Pod Insights a16z Podcast a16z Podcast | Where's My Stuff? -- The Lowdown on Logistics and Ops About this episode The ease with which we can order anything online masks the tremendous complexity of getting that item to your door in a day or two (or less). Andreessen Horowitz's Jeff Jordan leads a discussion with four experts in ecommerce logistics and operations -- the tough business of getting things from one place to another.
Transcript 10,638 words · 11 speakers
0:00 Speaker 5 Welcome to the A16Z Podcast. I'm Michael Copeland. This segment of the pod centers on all things logistics. How packages get from point A to point B. The warehousing in between, the security, the networking, the data, the predictions, the whole ball of wax. Moderating this discussion is Andreason Horwitz's own Jeff Jordan. And joining him are four experts in the space. Mike Smith, COO from Stitchfix. Laura Duke, who runs enterprise supply chain at Zazzle, Mike Swartz, who's a consultant and a longtime expert in this space. And then finally Dwight Gaston, Chief Delivery Officer at Julep. Jeff Jordan kicks it off, and then Mike Smith comes back with the first response.
0:46 Speaker 4 For each of you, can you spend a little time describing kind of what your business does and why your job is hard. So what unique attributes make it hard. And uh Mark, you can pick whatever experience you want to lob in on that one. Yeah.
1:02 Speaker 0 Okay. Yeah. Uh so our business is again, it's a personalized styling service for women. We use data and analytics uh to be able to uh understand a woman's style, and we send them a box of five things that should work for them, and they can try it on in but
1:16 Speaker 4 like highly personalized.
1:17 Speaker 0 Highly personalized. There's not been one fix of all the fixes that we've sent out that have been the same. So there's always uh different. And uh it is it is meant to be able to try it on the comfort of your own home, the context of your own closet, almost like a fitting room, like a personal fitting room, and you uh whatever you keep uh or whatever you like and you want to buy online, you you buy online and you keep and you send back what you don't want. You get a prepaid envelope to send back what uh what you don't want to do. That doesn't work for you. No. Uh it's funny that you say that because it, you know, I think uh in its in its form, sort of sending stuff out, getting stuff back and having it be sort of this whole reverse logistics thing, and a lot of investors passed on us because they're like, that's a big returns problem.
2:06 Speaker 0 And we're like returns is part of the model, like everyone has returns, and we just sort of operationalize that and it works for us. There's so much of my job today that seems so less complicated than the job I had at Walmart because of the range of product that we had at Walmart from diamonds to ready to assemble furniture to big screen TVs to chimneys. We sold chimneys to kayaks. I once woke up in the middle of the night thinking we were selling yachts, and I'm glad we never sold that. But there's there is a complexity to the send out, get back and making sure that the stuff that comes back is hasn't been like worn out at a party. Um so we have good quality kind of checks as it's coming back. Uh and then the inventory turns really, really fast. And so you have to have a pretty dynamic warehouse.
2:53 Speaker 0 And I think the other complexity of the model is just managing the growth. It's been uh crazy growth. We've gone from, I'll give you just the warehouse numbers. When I first started the company, uh we did pick pack and ship out of 500 500 square feet, um which was half of our office. Then two months later, uh we moved to uh a thousand square feet, then two months after that we moved to 9,000 square feet, and now we're in uh 450,000 square feet uh in two different warehouses. So managing that growth and scale and making sure that you know you you make really good decisions about what when you change your process or think about adding systems into when you're growing that much and the business is changing as much as it has, uh, I think has made the job complex and keeps me up at night.
3:39 Speaker 0 So
3:40 Speaker 6 good to hear
3:42 Speaker 3 so um so Zazzle. I work at Dazzle and um Zazzle is a customized marketplace. So it's um a marketplace or a platform where you can put any of the world's content on any of the world's products. So you can imagine um taking a picture of your children, putting it on your iPhone case, you know, taking artwork and putting it on a print. And and um really it's a very dynamic model that has um a few tiers. It has it has uh our makers, which are the um outsource partners, as well as our own internal production, which still does a large percentage of our business, um balanced out by so we have the production being made, and then we also have our contributors that are contributing art, and that can be anybody from our large partners like Disney or Marvel to um to individual contributors that are contributing artwork and they're putting artwork on um on those different products, right?
4:44 Speaker 3 And um so um I think the complicated part about um my job ends up being, you know, we have everything from a very large organization that is shipping that product to to very small makers that are potentially only shipping a couple items a week, and we want to create really an end-to-end customer experience that um is the same for everybody regardless, and so that those small makers can use um this same logistics model, which really is um uh creating more of kind of a peer-to-peer uh logistics system that doesn't really exist much out in the marketplace today, so that a customer or a small maker can ship to a customer and leverage um the contracts and the technology of the larger platform and um really kind of have this great end-to-end customer experience.
5:37 Speaker 3 So it's um that part is pretty complex, but my returns is easy because customizable market doesn't have much in the way of returns. So that just kind of shows the difference in um different types of operations and complexities and how different they can be.
5:56 Speaker 4 What company you want to pick?
5:58 Speaker 1 Well, you know, I I can talk in general. I it um you know, I I'm in I I do a lot of problem solving, and I get called into a lot of situations that are challenging, both culturally um and just the environment so different, especially when I go to different countries. And um what's counterintuitive is that it's really the soft stuff that's hard, it's not the hard stuff that's hard. So, you know, figuring out the process and figuring out what needs to get done is actually the easy part of the job. The hard part is getting the buy-in and getting people to recognize for themselves what needs to get done, getting them to acknowledge that there's a problem. Um and then the execution on that. And I I find over and over again that uh especially with startup companies, that where they tend to fall short and really need a lot of support is in something that they think they're good at, which is execution.
