out to Ricardo, our mutual friend who helped make this conversation happen. There's this like holy grail of experiments that I think people are always looking for of like one, you know, hour of work and it creates this massive result. I imagine this is very rare and don't expect this to happen. I guess in your experience, how often do you find kind of one of these gold nuggets just lying around?

• 10:56Speaker 00

  Yeah. So again, this is a topic that's uh near and dear to my heart. Everybody wants these amazing results and you know, I show them in chapter one in my book, multiple of these, you know, small efforts, huge gain. But as you said, they're very rare. I think most of the time the winnings are made sort of this inch by inch. And there's a graph that I show in my book, a real graph of how Bing Ads has managed to improve their revenue per thousand searches over time. And every month you can see a small improvement and a small improvement, sometimes a degradation because of

• 11:37Speaker 00

  legal reasons or other things. We, you know, there was some concern that we were not marking the ads properly. So you have to suddenly do something that you know is going to hurt revenue. But yes, I think most results are inch by inch. You improve small amounts, lots of them. I think that the best example that I can say is a couple of them that I can speak about. One is at Bing, the relevance team. Hundreds of people all working to improve Bing relevance. They have a metric, we'll talk about, oh, we see the overall evaluation criterion, but they have a metric that their goal is to improve it by 2% every year.

• 12:16Speaker 00

  It's a small amount and that 2%, you can see here's a point one, here's a point one five, here's a point two. And then they add up to around 2% every year, which is amazing. Another example that I am allowed to speak about uh from Airbnb is the fact that we ran some 250 experiments.

• 12:38Speaker 00

  In my tenure there in search relevance. And again, small improvements added up. So this became overall a 6% improvement to revenue. You know, so when you think about 6%, it's a big number, but it became, it came out not of one idea, but many, many smaller ideas that each gave you a small gain. And in fact, I, again, there's another number I'm allowed to say of these experiments, 92% failed to improve the metric that we were trying to move. So only 8% of our ideas actually were successful at moving the key metrics.

• 13:17Lenny Rachitsky?

  There's so many threads I want to follow here, but let me follow this one right here. You just mentioned of 92% of experiments failed. Is that typical in your experience running, seeing experiments run a lot of companies? Like, what should people expect when they're running experiments? What percentage should they expect to fail?

• 13:31Speaker 00

  Well, first of all, I published three different numbers for my career. So overall at Microsoft, about 66%, two thirds of ideas fail, right? And don't think the 66 is accurate, like, you know, it's about two thirds. At Bing, which is a much more optimized domain, after we've been optimizing it for a while, the failure rate was around 85%. So it's harder to improve something that you've been optimizing for a while. And then at Airbnb, this 92% number is, you know, the highest failure rate that I've observed. Now, I've quoted other sources that, you know, it's not that I worked at groups that were particularly bad.

• 14:15Speaker 00

  Booking, Google Ads, other companies published numbers. There are around 80 to 90% failure rate of ideas. This is where it's important of experiments. It is important to realize that when you have a platform, it's easy to get this number. You look at how many experiments were run and how many of them launched. Not every experiment maps to an idea. So it's possible that when you have an idea, your first implementation, you start an experiment, boom, it's egregiously bad because you have a bug. In fact, 10% of experiments tend to be aborted on the first day. Those are usually not that the idea is bad, but that there is an implementation issue

• 14:57Speaker 00

  or something we haven't thought about that forces an abort. You may iterate and pivot again, and ultimately, if you do two or three or four pivots or bug fixes, you may get to a successful launch. But those numbers of 80 to 92% failure rate are of experiments. Very humbling. I know that every group that starts to run experiments, they always start off by thinking that somehow they're different and their success rate is going to be much, much higher. And they're all humbled.

• 15:29Lenny Rachitsky?

  You mentioned that you have this pattern of clicking a link and opening a new tab is a thing that has worked at a lot of different places.

• 15:36Speaker 00

  Yeah.

• 15:37Lenny Rachitsky?

  Are there other versions of this? Do you collect kind of a list of like, here's things that often work when we want to move? Yeah. Are there some you, some you could share? I don't know if you have a list in your head.

• 15:47Speaker 00

  I can give you two resources. One of them is a paper that we wrote called Rules of Thumb. And what we tried to do at that time at Microsoft was to just look at thousands of experiments that run and extract some patterns. And so that's, that's one paper that we can then put in the notes.

• 16:07Lenny Rachitsky?

  Perfect.

• 16:08Speaker 00

  But there's another more, more accurate, I would say, resource that's useful that I recommend to people, and it's a site called goodui.org. And goodui.org is exactly the site that tries to do what you're saying at scale. So guy's name is Jacob Blinovsky. He asks people to send them results of experiments and he derives, he puts them into patterns. There's probably like 140 patterns, I think, at this point. And then for each pattern, he says, well, who hasn't helped? How many times and by how much? So you have an idea of, you know, this worked three out of five times and it was a huge win.

• 16:51Speaker 00

  In fact, you can find that open in your window in there.

• 16:54Lenny Rachitsky?

  I feel like you feed that into chat GPT and you have basically a product manager creating a roadmap. Tool.

• 17:01Speaker 00

  In general, by the way, this is all about, a lot of that is institutional memory. Right. Which is can you document things well enough? So that the organization remembers the successes and failures and learns from them. I think one of the mistakes that some company makes is they launch a lot of experiments and never go back and summarize the learnings. So I've actually put a lot of effort in this idea of institutional learning, of doing the quarterly meeting of the most surprising experiments. By the way, surprising is another question that people often are not clear about. What is a surprising experiment? To me, a surprising experiment is one where the estimated result beforehand

• 17:47Speaker 00

  and the actual result differ by a lot. So that absolute value of the difference is large. Now, you can expect something to be great and it's flat. Well, you learn something, but if you expect something to be small and it turns out to be great, like that ad title promotion, then you've learned a lot. Or conversely, if you expect that something will be small and it's very negative, you can learn a lot by understanding why this was so negative. And that's interesting. So we focus not just on the winners, but also surprising losers. Things that people thought would be a no brainer to run. And then for some reason it was very negative. And sometimes it's that negative that gives you insight.

• 18:31Speaker 00

  I'll actually, you know, I'm just coming up with one example of that that I should mention. We were running this experiment at Microsoft to improve the Windows indexer. And the team was able to show on offline test that it does much better at indexing. And, you know, they showed some relevance is higher and all these good things. And then they ran it as an experiment and you know what happened? Surprising result. Indexing, the relevance was actually high, but it killed the battery life.

• 19:02Lenny Rachitsky?

  Hmm.

• 19:03Speaker 00

  So here's something that comes from left field that you didn't expect. It was consuming a lot more CPU on laptops. It was killing the laptops. And therefore, okay, we learned something. Let's document it. Let's remember this so that, you know, we now take this other factor into account as we design the next iteration.

• 19:23Lenny Rachitsky?

  What advice do you have for people to actually remember these surprises? You said that a lot of it is institutional. What do you recommend people do so that they can actually remember this when people leave, say, three years later?

• 19:34Speaker 00

  Document it, you know, right? You know, we had a large deck internally of these successes and failures. And we encourage people to look at. The other thing that's very beneficial is just to have your whole history of experiments and do some ability to search by keywords, right? So I'm, I have an idea, type a few keywords and see if, from the thousands of experiments that ran, and by the way, these are very reasonable numbers. At Microsoft, just to let you know, when I left in 2019, we were on a rate of about 20 to 25,000 experiments every year. So every working day, we were starting something like 100 new treatments. Big numbers.

• 20:18Speaker 00

  So when you're running in a group like Bing, which is running thousands and thousands of experiments, you want to be able to ask, has anybody did an experiment on this or this or this? And so that searching capability is in the platform. But more than that, I think just doing the quarterly meeting of the most successful, most interesting, sorry, not just successful, most interesting experiments is very key. And that also helps the flywheel of experimentation.

• 20:45Lenny Rachitsky?

  This is a good segue to something I wanted to touch on, which is there's often a, I guess, a weariness of running too many experiments and being too data-driven and the sense that experimentation just leads you to these micro optimizations and you don't really innovate and do big, do big things. What's your perspective on that? And then just, can you be too experiment-driven in your experience?

• 21:07Speaker 00

  I'm very clear that I'm a big fan of test everything, which is any code change that you make. Any feature that you introduce has to be in some experiment because again I've observed this sort of surprising result that even small bug fixes, even small changes can sometimes have surprising unexpected impact. And so I don't think it's possible to experiment too much. I think it is possible to focus on incremental changes. Because some people say, well, you know, if we only tested 17 things around this, and you have to think

• 21:48Speaker 00

  about it's not just, it's like in stock, you need a portfolio, you need some experiments that are incremental that move you in the direction that you know you're going to be successful over time if you just try enough. But some experiments have, you have to allocate sometimes to these high risk, high reward ideas. We're going to try something that's most likely to fail, but if it does win, it's going to be a home run. And so you have to allocate some efforts to that. And you have to be ready to understand and agree that most will fail. Most of these, and it's amazing how many times I've seen people come up with new designs, or a radical new idea, and they believe in it, and that's okay. I'm

• 22:33Speaker 00

  just cautioning them all the time to say, if you go for something big, try it out, but be ready to fail 80% of the time. Right? And one true example that again, I'm able to talk about, because we put it in my book, is we were at Bing, trying to change the landscape of search. And one of the ideas, the big ideas was, we're going to integrate with social. So we hooked into the Twitter Firehose feed, and we hooked into Facebook, and we spent 100 person years on this idea.

• 23:14Speaker 00

  And it failed. You don't see it anymore. It existed for about a year and a half and all the experiments were just negative to flat. And, you know, it was an attempt. It was fair to try it. I think it took us a little long to fail to decide that this is a failure. But at least we had the data. We had hundreds of experiments that we tried. None of them were a breakthrough. And I remember sort of mailing Chi Lu with some statistics showing that, you know, it's time to abort. It's time to fail on this. And, you know, he decided to continue more. It's a million dollar question, you know, do you continue?

• 23:58Speaker 00

  And then maybe the breakthrough will come next month. Or do you abort? And a few, a few months later, we, we aborted.

• 24:07Lenny Rachitsky?

  That reminds me of at Netflix, they tried a social component that also failed. At Airbnb, early on, there was a big social attempt to make like, here's your friends have stayed at these Airbnbs. Completely not had no impact. So maybe that's one of these learnings that we should document.

• 24:21Speaker 00

  Yeah, this is hard. This is hard. And, uh, but that's again, that's the value of experiments, which are this oracle that gives you the data. You may be excited about things. You may believe it's a good idea. But ultimately the arbiter, the oracle is the controlled experiment. It tells you whether users are actually benefiting from it. Whether you and the users, the company and the users.

• 24:48Lenny Rachitsky?

  There's obviously a bit of overhead and downsides running an experiment, setting it all up and making sure, you know, analyzing the results. Is there anything that you ever don't think is worth A B testing?

• 24:59Speaker 00

  First of all, there are some necessary ingredients to A B testing. And I'll just say outright, not every domain is amenable to A B testing, right? You can't A B test mergers and acquisitions, right? It's something that happens once. You either acquire, you don't acquire. So you do have to have some necessary ingredient. You need to have enough units, mostly users in order for the statistics to work out. So yeah, if you're too small. It may be too early to A B test. But what I find is that in software, it is so easy to run A B testing and it is so easy to build a platform. I don't say it's easy to build a platform, but once you've built a platform,

• 25:43Speaker 00

  the incremental cost of running an experiment should approach zero. And we got to that at Microsoft where after a while, the cost of running experiments was so low that nobody was questioning the idea that everything should be experimented with. Now, I don't think we were there at Airbnb, for example. The platform at Airbnb was much less mature and required a lot more analysts in order to interpret the results and to find issues with it. So I do think there's this trade off. You're willing to invest in the platform. It is possible to get the marginal cost to be close to zero. But when you're not there, it's still expensive and there are, there may be reasons why not to run A B tests.

• 26:28Lenny Rachitsky?

  You talked about how you may be too small to run A B tests. And this is a constant question for startups is when should we start running A B tests? Do you have kind of a heuristic or a rule of thumb of just like, here's a time you should really start thinking about running an experiment?

• 26:42Speaker 00

  Yeah, the million dollar question that everybody asks. So I actually, we'll put this in the notes, but I gave a talk last year. What I called it is practical defaults. And one of the things I show there is that unless you have at least tens of thousands of users, the math, the statistics just don't work out for most of the metrics that you're interested in. In fact, you know, I gave an actual practical number of a retail site with some conversion rate, trying to detect changes that are at least, you know, 5% beneficial, which is something that startups should focus on. They shouldn't focus on the 1%, they should focus on the 5 and 10%. Then you need something like 200,000 users.

• 27:24Speaker 00

  Right? So start experimenting when you're in the tens of thousands of users. You'll be only be able to detect large effects. And then, once you get to 200,000 users, then the magic starts happening. Then you can start testing a lot more. Then you have the ability to test everything and make sure that you're not degrading and getting value out of its competition. So you asked for rule of thumb, 200,000 users, you're magical. Below that, start building the culture, start building the platform, start integrating. So that as you scale, you you start to see the value.

• 28:00Lenny Rachitsky?

  Love it. Coming back to this kind of concern people have of experimentation keeps you from innovating and taking big bets. I know you have this framework, overall evaluation criterion. And I think that helps with this. Can you talk a bit about that?

• 28:14Speaker 00

  The OEC or the overall evaluation criterion is something that I think many people that start to dabble in A B testing miss. And the question is, what are you optimizing for? And it's a much harder question that people think because it's very easy to say, we're going to optimize for money, revenue. But that's the wrong question because you can do a lot of bad things that will improve revenue. So there has to be some countervailing metric that tells you, how do I improve revenue without hurting the user experience? Okay, so let's take a good example with search.

• 28:58Speaker 00

  You can put more ads on a page and you will make more money. There is no doubt about it. You will make more money in the short term. The question is, what happens to the user experience? And how is that going to impact you in the long term? So we've run those experiments and we were able to map out, you know, this number of ads causes this much increase to churn. This number of ads causes this much increase to the time that users take to find a successful result. And we came up with an OEC that is based on these metrics that allows you to say, okay, I'm willing to take this additional money if I'm not hurting the user experience by more than this much, right? So there's a trade off there.