6:54 Speaker 1 So I spend a lot of time of uh if I look at the total time spent with a client, um spend a fair amount of time building trust and building rapport, a little bit of time identifying problems, and a lot of time helping them through the soft skills side. You know, execution, buy-in, uh the organizational change part of it. That hasn't changed since my Amazon days. It's the same, same issues there.
7:19 Speaker 4 Yep,
7:20 Speaker 1 I bet. But uh that's probably the biggest challenge that I face. Interesting.
7:24 Speaker 4 Right.
7:25 Speaker 2 So for me, I think uh a lot of the current complexity is just the stage we're at. Um I I joined six months ago, and before that there was nobody really focused on operations, and so the the basics just aren't there. So we're we're yeah, everything's breaking. So just trying to figure out how do we build out our you know call center and get that you know, the the leadership and and structure in place to handle calls appropriately. How do we uh deal with fulfillment? We were operating out of a small fulfillment center, about 11,000 square feet in Seattle, and uh no systems. It was just a really really um really uh struggling. So we made the decision on my second day there to outsource fulfillment, and so um now we're shipping out of a a third-party fulfillment um uh partner in in Columbus.
8:17 Speaker 2 And so that's a you know a big um big uh time suck right now. But um longer term, I think what's gonna be really fun to really um uh figure out is is the supply chain because we've got uh we're we're generating about 30 products a month that are brand new, and some of those are just new nail colors, but but a lot of those are new other cosmetic products and which
8:41 Speaker 4 by the way is revolutionary in the beauty business. Typically beauty companies do two sets a year, maybe because they you know the you know the we're not need to manufacture it, need to support it with marketing, you know, just so done, you know, it's a beautiful target for digital disruption.
8:55 Speaker 2 Right. So it's um so it's it's it's a huge part of our model, but it's uh we we just need to make sure we're we continue to get it right. And um the other challenge with that is that we have a hard deadline. So every month we put our new products into our subscription box, and so every month, uh you know, by the middle of the month, we've got half that new product in our warehouse ready to ship out. And um and so that is just you know fraught with you know risk. So uh
9:21 Speaker 4 a lot of nail polish flew first class before you came. Yes, it's trying to try to put
9:28 Speaker 2 it back
9:29 Speaker 4 on the
9:30 Speaker 2 my
9:30 Speaker 4 favorite julep challenge is the the community is a rabid community, and they're and they're just wait they've ordered their ni, you know, they've customized their box at the last minute, and then they just won it three days later. And one thing that happens is nail polish technically is a hazmat, I guess. And so they found out was using air to deliver, and they went crazy and and stopped them. You know, no, no, no, you can't use air, you have to drive it. And it's just all of a sudden all the Mavens who are waiting for their product are you know, you were it's not getting there in time because The you you know the the postmaster general said he could they can't fly it. It was a disaster. That was day four of my board member I found out yesterday.
10:10 Speaker 2 Yeah,
10:10 Speaker 0 it's no that's
10:12 Speaker 2 so uh yeah, so it turns out you know the first class mail, which is a great way to get things across the country from Seattle at a low cost um at low cost, um, suddenly was no longer feasible. So we yeah, we had to ship our next subscription around using our our other low-cost method, which is DHL uh global mail, which takes up to 10 days to get across the country. So we had this fur this like rabid um customer base was just up in arms. They actually started a petition on change.org um demanding that Julep break up with DHL. So we couldn't get it to uh get our business to ship to Ohio soon enough. Um
10:56 Speaker 4 so a lot of you came from fairly large enterprises and joined your enterprises when they were fairly young. Uh talk about that adjustment.
11:05 Speaker 0 Yeah, for me it was uh it wasn't that hard. I mean, uh again, my exit was a little complicated, but uh most of the things I'd worked on at Walmart.com again, I joined when we were there about a hundred people, we had just done 150 million dollars in revenue. I left, it was a lot bigger, but um most of what I worked on were small company things, and I just realized after nine years there, you know, sort of getting like Mike talks about getting things done through influence and change management that I had had enough. Uh I really wanted to have more direct impact. Um the you know, sort of warehouse staff at Walmart.com versus where I was in my position, there was at least eight layers, and so I was being held accountable to things that you know would happen in a warehouse that you know they didn't even know my name and I didn't really know their name.
11:57 Speaker 0 Uh and that became more challenging. So it wasn't a big adjustment to go small for me. Um I took a huge pay cut. That was probably the the biggest adjustment that I had done the math and said I can make it work and I made it work, but boy, the first couple months of paychecks, I'm a big wine person. Um and so not being able to buy my wine was a little bit.
12:18 Speaker 4 They sell it in boxes.
12:21 Speaker 0 I have heard that. My parents still buy it that way. Yeah, yeah, you've got
12:28 Speaker 4 gone to hell.
12:29 Speaker 0 Um but but you know, sort of sitting behind like I sat right next to our CTO for the first six months. And so if something was breaking in the warehouse operation in the systems, I'm like, Jeff, we got to get this fixed. And that was so empowering and and much better than you know, trying to get something on the product roadmap at Walmart.com and waiting for six months to a year or longer to actually see it happen. Uh so it wasn't a big adjustment for me. It's pretty it was pretty easy. Was it a big
12:57 Speaker 4 adjustment for anyone else?