• 29:42Speaker 00

  One of the nice ways to phrase this as a constraint optimization problem. I want you to increase revenue, but I'm going to give you a fixed amount of average real estate that you can use. Right, so you can, for one query, you can have zero ads. For another query, you can have three ads. For a third query, you can have wider, bigger ads. I'm just gonna count the pixels that you take, the vertical pixels. And I will give you some budget. And if you can, under the same budget, make more money, you're good to go. Right, so that, to me, turns the problem from a badly defined, let's just make more money.

• 30:23Speaker 00

  Right, any page can start plastering more ads and make more money short term, but that's not the goal. The goal is long-term growth and revenue. Then you need to insert these other criteria. And what am I doing to the user experience? One way around it is to put this constraint. Another one is just to have these other metrics. Again, something that we did to look at the user experience. How long does it take the user to reach a successful click? What percentage of sessions are successful? These are key metrics that were part of the overall evaluation criteria that we've used. I can give you another example, by the way, from, you know, the hotel industry or Airbnb that we both worked at. You can say, I want to improve conversion rate.

• 31:05Speaker 00

  But you can be smarter about it and say, it's not just enough to convert a user to buy or to pay for a listing. I want them to be happy with it several months down the road when they actually stay there. Right, so that could be part of your OEC to say, what is the rating that they will give to that listing when they actually stay there? And that's a, that causes an interesting problem because you don't have this data now. You're gonna have it three months from now when they actually stay. So you have to build the training set that allows you to make a prediction about whether this user, whether Lenny is gonna be happy at this cheap place or whether, no, I should offer him something more expensive

• 31:46Speaker 00

  because Lenny likes to stay at nicer places where the water actually is hot and comes out of the faucet.

• 31:52Lenny Rachitsky?

  That is true. Okay, so it sounds like the core to this approach is basically have a kind of a drag metric that makes sure you're not hurting something that's really important to the business. And then being very clear on what's the long-term metric we care most about.

• 32:05Speaker 00

  To me, the key here, the key word is lifetime value. Which is, you have to define the OEC such that it is causally predictive of the lifetime value of the user.

• 32:17Speaker 00

  Right? And that that's what causes you to think about things properly, which is, am I doing something that just helps me short term? Or am I doing something that will help me in the long term? Once you put that model of lifetime value? People say, okay, what about retention rates? We can measure that. What about the time to achieve a task? We can measure that. And those are these countervailing metrics that make it make the OEC useful.

• 32:42Lenny Rachitsky?

  And to understand these longer term metrics, what I'm hearing is use kind of models and forecast and predictions? Or would you suggest sometimes use like a long term holdout or some other approach? Like, what do you find is the best way to think, see these long?

• 32:57Speaker 00

  So there's two ways that I like to think about it. One is you can run long term experiments for the goal of learning something. So I mentioned that at Bing, we did run these experiments where we increased the ads and decreased the ads. So that we will understand what happens to key metrics. The other thing is you can just build models that use some of our background knowledge or use some, you know, data science to look at historical. I'll give you another good example of this. When I came to Amazon, one of the teams that reported to me was the email team. That it was not the transactional emails when you buy something, get an email, but was the team that sent these recommendations.

• 33:41Speaker 00

  You know, here's a book by an author that you bought. Here's a product that we recommend. And the question is, how do we give credit to that team? And the initial version was, well, whenever a user comes from the email and purchases something on Amazon, we're going to give that email credit. Well, it turned out this had no countervailing metric. The more emails you send, the more money you're going to credit the team. And so that led to spam. Literally a really interesting problem. The team just ramped up the number of emails that they were sending out and claimed to make more money and their fitness function improved. And and then so then we backed up and then we said, okay, there we can either phrase this as a constraint satisfaction problem.

• 34:26Speaker 00

  You're allowed to send user an email every X days.

• 34:30Speaker 00

  Or, which is what we ended up doing is let's model the cost of spamming the users. Okay, what's that cost? Well, when they unsubscribe, we can't mail them.

• 34:43Speaker 00

  Okay, so we did some data science study on the side and we said, what is the value that we're losing from an unsubscribe, right? And we came up with a number of a few dollars. But the point was now we have this counter veiling metric. We say, here's the money that we generate from the emails. Here's the money that we're losing a long-term value. What's the trade-off? And then when we started to incorporate this formula, more than half the campaigns that were being sent were negative.

• 35:13Speaker 00

  Okay, so it was a huge insight at Amazon about how to send the right campaigns. And this led, and this is what I like about these discoveries. This fact that we integrated the unsubscribe led us to a new feature to say, well, let's not lose their future lifetime value through email. When they unsubscribe, let's offer them by default to unsubscribe from this campaign. So when you get an email, you know, there's a new book by the author, the default to unsubscribe would be unsubscribe me from author emails. And so now the the negative, the counter veiling metric is much smaller.

• 35:57Speaker 00

  And so again, this was a breakthrough in our ability to send more emails and understand based on what users were unsubscribing from, which ones are really beneficial.

• 36:06Lenny Rachitsky?

  I love the surprising results.

• 36:08Speaker 00

  We all love them. Lots of, I mean, this is, this is the humbling reality. And, you know, people talk about the fact that A B testing sometimes leads you to incremental. I actually think that many of these small insights lead to fundamental insights about, you know, which areas to go, some strategies we should take, some things we should develop. Helps a lot.

• 36:31Lenny Rachitsky?

  This makes me think about how every time I've done a full redesign of a product, I don't think ever has it ever been a positive result. And then the team always ends up having to claw back what they just hurt and try to figure out what they messed up. Is that your experience too?

• 36:47Speaker 00

  Absolutely. Yeah. In fact, I I've published. Uh, some of these in, in LinkedIn posts showing a large set of, you know, big launches that redesigns that dramatically failed. Uh, and it happens very often. So the right way to do this is to say, yes, we want to do a redesign, but let's do it in steps and test on the way and adjust. So you don't need to take 17 new changes that many of them are going to fail. Start to move incrementally in a direction that you believe has been official. Adjust on the way.

• 37:24Lenny Rachitsky?

  The worst part of those experiences I find is it took like, I don't know, three, six months, three or six months to build it. And by the time it's launched, it's just like, we're not going to unlaunch this. Everyone's been working in this direction. All the new features are assuming this is going to work and you're basically stuck, right?

• 37:40Speaker 00

  I mean, this is a sunk cost fallacy, right? We invested so many years in it. Let's launch this even though it's bad for the user. No, that's terrible. Yeah. Yeah. So, uh, this is, this is the other advantage of recognizing this humble reality that most ideas fail, right? If, if you believe in that statistics that I published, then doing 17 changes together is more likely to be negative. Do them in smaller increments, learn from, it's called OFAT, one factor at a time. Do one factor, learn from it and adjust of the 17, maybe you have four good ideas. Those are the ones that will launch and be positive.

• 38:21Lenny Rachitsky?

  I generally agree with that and always try to avoid a big redesign, but it's hard to avoid them completely. There's often team members that are really passionate and like, we just need to rethink this whole experience. We're not going to incrementally get there. Have you found anything effective in helping people either see this perspective? Or just making a larger bet more successful?

• 38:41Speaker 00

  By the way, I, I'm not opposed to large redesigns. I try to give the team the data to say, look, here are lots of examples where big redesigns fail. Try to decompose your redesign. If you can't decompose it to one factor at a time, to a small set of factors at a time. And learn from these smaller changes, what works and what doesn't. Now, it's also possible to do a complete redesign. I'm just, as you said yourself, they be ready to fail.

• 39:15Speaker 00

  Right? I mean, do you, do you really want to work on something for six months or a year? And then run the A B test and realize that you've hurt revenues or other key metrics by several percentage points and a data-driven organization will not allow you to launch? What are you going to write in your annual review?

• 39:32Lenny Rachitsky?

  Yeah, but nobody ever thinks it's going to fail. They can know we got this or talk to so many people.

• 39:36Speaker 00

  But I think organizations that start to run experiments are humbled early on from the smaller changes.

• 39:43Lenny Rachitsky?

  Yeah.

• 39:43Speaker 00

  Right? You're right. Nobody, I'll tell you a funny story. When I came from Amazon to Microsoft, I joined the group and for one reason or another, that group disbanded a month after I joined. And so people came to me and said, look, you just joined the company. You're a partner level. You figure out well, how you can help Microsoft. And I said, I'm going to build an experimentation platform because nobody at Microsoft is running experiments and 50%, more than 50% of ideas in Amazon that we tried, failed. And the classical response was we have better P M's here.

• 40:26Speaker 00

  Right? There was this complete denial that it's possible that 50% of ideas that Microsoft is implementing in a three-year development cycle. By the way, this is how long it took Office to release. It was a classical, every three years we release. And the, the data came about showing that, uh, yeah, you know, Bing was the first to truly implement experimentation at scale and we shared with the rest of the companies, the surprising results. And so when Office was, and this was, you know, credit to Chi Lu and Sachin Adella. They were ones that says, Ronny, you know, you try to get Office to run experiments.

• 41:09Speaker 00

  We'll give you the air support and it was hard, but we did it, you know, it was, it took a while, but Office started to run experiments and they realized that many of their ideas were failing.

• 41:20Lenny Rachitsky?

  You said that there's a site of failed redesigns. Was that, is that in your book? Or is that a site that you can point people to, to kind of help build this case?

• 41:28Speaker 00

  It's, I teach this in my class, but I, I think I've posted this on LinkedIn and answered some questions. I'm happy to put that in the notes.

• 41:36Lenny Rachitsky?

  Okay, cool. We'll put that in the show notes because I think that's the kind of data that often helps convince a team. Maybe we shouldn't rethink this entire onboarding flow from scratch. Maybe we should kind of iterate towards and learn as we go. This episode is brought to you by EPO. EPO is a next generation A/B testing platform built by Airbnb alums from modern growth teams. Companies like DraftKings, Zapier, ClickUp, Twitch and Cameo rely on EPO to power their experiments. Wherever you work, running experiments is increasingly essential, but there are no commercial tools that integrate with a modern growth team stack. This leads to waste of time building internal tools or trying to run your own experiments through a clunky marketing tool. When I was at Airbnb, one of the things that I loved most about working there was

• 42:18Lenny Rachitsky?

  our experimentation platform, where I was able to slice and dice data by device types, country, user stage. EPO does all that and more, delivering results quickly, avoiding annoying prolonged analytic cycles and helping you easily get to the root cause of any issue you discover. EPO lets you go beyond basic click-through metrics and instead use your North Star metrics like activation, retention, subscription and payments. EPO supports tests on the front end, on the back end, email marketing, even machine learning clients. Check out EPO at geteppo.com. That's geteppo.com and 10x your experiment velocity. Is it ever worth just going, let's just rethink this whole thing and just give it a shot to break out of a local minima or a local maxima essentially?

• 43:03Speaker 00

  So I think what you said is fair. I mean, I do want to allocate some percentage of resources to big bets. As you said, we've been optimizing this thing to hell. Could we completely redesign it? It's a very valid. Idea, you may be able to break out of a local minimum. What I'm telling you is 80% of the time you will fail. So be ready for that, right? What people usually expect is my redesign is going to work. No, you're most likely going to fail. But if you do succeed, it's a breakthrough.

• 43:35Lenny Rachitsky?

  I like this 80% rule of thumb. Is that just like a simple way of thinking about it? 80% you're experimenting with.

• 43:40Speaker 00

  Yeah, rule of thumb. And you know, you had, uh, you know, I've heard people say it's 70% or 80%, but it's in that area where I think, you know, when you talk about how much to invest in the known versus the high risk, high reward, that's usually the right percentage that most organizations end up doing this allocation, right? You interviewed Trey, I think he mentioned that, uh, you know, that at Google is like 70%, you know, the search and ads and it's a 20% for some of the apps and new stuff. And then it's the 10% for infrastructure.

• 44:16Lenny Rachitsky?

  Yeah, I think the, the most important point there is if you're not running an experiment, 70% of stuff you're shipping is hurting your business.

• 44:22Speaker 00

  Well, not hurting, it may, it's flat to negative. Some of them are flat. Uh, and by the way, flat to me, if something is not that sick, that's a no ship because you've just introduced more code. There is a maintenance overhead to shipping your stuff. I've heard people say, look, we already spent all this time. The team will be demotivated if we don't ship it. And I'm no, that's wrong guys, right? You know, let's make sure that we understand that shipping this project has no value is complicating the code base. Maintenance costs will go up. You don't ship on flat unless it's a sort of a legal requirement, right? When legal comes along and says you have

• 45:03Speaker 00

  to do X or Y, you have to ship on flat or even negative. And that's understandable. But again, I think that's something that a lot of people make the mistake of saying legal told us we have to do this. Therefore, we're going to take the hits. No, legal gave you a framework that you have to work under. Try three different things and ship the one that hurts the least.

• 45:25Lenny Rachitsky?

  I love that. It reminds me when Airbnb launched the rebrand, even that they ran as an experiment with the entire homepage redesign, the new logo and all that. And I think there's a long-term holdout even. And I think it was positive in the end, from what I remember.

• 45:41Lenny Rachitsky?

  Speaking of Airbnb, I want to chat about Airbnb briefly. I know there's, I know you're limited in what you can share, but it's interesting that Airbnb seems to be moving in this other direction, where it's becoming a lot more top-down, Brian vision-oriented, and Brian's even talked about how he's less motivated to run experiments. He doesn't want to run as many experiments as they used to. Things are going well, and so, you know, it's hard to argue with the success potentially. You worked there for many years. You ran the search team, essentially. I guess, just what was your experience like there? And then roughly, what's your sense of how things are going, where it's going?

• 46:15Speaker 00

  As you, as you know, and I'm restricted from talking about Airbnb, I will say a few things that I am allowed to say. One is, in my team, in search relevance,

• 46:27Speaker 00

  everything was A/B tested.

• 46:30Speaker 00

  So while Brian can focus on some of the design aspects, the people who are actually doing, you know, the neural networks and the search, everything was A/B tested to hell. So nothing was launching without an A/B test. We had targets around improving certain metrics, and everything was that A/B test. Now, other teams, some did, some did not. I will say that, you know, when you say things are going well, I think we don't know the counterfactual. I believe that had Airbnb kept people like Greg Reilly, which was pushing for a lot more data-driven, and had Airbnb run more experiments, it would have been in a better state than today. But it's the counterfactual we don't know.