12:59 Speaker 3 I I think there was there was adjustments. I mean, if for anybody
13:03 Speaker 4 in HP. For
13:05 Speaker 3 anybody that has gone to a startup, I mean, you guys all know it's not all sunshine and unicorns, right? I mean, uh there's challenges, and I think I think ironically, the biggest challenge for me was actually what I went looking for, which was creativity. Now, granted, I I left um I left eBay, I came to Zazzle, and there was only 30 something employees and you know, a few million in revenue, and it looked a lot more like chaos than creativity to me. So I think that adjustment of being able to, you know, even though I I had this kind of mindset in this of um big company and working through significant processes that you know were well ingrained and a lot of them were holding us back and slowing us down, but at the same time, um I could really had to kind of sit back, resist the urge to fix everything and actually kind of really recognize the creativity and build on that.
14:03 Speaker 3 Now I do think that there's a fine line between um creativity and chaos, and I think um kind of you know, managing that line still takes some finesse. But um but there's there's uh you know Dazzle's a lot bigger now, but um but we're still kind of heart and soul startup, and I can't say that the chaos is gone from my life, but uh but I definitely have uh found an ability to kind of eliminate those chaotic elements and kind of focus on the creative side. So
14:34 Speaker 4 Megan Dwayne, when you first show up small company, there's no team. So ha how do you go out about building a team? And then uh you know, particularly how do you uh balance the rewarding the people who brought you to the party, you know, the young aggressive, you know, type versus you know, bringing in talent from the outside, you know. I
14:54 Speaker 1 I think one of the I was thinking about some of the challenges, it's kind of goes with your first question, is the skills and the people that got you there won't get you to the next level. You know, recognizing that, and I think having conversations with people early um is a big part of that challenge. You know, that it takes a certain type of person in the early stages of startup to succeed. But that person, if you especially if you're on a really uh high trajectory growth curve, they're not gonna be necessarily the same people that take you to the next level. And you can risk losing some really good people if you don't talk to them early about how their role is going to change and how they're gonna have to go maybe more narrow and deep and not be broad in their responsibilities.
15:42 Speaker 1 All they see is I'm losing control, I'm losing control, and I feel less impactful. Um so I think p I think people in general is the biggest shock, you know. Um even in the the time that I spent at Amazon and QBC were was still very entrepreneurial early stage kind of companies. They were growing fast. And I underestimated how much time you have to spend recruiting and hiring. You know, I I would spend 80% of my day sometimes interviewing and and trying to find the right people. And even at Amazon, after my facility was staffed, we started we had a five million square foot expansion in uh 1999, and um I was recruiting for other sites. So I would recruit all day long and then do my job, you know, from like six to midnight doing my regular job.
16:27 Speaker 1 So I think that was that was really a a big stressful part of it, you know, is putting the time into hiring and finding the right people, making sure you don't lose the good people that are there and helping them understand what their role is going to be going forward, and then doing your day job. Yeah.
16:46 Speaker 4 It sounds like you just walked into a green field.
16:48 Speaker 2 I did. Yeah. Yeah. And so um for for me, my I I had a team in each area, each of the three areas, and or I have, and so it wasn't so much that I needed to you know build a team from scratch. So I so that was nice because I've been able to kind of evaluate people uh in in customer service. Uh the team was so incredibly junior that it was just there was a clear need to bring somebody in. And the and the nice part was everybody loved that. The the junior people who were trying to you know deal with this um you know chaotic situation suddenly had this you know role model and mentor. So that was that was great. It was easy decision to bring somebody in because it was clearly necessary and it wasn't disruptive. Um for the other teams, what I'm trying to do is figure out who who should stay, who who can take different roles, and then others in the organization have have been coming expressing interest in joining operations.
17:41 Speaker 2 So it was kind of a I'm trying to not make decisions until I kind of suss out kind of what what internal candidates there are. Um so just trying not to, you know, um uh change things dramatically by kind of bringing in a bunch of new people because that can be.
17:57 Speaker 4 Uh Ben Horowitz has written a blog that I really like about how every the j in these hyper-growth companies, the job requirements change so fundamentally so quickly that it basically everyone in the organization has to almost reapply for their job each year. You know, just it's that dynamic. So that's something that stayed with me for a while. How do you how do you guys define success in your operation? I mean, just what what what is what is success? How do you how are you held accountable? How do you hold your teams accountable?
18:25 Speaker 2 You want to start
18:26 Speaker 4 it up? Yeah, why
18:27 Speaker 2 don't you go ahead? Um I mean I think it operates.
18:30 Speaker 4 Everyone doesn't have to answer every
18:31 Speaker 2 question. Let's
18:32 Speaker 4 go for it, yeah. Um
18:33 Speaker 2 this might be oversimplifying, but in operations, I think it all boils down to either you know cost control, um doing things faster or doing things at higher quality. And so it's determining for what each area, what are the priorities? I mean, all of those are good, but um s really deciding um which is most important and then um figuring out the right metrics and and and just a couple metrics can be really helpful to focus everybody, but um for me it's it's it's it's just setting those and then having everybody really focus on managing those and um and dealing with it as we grow.
19:07 Speaker 4 Trillion metrics. What do you pick?
19:09 Speaker 2 Um so for us, um I think we've got um the call center we wanted to measure customer satisfaction and then um service level. Um and those are the the two things that are most important. Right.
19:22 Speaker 4 Uh efficiency and quality.
19:25 Speaker 2 Right. And then uh in uh operations or in fulfillment, it would be on-time shipping and then cost per order to manage that. Uh and then we're still figuring out supply chain.
19:39 Speaker 1 I uh you know, I get asked this question a lot, and and the question usually gets uh phrases, well what what do they do at Amazon or what's Amazon's metric, or what's best in class in X, Y, and Z. And my response is always the same. It's like, why do you care? Right? So what's really important is I think the measure of success is do you Do you understand what's driving your business? Because what's driving yours might be different than what's driving Amazon or another company? Two is are you doing better today than you are than you were yesterday? And three, how aligned are the people in the organization around these two two fundamental questions. One is what's important. And if you ask that of everybody in the organization from the founder all the way down to the guy picking and packing the order, you should have alignment.