• 47:13Lenny Rachitsky?

  That's a really interesting perspective, yeah. Airbnb is such an interesting, natural experiment of a way of doing things differently. There's like de-emphasizing experiments, and also, they turned off paid ads during COVID, and I think, I don't know where it is now, but it feels like it's become a much smaller part of the growth strategy. Who knows if they've ramped it up to back to where it is today, but I think it's going to be a really interesting case study, looking back, I don't know, five, 10 years from now.

• 47:37Speaker 00

  It's a one-off experiment where it's hard to assign value to some of the things that Airbnb is doing. I personally believe it could have been a lot bigger and a lot more successful if it had run more controlled experiments. But I can't speak about some of those that I ran and that showed that some of the things that were initially untested were actually negative and could be better.

• 48:04Lenny Rachitsky?

  Right. Mysterious.

• 48:07Lenny Rachitsky?

  One more question Airbnb, you were there during COVID, which was quite a wild time for Airbnb. We had Sanjit on the podcast talking about all the craziness that went on when travel basically stopped. And there was a sense that Airbnb was done and travel is not going to happen for years and years. What's your take on experimentation in that world where you have to really move fast, make crazy decisions, make big decisions? What did you, what was like during that time?

• 48:34Speaker 00

  So I, I think actually in a state like that, it's even more important to run A B tests. Right? Because what you want to be able to see is if we're making this change, is it actually helping in the current environment? You know, there's this idea of external generalizability. Is it going to work out now during COVID? Is it going to generalize later on? These are things that you can really answer with the controlled experiments. And sometimes it means that you might have to replicate them six months down when COVID, say, is not as impactful as it is. Saying that you have to make decisions quickly. To me, I'll point you to the success rate.

• 49:16Speaker 00

  Like if, if in peacetime, you're wrong two thirds to 80% of the time, why would you be suddenly right in wartime, right? In COVID time? So I, I don't believe in the idea that because

• 49:33Speaker 00

  bookings went down materially, the company should suddenly, you know, not be data driven and do things differently. I think if Airbnb stayed the course, did nothing, the revenue would have gone up in the same way.

• 49:48Lenny Rachitsky?

  Fascinating.

• 49:49Speaker 00

  In fact, if you look at one investment, one big investment that was done at the time was online experiences. And the initial data wasn't very promising. And I think today it's a footnote.

• 50:01Lenny Rachitsky?

  Yeah. What a, another case study for the history books, Airbnb experiences. I want to shift a little bit and talk about your book, which you mentioned a couple of times. It's called Trustworthy Online Controlled Experiments. And I think it's basically the book on A B testing. Let me ask you just, what surprised you most about writing this book and putting it out and the reaction to it?

• 50:24Speaker 00

  I was pleasantly surprised that.

• 50:29Speaker 00

  It sold more than what we thought and more came more than what Cambridge predicted. So when when first we were approached by Cambridge after a tutorial that we did to write a book, I was like, I don't know, this is too small of an inch area. And I, you know, they were saying, so you'll be able to sell a few thousand copies and help the world. And I found, you know, my co-authors, which are great. And, you know, we wrote a book that we thought is not statistically oriented, has fewer formulas than you normally see, and focuses on the practical aspects and on trust, which is the key. The book, as I said, you know, was more successful.

• 51:13Speaker 00

  It sold over 20,000 copies in English. It was translated to Chinese, Korean, Japanese, and Russian. And so it's it's great to see that we helped the world become more data driven with experimentation. And I'm happy because of that. I was pleasantly surprised. By the way, all proceeds from the book are donated to charity. So if I'm pitching the book here, there is no financial gain for me from having more copies sold. I think we made that decision, which was a good decision. All proceeds go to charity.

• 51:46Lenny Rachitsky?

  Amazing. I didn't know that. We'll link to the book in the show notes. You talked about how trust, like it's trust is in the title. You just mentioned how important trust is to experimentation. A lot of people talk about how do I run experiments faster? You focus a lot on trust. Why is trust so important in running experiments?

• 52:03Speaker 00

  So to me, the experimentation platform is the safety net. And it's an oracle. So it serves really two purposes. The safety net means that if you launch something bad, you should be able to abort quickly, right? Safe deployments, safe velocity, there are some names for this. But this is one key value that the platform can give you. The other one, which is the more standard one, is at the end of the two week experiment, we will tell you what happened to your key metric and do many of the others surrogate and debugging and guardrail metrics. Trust builds up, it's easy to lose. And so.

• 52:45Speaker 00

  To me, it is very important that when you present this and say, this is science. This is a controlled experiment. This is the result. You better believe that this is trustworthy. And so I focus on that a lot. I think it allowed us to gain the organizational trust that this is really. And the nice thing is when we we built all this checks to make sure that the experiment is correct. If there were something wrong with it, we would stop and say, hey, something is wrong with the experiment. And I think that's something that some of the early implementations in other places did not do. And it was a big mistake. I've mentioned this in my book, so I can mention this here.

• 53:28Speaker 00

  Optimizely, in its early days, were very statistically naive. They sort of said, hey, we're real time. We can compute your P values in real time. And then you can stop an experiment when the P value is statistically significant. That is a big mistake. That inflates your, what's called type one error or the false positive rate materially. So if you think you've got a 5% type one error or you aim for that P value less than 0.05, using real time sort of P value monitoring to optimize the offered, you would probably have a 30% error rate. So what this led is that people that started using Optimizely

• 54:11Speaker 00

  thought that the platform was telling them they're very successful. But when they actually started to see, well, it told us this is positive revenue, but I don't see this over time, like by now, we should have made double their money. So their questions started to come up around the trust in the platform. There's a very famous post that somebody wrote about how Optimizely almost got me fired by a person who basically said, look, I came to the org and I said, we have all these successes. But then I said, something is wrong. And he tells of how he ran an AA test when there is no difference between the A and the B. And Optimizely told him that it was statistically significant too many times. Optimizely learned.

• 54:53Speaker 00

  Optimizely, you know, several people pointed, I pointed this out in my Amazon review of the book that the Optimizely authors wrote early on. I said, hey, you're not doing the statistics correctly. Other, you know, Ramesh Johari at Stanford pointed this out, became a consultant to the company, and then they fixed it. But to me, that's a very good example of how to lose trust. They lost a lot of trust in the market. They lost all this trust because they built something that had very much inflated error rate.

• 55:26Lenny Rachitsky?

  That is a pretty scary to think about. You've been running all these experiments and they weren't actually telling you accurate results. What are signs that what you're doing may not be valid? If you're starting to run experiments and then just how do you avoid having that situation? What kind of tips can you share for people trying to run experiments?

• 55:47Speaker 00

  You know, there's a whole chapter of that in my book, but I'll say maybe one of the things that is the most common occurrence by far, which is a sample ratio mismatch. Now, what is a sample ratio mismatch? If you design the experiment to send 50% of users to control and 50% of users to treatment, it's supposed to be a random number or, you know, a hash function. If you get something off from 50%, it's a red flag. So let's take a real example. Let's say you're running an experiment and it's large, it's got a million users and you got 50.2. So people say, well, I don't know, it's not going to be exactly the same as 50.2, reasonable or not? Well, there's a formula that you can plug in.

• 56:32Speaker 00

  I have a spreadsheet available for those that are interested. And you can tell, here's how many users are in control. Here's how many users are having treatment. My design was 50/50 and it tells you the probability that this could have happened by chance. Now, in a case like this, you plug in the numbers. It might tell you that this should happen one in half a million experiments. Well, unless you've run half a million experiment, very unlikely that you would get a 50.2 versus 49.8 split. And therefore something is wrong with the experiment, right? Now, people, I remember when we first implemented this check, we were surprised to see how many experiments suffered from this.

• 57:14Speaker 00

  Right? And there's a paper that was published in 2018 and where we share that at Microsoft, even though we'd be running experiments for a while, is around 8% of experiments that suffered from the sample ratio mismatch. And it's a big number. I think about this, you're running 20,000 experiments a year, so many of them, 8% of them are invalid. And we, somebody has to go down and understand what happened here. We know that we can't trust the results, but why? So over time, you begin to understand there's something wrong with the data pipeline, there's something that happens with bots. Bots are a very common factor for causing

• 57:55Speaker 00

  sample ratio mismatch. So there's a whole, that paper that was published by my team, talks about how to diagnose sample ratio mismatches. In the last, about a year and a half, it was amazing to see all these third-party companies implement sample ratio mismatches. And all of them were reporting, oh my God, you know, 6%, 8%, 10%. So yeah, they were, it's sometimes fun to go back and say, how many of your results that were in the past were invalid before you had this sample ratio mismatch test?

• 58:32Lenny Rachitsky?

  That is frightening. Is the most common reason this happens is you're assigning users in kind of the wrong place in your, in your code?

• 58:39Speaker 00

  So when you say most common, I think the most common is bots. Somehow they hit the control or the treatment in different proportions. Because you changed the website, the bot may fail to parse the page and try to hit it more often. That's a classical example. Another one is just the data pipeline. We've had cases where we were trying to remove bad traffic under certain conditions, and it was skewed because of the control and treatment. I've seen people that start an experiment in the middle of the site on some page, but they don't realize that some campaign is pushing people from the side. So there's multiple reasons. It is surprising how often this happens. And I'll tell you a funny story,

• 59:19Speaker 00

  which is when we first added this test to the platform, we just put a banner saying, you have a sample ratio mismatch, do not trust these results.

• 59:30Speaker 00

  And we noticed that people ignored it. They were starting to present results that had this banner. And so we blanked out the scorecard. We put a big, you know, red, can't see this result. You have a sample ratio mismatch. Click OK to expose the results. And why do we need that? Okay, we need that OK button because you want to be able to debug the reasons. And sometimes the metrics help you understand why you have a sample ratio mismatch. So we blanked out the scorecard. We had this button and then we started to see that people pressed the button. This still presented the results of experiments with sample ratio mismatches.

• 1:00:09Speaker 00

  So we ended up with an amazing compromise, which is every number in the scorecard was highlighted with a red line. So that if you took a screenshot, other people could tell you how the sample ratio mismatched.

• 1:00:24Lenny Rachitsky?

  Freaking product managers.

• 1:00:25Speaker 00

  This is, this is intuition. People just say, well, my S R M was small, therefore I can still present the results. People want to see success. I mean, this is a natural bias. And then we have to be very conscientious and fight that bias and say, when something looks too good to be true, investigate.

• 1:00:45Lenny Rachitsky?

  Which is a great segue to something you mentioned briefly, something called Twyman's Law. Can you talk about that?

• 1:00:51Speaker 00

  Yeah. So Twyman's Law, you know, the, the general statement is if any figure that looks interesting or different is usually wrong. Uh, it was first said by this person in the UK who worked in radio media, but I'm a big fan of it. And, uh, you know, my main claim to people is if the result looks too good to be true, if you suddenly moved your, you know, your normal movement of an experiment is under 1% and you suddenly have a 10% movement, hold the celebratory dinner.

• 1:01:24Speaker 00

  Like it was just your first reaction, right? Let's take everybody to a fancy dinner because we just improved revenue by millions of dollars. Hold that dinner, investigate, see, because there's a large probability that something is wrong with the result. And I will say that nine out of 10, when we call out Twyma's law, it is the case that we find some flaw in the experiment. Now, there are obviously outliers, right? That first experiment that I shared where we promoted the made long ad titles, that was successful. But that was replicated multiple times and double and triple checked and everything was good about it. Many other results that were so big turn out to be false.

• 1:02:05Speaker 00

  So I'm a big, I'm a big, big fan of Twyma's law. There's a deck, I can also give this in a note where I shared some real examples of Twyma's law.

• 1:02:13Lenny Rachitsky?

  Amazing. I want to talk about rolling this out of companies and things that you run into that fail. But before I get to that, I'd love for you to explain just P value. I know that people kind of misunderstand it and this may be a good time just to help people understand. What is it actually telling you at P value, say 0.05?

• 1:02:29Speaker 00

  I don't know if this is the right forum for explaining P values because the definition of a P value is simple. What it hides is very complicated. So I'll say one thing, which is many people assign one minus P value as the probability that your treatment is better than control. So you ran an experiment, you got a P value of 0.02. They think there's a 98% probability that the treatment is better than the control. That is wrong. So rather than defining P values, I want to caution everybody that the most common interpretation is incorrect.

• 1:03:08Speaker 00

  P value assumes it's a conditional probability or an assumed probability. It assumes that the null hypothesis is true and we're computing the probability that the data we're seeing.

• 1:03:23Speaker 00

  It matches the hypothesis. There's no hypothesis. In order to get the probability that most people want, we need to apply Bayes' rule and invert the probability from the probability of the data given the hypothesis to the probability of hypothesis given the data. For that, we need an additional number, which is the probability, the prior probability that the hypothesis that you're testing is successful or not. That's an unknown. What we do is we can take historical data and say, "Look, people fail 2/3 of the time or 80% of the time, and we can apply that number and compute that." We've done that in a paper that I will give in the notes so that you can assess the number that you really want,

• 1:04:07Speaker 00

  that what's called the false positive risk. So I think that's something for people to internalize, that what you really want to look at is this false positive risk, which tends to be much, much higher than the 5% that people think. Right? So if you're, I think the classical example in the Airbnb where the failure rate was very, very high is that when you get a statistically significant result, let me actually pull the note so that I know the actual number. If you're at Airbnb where the success rate of, or Airbnb search where the success rate is only 8%, if you get a statistically significant result with a p-value less than 0.05, there is a 26% chance that this is a false positive result.

• 1:04:50Speaker 00

  Right? It's not 5%, it's 26%. So that's the number that you should have in your mind. And that's why when I worked at Airbnb, one of the things we did is we said, "Okay, if you're less than 0.05 but above 0.01, rerun, replicate." When you replicate, you can combine the two experiments and get a combined p-value using something called Fisher's method or Stouffer's method, and that gives you the joint probability, and that's usually much, much lower. So if you get 2.05s or something like that, then the probability that you've got them is much, much lower.

• 1:05:26Lenny Rachitsky?

  Wow, I have never heard it described that way. Makes me think about how like even data scientists in our teams are always just like, this isn't perfect, like, we're not 100% sure this experiment is positive, but on balance, if we're launching positive experiments, we're probably doing good things. It's okay if sometimes we're wrong.