20:30 Speaker 1 And often you don't when you ask what's important. And two is how are you doing right now? And if people can't answer those two questions, you can have a lot of chaos. So I I define success as do you have alignment and are you doing better each day? And are you are you taking steps to understand and reduce variation in each of those metrics? So it's not the mean. You know, and on average, I can have a hand in a bucket of ice and one in a bucket of hot coals. On average, I'm comfortable. You know, that's not the truth. And uh even well-established companies I work with don't understand variation. They do everything off of means. Yeah. So
21:10 Speaker 3 the only other thing that I'd add from a metrics perspective and is that it's it's not only what you measure, but when you measure it. So if you're looking at, let's say you're looking at your metrics monthly and you're a startup, um, you're moving so quickly. That's like driving down the road at 90 miles per hour and only looking in your rear view mirror. You know, you've really got to look at the metrics as current as possible and even start to find metrics that are leading indicators. You know, so when you're looking at on-time delivery, not just looking at on-time delivery, also looking at backlog, which might tell you what delivery what what delivery times are going to be in the future and what customer inquiries are going to be in the future, so that you can start looking at metrics that actually start to look a little forward and as current as possible rather than just waiting a month when the damage is already done.
21:59 Speaker 4 Um what kind of uh mistakes do you see startups making on in the operational front? I mean, what are you know pattern recognition says?
22:09 Speaker 0 Well, it Dwight gave up a good a good one, is is just not focused on it fast enough, I think. I mean uh I got lucky in that, you know, Katrina, the founder and sort of the investors early stage said this needs to be a core competency because we're sending stuff out and getting stuff back, and I I mean it's shocking to me now, and I'm very grateful that they sort of hired a COO as employee number four, you know, and uh that's very early to do that. But I think it's not focusing on it fast enough. Um and then you know, there's times I think when startups and even bigger companies where you start to lose, like the senior people start to lose connection with the work that's being done, and I think that's a mistake too. There's got to be ways to that you're always staying connected so you understand both on the like what Mike said on the softer side, what kind of culture do I have, and uh are we um you know, do they understand sort of the vision and the purpose of what you're trying to accomplish every day because those jobs are really hard and keeping them motivated to do them and have low turnover is an important part of the job.
23:13 Speaker 0 Uh and I'd say the other part is just like understanding where you have process flaws or productivity, you know, where you're not meeting productivity goals. Um so you need to stay close to it. And I think there are some leaders that like it's not my problem, like it's my you know, supervisor on the floor problem. So I always walk the floor um and stay as close to it as I think. We're gonna come back to that one. Okay.
23:37 Speaker 4 Other uh no, I think it's really important. Other uh other uh common uh laws.
23:43 Speaker 3 Sure. I I think um one thing that I've seen is um uh not negotiating. So so when you have relationships, there's a kind of this with startups because we're moving so fast is that you you get this kind of grab and go philosophy as far as setting up vendors and suppliers and big huge partners that um you're gonna try to get something in place really quickly and um put it in place so you can focus on all the bigger issues. And um there's a lot of times you think you're small, you don't have a lot of volume, there's not a lot of leverage there for negotiation in the first place. But I think what we're missing is the fact that even the biggest partners know that um the value of startups in this economy and they can look years out. So if you just take the time to get the right people in the room and actually kind of talk through your vision, you can ask for what the big guys get and you'll be surprised.
24:37 Speaker 3 And it doesn't take much time, and um you'll get significant payback in terms of margins and uh customer experience.
24:45 Speaker 4 Set a twice a year meeting with UPS and just talk to them about how much how big how big you've gotten since the last time you were there. And it's it is fairly remarkable.
24:55 Speaker 1 Yeah. I there's two things I even in even in early stage startup companies, uh the two things that come to mind. One is um really understanding and staying focused on your customer. Who are they? You'd be surprised how often I ask who the customer is, and I get 50 different answers. Uh so getting clear about your customers and making sure that everything you're doing is lined aligned with their needs. And the second thing that I find quite often is um a poor uh or maybe the wrong attitude about defects and about what's you know, things that are wrong. People tend to hide them and uh they get in trouble if defects are present and they're always there, right? I mean, nobody's perfect. So I think that's a fatal mistake early is not having a uh not putting systems in place and metrics and really embracing problems for what they are.
25:48 Speaker 1 You know, when you drive them down, you don't see them, you can't do anything about them. So I think developing a strong culture or attitude toward making defects public, making them something everyone can center on is a very powerful thing.
26:04 Speaker 2 Yeah, and um when defects occur, um making sure that the pro the root cause analysis is not attacking the people involved, it's attacking the process. Is it it um it's so important to make it safe to raise these mistakes and talk about them because if you fix the fix the root cause is about fixing the process, it's not about you know somebody making a mistake. If you have that culture, you can really drive much better improvement.
26:27 Speaker 4 So to give putting a mirror up, any personal mistakes you've made along the way in the career that you think might be illustrative for people to learn from.