• 1:05:42Speaker 00

  By the way, it's true, on balance, you're probably better than 50/50, but people don't appreciate how much that 26% that I mentioned is high. And the reason that I want to be sure is that I think it leads to this idea of the learning, the institutional knowledge, what you want to be able to say is share with the org is success. And so you want to be really sure that you're successful. So by lowering the P value, by forcing teams to work with a P value maybe below 0.01 and do replication on hires, then you can be much more successful. And the false positive rate will be much, much lower.

• 1:06:20Lenny Rachitsky?

  Fascinating. And also shows the value of keeping track of just what percentage of experiments are failing historically at the company or within that specific product. Say someone listening wants to start running experiments, say they have 10s of thousands of users at this point. What would be the first couple of steps you'd recommend?

• 1:06:37Speaker 00

  Well, so if they have somebody in the org that has previously been involved in experiment, that's a good way to consult internally. I think that the key decision is whether you want to build or buy. There's a whole series of eight sessions that I posted on LinkedIn where I invited guest speakers to talk about this problem. So if people are interested, they can look at how, what the vendor say and what agency said about build versus buy question. And it's usually not a 0, 1, it's usually a both. You build some and you buy some. And it's a question of do you build 10% or do you build a 90%? I think for people starting, the third-party products that are available today are pretty good.

• 1:07:21Speaker 00

  This wasn't the case when I started working. So when I started building, running experiments at Amazon, we were building the platform because nothing existed. Same at Microsoft. I think today, there's enough. Vendors that provide good experimentation platforms that are trustworthy, that I would say not a good way to consider using one of those.

• 1:07:44Lenny Rachitsky?

  So you're at a company where there's resistance to experimentation and A B testing, whether it's a startup or a bigger company. What have you found works in helping shift that culture? And how long does that usually take, especially at a larger company?

• 1:07:57Speaker 00

  My general experiences with Microsoft where, you know, we, we went with this beachhead of Bing. We were running a few experiments and then we were asked to focus on Bing. And we scaled experimentation and built a platform at scale at Bing. Once Bing was successful and we were able to share all these surprising results. I think that many, many more people in the company were amenable. And it's all, it was also the case that helped a lot that, you know, there's the usual cross pollination, people from Bing move out to other groups. And that helped these other groups say, hey, there's a better way to build software. So I think if you're starting out, find a place, find a team where experimentation is easy to run.

• 1:08:41Speaker 00

  And by that, I mean they're launching often, right? Don't go with the team that launches every six months or, you know, Office used to launch every three years. Go with the team that launches frequently. You know, they're running on sprints. They launch every week or two, sometimes they launch daily. I mean, Bing used to launch multiple times a day. And then make sure that you understand the question of the O E C. Is it clear what they're optimizing for? Right? There are some groups where you can come up with a good O E C. Some groups are harder. You know, I remember one funny example was the Microsoft.com website, which this is not M S N, this is Microsoft.com, has like multiple different constituencies that are trying to determine this is a support site.

• 1:09:26Speaker 00

  And this is some, uh, the ability to sell software through this site. And it's, and, and warn you about, you know, safety and updates. It has so many goals. I remember when the team said, we want to run experiments. And I, and I brought the group in and some of the managers. And I said, do you know what you're optimizing for? It was very funny because they, they surprised me. They said, hey, Ronny, we read some of your papers. We know there's this term called O E C. We decided the time on site is our OEC. And I said, wait a minute. Some of your main goals is support site. Is people more spending more time on a support site a good thing or a bad thing?

• 1:10:08Speaker 00

  And then half the room thought that more time is better and half the room thought that more time is worse. So an OEC is bad if directionally you can't agree on it.

• 1:10:18Lenny Rachitsky?

  That's a great tip. Along these same lines, I know you're a big fan of platforms and building a platform to run experiments versus just one-off experiments. Can you just talk briefly about that to give people a sense of where they probably should be going with their experimentation approach?

• 1:10:32Speaker 00

  Yeah, I mean, so I think the motivation is to bring the marginal cost of experiments down to zero. So the more you self-service, right? Go to a website, set up your experiment, define your targets, define the metrics that you want, right? People don't appreciate that the number of metrics starts to grow really fast if you're doing things right. And being you could define 10,000 metrics that you wanted to be in your scorecard. Big number. So it was so big and people said it's computationally inefficient. We broke them into templates so that if you were launching a UI experiment, you would get this set of 2,000. If you're doing a revenue experiment, you would get this set of 2,000 if you're doing.

• 1:11:15Speaker 00

  So the point was build a platform that can quickly allow you to set up and run an experiment and then analyze it. I think, you know, one of the things that I will say at Airbnb is the analysis was relatively weak. And so lots of data scientists were hired to be able to compensate for the fact that the platform didn't do enough. And so, and this happens in other organizations too, where there's this trade-off. But if you're building a good platform, invest in it so that more and more automation will allow people to look at the analysis without the need to involve a data scientist. We published a paper, again, I'll give it in the notes with this, you know, sort of a

• 1:11:57Speaker 00

  nice matrix of six axes and how you move from crawl to walk to run to fly and what you need to build on those six axes. So if, you know, one of the things that I do sometimes when I consult is I go into the org and say, where do you think you are on these six axes? And that should be the guidance for what are the things you need to do next?

• 1:12:21Lenny Rachitsky?

  This is going to be the most epic show notes episode we've had yet. Maybe a last question. We talked about how important trust is to running experiments and how even though people talk about speed, trust ends up being most important. Still, I want to ask you about speed. Is there anything you recommend for helping people run experiments faster and get results more quickly that they can implement?

• 1:12:40Speaker 00

  So I'll say a couple of things. One is if your platform is good, then when the experiment finishes, you should have a scorecard soon after. Maybe takes a day, but it shouldn't be then you have to wait a week for the data scientist. To me, this is the number one way to speed up things. Now, in terms of using the data efficiently, there are mechanisms out there under the title of variance reduction that help you reduce the variance of metrics so that you need less users so that you can get results faster. Some examples that you might think about are capping metrics. So if your revenue metric is very skewed,

• 1:13:20Speaker 00

  maybe you say, well, if somebody purchased over a thousand dollars, let's make that a thousand dollars. At Airbnb, one of the key metrics, for example, is nights booked. Well, it turns out that some people book tens of nights. They're like an agency or something, hundreds of nights. You may say, okay, let's just cap this. It's unlikely that, you know, people book more than 30 days in a given month. So that variance reduction technique will allow you to get statistically significant results faster. And a third technique is called Cupid, which is an article that we published. Again, I can give it in the notes, which uses the pre-experiment data to

• 1:14:02Speaker 00

  adjust the result and we can show that you get the result as unbiased, but with lower variance in health, hence it requires fewer users.

• 1:14:10Lenny Rachitsky?

  Ronny, is there anything else you want to share before we get to our very exciting lightning round?

• 1:14:14Speaker 00

  Uh, no, I think we've asked a lot of good questions. Uh, hope people enjoy this.

• 1:14:20Lenny Rachitsky?

  I know they will.

• 1:14:21Speaker 00

  Lightning round.

• 1:14:22Lenny Rachitsky?

  Lightning round. Here we go. I'm just gonna roll right into it. What are two or three books that you've recommended most to other people?

• 1:14:29Speaker 00

  There's a fun book called Calling Bullshit, which is despite the sort of name, which is a little extreme, I think, for the title, it actually has a lot of amazing insights that I love. And it sort of embodies, in my opinion, a lot of the Twi Man's law of showing that things that are too extreme, your bullshit meter should go up and say, hey, I don't believe that. So that was, that's my number one recommendation. There's a slightly older book that I love called Hard Facts, Dangerous Half-Truths and Total Nonsense,

• 1:15:05Speaker 00

  by the Stanford professors from the Graduate School of Business. Very interesting to see many of the things that we grew up with as sort of well-understood, turn out to have no justification. And then so with Stranger Book, which I love, sort of on the verge of psychology, it's called Mistakes Were Made, but not by me.

• 1:15:30Speaker 00

  About all the fallacies and that we fall into and the humbling results from that.

• 1:15:37Lenny Rachitsky?

  The titles of these are hilarious and there's a common theme across all these books.

• 1:15:43Lenny Rachitsky?

  Next question, what is a favorite recent movie or TV show?

• 1:15:46Speaker 00

  So I recently saw a short series called Chernobyl on the disaster. It, I thought it was amazingly well done.

• 1:15:56Lenny Rachitsky?

  Yeah.

• 1:15:57Speaker 00

  Highly recommend it, you know, based on true events. You know, as usual, there's some freedom for the artistic movie. They, it was kind of interesting. At the end, they say, this woman in the movie wasn't really a woman. It was a bunch of 30 data scientists, not data scientists, 30 scientists that in real life presented all the data to the leadership of what to do.

• 1:16:22Lenny Rachitsky?

  I remember that. Fun fact, I was born in Odessa, Ukraine, which was not so far from Chernobyl. And I remember my dad told me he had to go to work. They called him into work that day to clean some stuff off the trees, I think ash from the explosion or something. It was like far away where I don't think we were exposed. But, uh, but yeah, we were in the vicinity. That's pretty scary. My wife thinks I've, every, every time something's wrong with me, she's like, oh, that must be a Chernobyl, Chernobyl thing.

• 1:16:51Lenny Rachitsky?

  Okay, next question. Favorite interview question you like to ask people when you're interviewing them?

• 1:16:55Speaker 00

  So it depends on the interview, but I'll, I'll give you, when I do a technical interview, which I do less of, but, uh, one question that I love that is amazing how many people it throws away for languages like C++ is tell me what the static qualifier does. And for, for multiple, you know, you can do it for a variable, you can do it for a function. Uh, and it is just amazing that I would say more than 50% of people that interview for an engineering job cannot get this and get it awfully wrong.

• 1:17:31Lenny Rachitsky?

  Definitely the most technical, uh, answer to this question yet. I love it. Okay. What's a favorite recent product you've discovered that you love?

• 1:17:39Speaker 00

  Blink cameras.

• 1:17:42Speaker 00

  So, uh, a blink camera is this small camera. You stick in two AA batteries and it lasts for about six months. They claim up to two years. My experience is usually about six months. But it was just amazing to me how you could throw these things around in the yard and see things that you would never know otherwise. You know, some animals that go by. We had a skunk that we couldn't figure out how it was entering. So I threw five cameras out and I saw where he came in.

• 1:18:18Lenny Rachitsky?

  Where'd it come in?

• 1:18:19Speaker 00

  He came in under a hole in the fence that was about this high. I cannot, I have a video of this thing just squishing underneath. We never would have assumed that it came from there, from the neighbor. But yeah, it's, these things have just changed. And when you're away on a trip, it's always nice to be able to say, you know, I can see my house, everything's okay. At one point, we had a false alarm and the cops came in and had this amazing video of how they're entering the house and pulling the guns out.

• 1:18:56Lenny Rachitsky?

  You gotta share that on TikTok. That's good content.

• 1:19:00Lenny Rachitsky?

  Wow. Okay. Blink cameras. We'll set those up in my house ASAP.

• 1:19:04Speaker 00

  Yes.

• 1:19:05Lenny Rachitsky?

  What is something relatively minor you've changed in the way your teams develop product that has had a big impact on their ability to execute?

• 1:19:13Speaker 00

  I think this is something that I learned at Amazon, which is a structured narrative. So Amazon has some variants of this, which sometimes they go by the name of a six pager or something. But when I was at Amazon, I still remember that email from Jeff, which is, "No more PowerPoint. I'm going to force you to write a narrative." I took that to heart and many of the features that the team presented, instead of a PowerPoint, you start off with a structured document that tells you what you need, the questions you need to answer for your idea, and then review them as a team. At Amazon, these were like paper-based. Now it's all, you know, based on Word or Google Docs

• 1:19:56Speaker 00

  where people comment. And I think the impact of that was amazing. I think the ability to give people honest feedback and have them appreciate and have it stay after the meeting was, you know, and these notes on the document, just amazing.

• 1:20:13Lenny Rachitsky?

  Final question. Have you ever run an A/B test on your life, either your dating life, your family, your kids? And if so, what did you try?

• 1:20:20Speaker 00

  So there aren't enough units. Remember I said unique 10,000 or something to run true A/B tests? I do, I will say a couple of things. One is I try to emphasize to my family and friends and everybody this idea called the hierarchy of evidence. When you read something,

• 1:20:39Speaker 00

  There's a hierarchy of trust levels. If something is anecdotal, don't trust it. If there was an experiment, it was observational. Give it some bit of trust. As you get more up and up to a natural experiment and controlled experiments and multiple controlled experiments, your trust level should go up. So I think that that's a very important thing that a lot of people miss when they see something in the news is where does it come from? I have a talk that I I've shared. Of all these observational studies that people made that were published. And then somehow a controlled experiment was run later on and proved that it was directionally incorrect.

• 1:21:21Speaker 00

  So I I think there's a lot to learn about this idea of the hierarchy of evidence and share it with our family and kids and and friends. And there's a now, I think there's a book that's based on this, like how to read a book.

• 1:21:34Lenny Rachitsky?

  Well, Ronny, the experiment of us recording a podcast, I think, is 100% positive. P value 0.0. Thank you so much for being here.

• 1:21:44Speaker 00

  Thank you so much for inviting me and for great questions.

• 1:21:47Lenny Rachitsky?

  Amazing, I appreciate that. Two final questions, where can folks find you online if they want to reach out? And is there anything that listeners can do for you?

• 1:21:55Speaker 00

  Finding me online is easy, it's LinkedIn. And what can people do for me? You know, understand the idea of controlled experiments as a mechanism to make the right data-driven decisions, use science. You know, learn more by reading my book if you want. Again, all proceeds go to charity. And if you want to learn more, there's a class that I teach every quarter on Maven. We'll put in the notes how to find it and some discount for people who managed to stay all the way to the end of this podcast.

• 1:22:30Lenny Rachitsky?

  Yeah, that's awesome. We'll include that at the top so people don't miss it. So there's gonna be a code to get a discount on your course. Ronny, thank you again so much for being here. This was amazing.

• 1:22:39Speaker 00

  Thank you so much.

• 1:22:40Lenny Rachitsky?

  Bye, everyone. Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com. See you in the next episode.