26:39 Speaker 0 Yeah, I've got a really direct one. I I got on the phone with a customer when I uh was not in the what not in a good place to be talking to a customer. I was too sort of logical, rational about dealing with what they were asking for, and I would it wasn't good. I mean, it almost turned into sort of a big PR issue for Walmart, and so I was uh glad that uh dish fix and opportunity came along. That's you. No, I just so I think yeah, I was it was it was a challenging situation. It was, you know, I just I wasn't trained to handle an escalated call at that level, and I should have opted out. Yeah. So
27:23 Speaker 2 I think overdelegating is one mistake I made once, and it was pretty painful. Um it relates to like understanding as you're growing fast who's kind of keeping up with the the needs of the company and who's falling behind. And we were doing a uh a big expansion in our fulfillment center, Blue Nile, and um we needed to buy a second vault, and we have these big you know uh very secure balls we built that we built. And long story short, the person I put in charge of that didn't you know do a lot of research and the salesperson didn't do a lot of research, and they bought a vault that would have fallen through the um floor of our warehouse. Um what
28:02 Speaker 4 was the problem with that?
28:05 Speaker 2 So you discovered this when it when the 25 trucks were driving from Colorado with all of the pieces and um and after a really ugly negotiation with the vault manufacturer, we had it sitting outside our warehouse for six months while they tried to find another buyer. So it was a nice reminder of a huge expensive mistake with the
28:23 Speaker 4 Which is always good in Seattle. It doesn't rush for it as well.
28:29 Speaker 1 I I think one of my the one of the lessons that come to mind uh very quickly is you know I used to have an advertis adversarial relationships with finance, right? And to the point where I'd call them CF nos instead of CFOs. And uh I learned early on that it's better to involve them early and get them to partner in terms of coming up with what your true costs are and um and making them part of the of the solution rather than somebody that you present to at the end to debate with you. And that that was very helpful on a lot of levels because they were more involved. I think we got better results. Yep, absolutely. And uh it was better for the business.
29:12 Speaker 4 Good. Um companies almost take operations for granted, it's a bit out of sight, out of mind, you know, it just it i it it happens till it doesn't. And uh yes, what are great what are some mechanisms to try to keep, you know, the had to have the entire company keep at that operational uh vantage point and you know, thoughts in mind that is that it's a holistic company that you know that you know and not a detached, you know work still doesn't work.
29:42 Speaker 0 For me, two things come to mind. Uh one is just making sure that at sort of the executive table that you you are seen at the same sort of light and peer as like the chief merchant or chief marketing officer. A lot of that is like your own leadership style and and how you You present how important your group is, not in a political self-promoting kind of way, but more just how this is impacting the business and why it matters. That's one. And then a very tangible one that I've I've wanted to do in my old job and have done at Stitchfix is having customer service agents that are always in headquarters so that people and sit amongst the business so they can everyone can hear defects and exceptions and things that aren't working.
30:29 Speaker 0 So that also keeps it front and center.
30:31 Speaker 4 That feedback loop is critical.
30:33 Speaker 0 Yeah.
30:34 Speaker 3 I think I think another thing that has worked for me personally is having some discussions around a shared metric, using something like gross margin, or gross margin isn't owned by one group alone. It's owned by so many and having all the department heads and key stakeholders in discussions so that marketing as they're trying to hit their revenue target, they can have that discussion and say, I, you know, I want to send out this promotion with this discount and operations is right there in the room to say, okay, well, this is what that does in terms of capacity. Maybe we could talk about using a different product that we have more capacity on, things like that. So you can have those kind of cross discussions in the same room and they can understand the challenges, and then they get to considering those as they're driving the business, even when they're not in the same room.
31:25 Speaker 7 Yeah.
31:25 Speaker 1 I'm going to echo one of the things Mike said. I I think transparency is huge. And if you look at call centers, uh a lot of companies look at call centers or contact centers like they're a cost center. And really call centers don't generate calls. They're responding to things that go wrong somewhere else in the company. So I think creating transparency there is huge. The other thing is finding opportunities, even as you get very large for everyone to participate in what goes on in operations. So during holiday peaks, making sure that they be that inviting people from the front office, that technology people especially invite them into the warehouse, invite them into operations. Make sure that your merchandising team, that somebody's there to physically touch and see the products and how it's coming in from the vendors at least one day a week.
32:17 Speaker 1 Making sure you just kind of keep the doors to your operations open is huge. I know when I was at Amazon, more code got written in the fourth quarter when we would get deployed to the fulfillment centers than any other time of the year. More things were fixed. So very valuable to keep those models.
32:36 Speaker 3 I laugh only because be careful what you ask for and how you implement, because I can tell you our CFO is really bad at making ties. So when we do that, we just have to be very careful about what you give them if you need some jobs. That's a good point.
32:51 Speaker 0 I got a good quick story on on this point too. Uh so uh it was a Friday afternoon, and one of our chief engineers was in town and he went down to the warehouse and just wanted to kind of pick pack and ship and do some returns. Um he came to me at like 5 30 that day and was really upset that we had put you know these associates through a really bad, very manual returns process. Uh by Saturday, he had uh RF gun that he had ordered. Uh by Sunday he had written code that basically could return process returns in a system wide way. Monday we rolled it to the whole warehouse. At that point, we only had probably eight people working returns. But when you get that level of engagement, you actually get to see it, there's huge returns. Um and he just thought I was the worst operator ever for letting them go through that for even you know for six weeks, and uh there's just huge dividends for take it being transparent.
33:45 Speaker 0 Take it to the floor,
33:46 Speaker 1 don't take it out of the conference room. Yeah. That
33:48 Speaker 4 would be great. And I'll open up for uh questions after this uh last one. Um security privacy, uh brave new world out there. Uh how's that uh what what complications does that bring?
34:00 Speaker 2 Uh so at Blue Nile, we had a big uh you know we were a big target for fraud, so credit card fraud and um and is that kind of part of the yeah and um so we really had to be world class at that. And um there are a lot of tools out there now. People
34:15 Speaker 4 wanted to steal diamonds? They did, yeah.
34:18 Speaker 2 Who'd have thought? Um and so for us uh I think most companies have a kind of automated if they have a fraud prevention process, there's an automated screen where orders go through and kind of get a score to see you know how risky they are. And we had that set to be very, very sensitive so that almost all of our orders would get kicked into a manual review team. And so we had four to five people who would look at these orders and they would have they had uh had some kind of ad hoc tools to kind of do some further research. Um we would sometimes call the cardholder, um, sometimes we would find a um a different number to call that person to make sure that we didn't call the number that the person who's ordered it and gave.
35:04 Speaker 2 Um we um so it was just we had a a whole very um you know kind of hardcore approach. But I would say that there are a lot of tools out there now, third-party tools that can help you uh configure configure the the rules and such to match your business needs.
35:21 Speaker 1 Yeah I I have two comments. I mean, I ran a jewelry vault for many years and uh you know always had you know between 150 and 250 million dollars in inventory. So I used to joke that you know my job was to walk around every day thinking like a thief. You know, how would I steal? And even though uh that was my those are my times at QVC, even though the company invested millions in surveillance tools and all kinds of seen and unseen cameras, we had uh metal detectors that would weigh your both your ferris and non-ferrous metal in your body content. And it was programmed into your badge. And um even with all those tools, the most effective tools were human.
35:59 Speaker 4 Whoa! Programmed into your badge. Yeah, so you
36:02 Speaker 1 when you were when you were signed when you got your access badge, your base metallic weight was programmed into your card. So if you were over or under, you would go through a secondary screen. Wow. And it the structure had to be three feet off the ground, it couldn't be near swinging doors. It was crazy. But even with all that money and proximity alarms on doors, you couldn't just hit uh an exit door. Um there were delays and all sorts of all sorts of things. But the most effective security systems were people. And because nothing does a better job of pattern recognition than a human being. So having tools for humans to use and really relying on people systems is probably the best way to secure a facility.
36:49 Speaker 1 And the the other thing I took away from from that was not designing your security systems and your hooks for the 99% that aren't a problem, right? You don't want to punish 99% of your customers for the one percent that's gonna get you. And uh, you know, kind of keeping that in mind that you don't want to over-engineer it so that you're creating a bad experience for your customer. Right.
37:14 Speaker 2 On the uh network security side, Blue Nile um used to have hackers that would hire hackers to come in and try to uh you know get through our systems, and that was a real helpful way to get a good um, there was more vulnerability.
37:28 Speaker 4 We started now marketplaces or white hat hackers come and pound on your uh on on your infrastructure. I mean smart. Now, as an eBay and PayPal, I mean hundreds of millions of dollars were running out the door, and so it was it was hand-to-hand combat. Yeah. It was really entertaining. That open it up to the floor. Questions.
37:47 Speaker 7 Um, I guess there's one trend in e-commerce that is about customization. There's another trend clearly on having your stuff same day. It doesn't really go together. We do both on an HP, and if we customize it up to get it to you tomorrow. Um, how do you think about that from a customer's perspective? I'm not so worried about the supply chain, but uh how do you determine customer value? I can give you a choice, but having a customer choose doesn't always necessarily work either.
38:10 Speaker 3 So everybody's looking at me. I think um I think that we have so we have that dilemma all the time. Um the immediate kind of instant gratification of this current on-demand economy, um, everybody is um an Amazon Prime customer and is used to getting things, you know, next day, two days. And um sometimes when you're actually making it, it takes longer than that. So I think we've done um from a customer perspective, I think we have found that um it's a lot about expectation setting. So it's about getting out there and really letting them know, hey, it's it's um it's so much storytelling at this point. What we're really getting to is now is saying this is coming from this maker, and you know, it's a small organization in Kenya that is going to be making this shirt, and that is what you're going to get.
39:04 Speaker 3 So that storytelling is something that really, you know, we've seen this happen in a lot of industries. We've seen this happen um in food from the you know um where people want something that's locally grown and then move to their table. Um and and that kind of storytelling you can really latch on to, and the customer gets that. So if you use that to really explain, here's some of the things that you're getting and where you're getting it from, you can then explain the extended, the extended time of customization a little bit better. Um, but at the same time, so one is setting expectations and explaining it, but then the other is shortening those times. So we um actively every um year and constantly are working on shortening our lead times for all of our production so that we can actually produce things um Most of them within 24 hours, a lot of them same day and get them there are a lot of products that we actually do deliver next day at this point.
40:08 Speaker 3 So I think it's striving to do a little bit of both.
40:12 Speaker 4 So HP can say you're manufacturing your computers in small villages in
40:17 Speaker 1 Canada. I think I think it's also creating some visibility into your supply chain and letting them see how it's progressing. It's you know, they don't want to just wait and have it, you know, magically show up on their doorstep. If it's going through different phases at appropriate times, you know, without overwhelming them. I think Domino's does as well.
40:37 Speaker 0 Yeah, I was gonna say, I mean, I I think that's the best app, and I think sort of setting expectations and and turning your supply chain out and showing the customer. I know it sounds crazy.
40:47 Speaker 4 Domino's is the best app.
40:48 Speaker 0 I
40:48 Speaker 4 well it's the best app for this. Yeah, yeah. Like sharing
40:51 Speaker 0 for sharing it. On our product roadmap, we have a dom Domino's you know, pizza tracker, Stitchfixing told. Totally. And I do think it's about setting expectations, and I do think they are misaligned. And I'm not a huge uh believer in the same day. I've I am a very contrarian, I feel like on same day. Uh I just it's hard for me. We did a same day test in 2003, 2004. Um we almost invested in a same day company way back then. They had already tested a lot of different products. I don't buy that there's a lot of product demand for same day. Uh I think that Amazon will do the same day for grocery because it makes sense on fresh. I don't I believe they'll do next day on everything else.
41:34 Speaker 4 Yeah. So you could do Apple and just promise four-week delivery. Yeah. Yeah. Yeah, yeah. Yeah.
41:42 Speaker 8 Yeah, I had the question regarding the leadership style. Everybody worked on big companies and then you guys moved to the startup environment. How did you change your leadership style in these new environments? If you did and
41:58 Speaker 0 one uh sorry, quick, because I was going to talk about this earlier, so it's top of mind. Um I had to change a lot because I was managing uh two folks initially, we were really small company that were 22 or 23 years old. Um and I think they when they heard the COO of Walmart.com was joining the company, they thought that I knew everything. And I'd have spent a lot of time telling them, and it's hard because it's like, you know, I'm telling you that I don't know everything. I'm telling you also I'm not an idiot. So like there's a balance between saying, I want you to challenge me. Uh I want you to sort of ask good questions and make sure we're making good decisions. But when we make a decision, you need to trust that I know enough that this is the right decision. Uh at Walmart, I was used to managing VP level people that you know, they challenged me and that was just the way it was.
42:49 Speaker 0 Uh I had to train and teach really young people how to be challenging in the right way. So it was a it was a definitely a style difference. Well,
42:57 Speaker 4 I mean we uh my experience um you when you're running big functions at big companies, you actually forget how to be proactive. Um you are reacting constantly, and the level amount of stimuli you're getting is just keeping the plate spinning is a major accomplishment. You show up a startup and you know it's all about proactive. You know, and so I I rem when I went from PayPal to open table, it's like, hey, someone should run the you know, the the financial planning team should run a model on this crickets. You're like, oh, there isn't one. I'm back on Excel. I mean, and that that would that's a tough you know, it's a tough adjustment.
43:30 Speaker 1 You know, at the risk of sounding like a contrarian, I I would say my leadership style didn't change. Um that but you know, I get this question a lot from guys, should I be um should I be a directive or should I be a uh collaborative kind of leader? And you know, I think it's situational and you have and maybe in a startup environment you have to be more directive sometimes and collaborative or vice versa. And I think it's just recognizing when which approach is most effective. You may have a bias toward one way or another, but you know, if the house is burning, you don't have time to call a meeting and say, what does everybody think? Which exit should we use? You know, you you gotta get out of the burning building. So um, you know, I would say personally, from my own perspective, you know, I've always been very hands-on, I've always gone to where the problems are.
44:20 Speaker 1 Uh I've always tried to communicate, and I'm always direct. You know, if I took the you know, I've played a lot of athletics and I think I learned a lot from the coaches I had. I never loved my coaches, but I respected them because they always told you if you're doing well quickly, and they told you if you're doing badly quickly. And if you did something that was going to hurt the team, you heard right away, right? Too often I think leaders lose that perspective, and they only give that kind of feedback when it's time for the annual appraisal or time for the annual performance review or semi-annual. And I think it's important to do that every single day. When you do this, this is what happens. These are the consequences. Don't do this, do more of that, you know. It's pretty basic.
45:06 Speaker 2 Yeah, I would agree it's situational. For me, it was it was uh my concern coming into a company after being somewhere for 15 years for 15 years was like how how do I balance the need to learn before you know deciding or recommending things versus the need to get things done. And um for me, I I quickly realized I was the been there, done that guy for customer service and fulfillment. So that was easy to just go. And then I've been more in the question-asking mode on the on the other areas. So trying to kind of adjust that way. Yeah.
45:37 Speaker 9 If you look 10 years out, when it comes to supply chain and ops, what do you think is gonna be the biggest change that'll hit your functional area?
45:48 Speaker 3 Yeah, uh, biggest change. That's stuff. I mean, I think everything is changing right now. That's what's so exciting about supply chain and operations right now is that in this kind of on-demand economy, it's a complete blank canvas for new operations and creative operation solutions. Um I mean, if you think about a lot of the operation solutions, a lot of the methodology was um built back in, you know, the 80s and 90s of the mass production um retail era. And we've been kind of working off of that. And so I think that the changes that are happening right now are amazing, you know, from peer-to-peer um logistics to personal shopping to same-day delivery to drones, you know, there's just there's a lot of things that are changing. So I don't know that I can say what I think it's going to be.
46:36 Speaker 3 I just think that what's happening is that from an operations perspective, rather than seeing operations kind of behind the scenes doing the optimization, we're starting to see operations come out and really drive revenue and be entire business models. And um I think it's gearing up to be a major disruption, like kind of the next major disruption, I think, in the e-commerce space, that I don't know that I can nail it down to what the biggest thing is gonna be. But I think I think there's going to be a lot of them. I think that's why. Maybe somebody else has a hasn't nailed it.
47:11 Speaker 1 I think if I knew I wouldn't be sitting here, you'd be building it. I'd be getting funding for the next big thing. I I do think what's what a trend that you see happening is that uh smaller startups are starting to are enabled or being enabled to provide some of the same levels of service that larger, more established companies have only been privy to in the past. So, you know, the everything from the technologies and material handling systems, having um, you know, cloud-based uh uh telephone systems, you know, used to have to have a big uh box in your warehouse, now you can just subscribe to a service. And I think all those kinds of disruptive technologies are making it easier to provide first class service without having to have really deep pockets.
47:59 Speaker 1 So I think that's a trend, which I think is kind of exciting. It's uh you know, I think uh the Amazons of the world are gonna have a lot of competition from smaller startups on the service front.
48:12 Speaker 2 I think in physical goods, um lead times have to get shorter. So in jewelry and in beauty products, we're just dealing with um this old school industry that operates on the model of you know, it's gonna take six months. And um the the retailers are all you know, just just trying to uh act as fast as possible. That just doesn't work. Like we're we're trying to disrupt our space and and we're gonna be pushing our suppliers to figure out how to get faster and faster. And I imagine others are too. And I just feel like you know the the the world of a you know 18-month product launch cycle just can't maintain itself. Something's gotta break. I think it's gonna um force companies in through that entire supply chain to get more flexible, to get more nimble, and then they'll you know they'll be able to win.
49:00 Speaker 4 I mean it's interesting. All the the three companies up here are all personalized. Um you know, they they're sending out a personalized product, shutterflies in the house, uh sending out a personalized product. Uh we're seeing a trend towards not you know, to not towards mass customization, but to ends of one. You know, the more and more business models are being pitched to us, our you know, personalized service to every consumer. And so that has massive operational implications. Shout it out.
49:25 Speaker 6 Well, legal issues are the biggest pain points right now currently. What legal issues are merging, then you know you'll have to deal with maybe the only lawyer.
49:38 Speaker 2 Um patent trolls. If you if you leave, we'll address the uh patent trolls are uh certain a pain point that blue Nile we dealt with, you know, IP, you know um, you know litigation threats and Julia, you know, we got our first letter, so that's top of mind. Because it's current, but that just seems like a you know an ongoing problem.
50:08 Speaker 6 A lot of the people.
50:10 Speaker 4 Can
50:10 Speaker 6 we finish that?
50:11 Speaker 4 Sorry, a lot of the people marketplace things we work with are running into big labor law. You know, uh thing you'll give it a contractor as an employee, what you know, or competing with the union labor force. So that's seeing a lot more of that. Yeah.
50:25 Speaker 6 Uh a little bit on more on the infrastructure side. What systems are you using that maybe other companies don't know about? You know, something beyond your ERP or what are you doing to enable your operations pieces that you think we ought to be looking at?
50:41 Speaker 0 I I don't have we're building it all ourselves, and so uh we use Zendesk for customer service, uh, but I'd say the the tie-ins to Zendesk are very proprietary in terms of like how tickets um how we think about tickets related to customer segmentation or how we think about tickets related to holding business owners more accountable to better operations. But uh it's not that we're so uh, you know, it has to be us that builds it. It's just that when we have researched outside parties to talk to us about how they can fit into our model, our model is funky enough and different enough that it doesn't fit. And so we've just sort of said we'll listen, but likely have to build our own stuff to run the business the way we want to run
51:28 Speaker 1 it.
51:30 Speaker 0 And that's really uh empowering to engineers. I don't like one of the challenges I think we've had in our businesses getting in front of engineers uh because it's a women's fashion company, and that's just not where we're getting as much kind of um in the front end of the funnel, people coming through. Um but once we show engineers what we're doing, there's a lot of excitement because of the systems that we're building and also just the data and analytics and how engineers can work with the data science team in a very proprietary and cool way. Um it it's for us a competitive advantage once we get in front of an engineer building our own stuff.
52:10 Speaker 4 I'm seeing a lot of that in my companies. There's yeah, well a lot more is built than bought in just because the models are different, you know. Zulilli and Fab built almost everything.
52:19 Speaker 1 It's a big commitment though, once you decide to go that route. Now you're you know now you're saddled with maintaining it and um that's that's a a global issue. Um, finding platforms that handle most of your needs and resisting, either resisting or going both feet in on uh owning the code is uh is a big decision.
52:42 Speaker 4 One more question.
52:44 Speaker 10 Can you talk a little bit about sort of in-season demand planning and sort of who should own that, where you've seen it owned successfully, and how that sort of how what the interactions are?
52:57 Speaker 2 Uh both Blue Nile and Julep, the demand plan is is with uh the merchandising team. The the they're they're trying to drive the revenue with the product assortment and the promotions and that makes sense to me. And then we couple very tightly with that though on operation side, and and then we'll take that and do our own translation into orders, units, calls, and that sort of thing. So you know they're closer to what we're gonna sell, and then we take it from there.
53:29 Speaker 1 Well, and it's also being responsive. I mean, because the plan is just the plan. And um you know, that's that's the um that's a long and probably a never-ending battle between the people who do the demand planning and the operators, right? So what you describe is very common. You you get a sales plan that's expressed in dollars, and then somebody tries to convert that to orders and then they may extrapolate that to units, and they may take a monthly sale plan and then try to water it down to a weekly or a daily or sometimes by shift plan, it's always wrong. By definition, it's always wrong. So the biggest challenge from an operation standpoint is how do you build enough flexibility into your capacities to handle um not just the plan but both sides of that plan.
54:19 Speaker 1 You know, how do you respond downward so you don't throw a lot of money away? And how do you flex upwards so you don't blow customer service during those periods? Yeah,
54:30 Speaker 2 that's gotta be part of it. So uh what we found is that um when we've had like for a while we had we asked the the you know planning team on the merchandising side to generate the units and orders, and uh what we found is they they have less sort of skin in the game in in terms of that accuracy, and and so we we wanted to own that because we could make the better you know, better estimates and also understand the the kind of margin of error better so that we could we could understand what variability to build in.
55:01 Speaker 1 Yeah, and and depending on the business you're in, you really do have to build a church for Easter Sunday. Because you have to be able to handle the worst case scenarios or or blow sales that you'll never get back, possibly lose customers that you'll never recover. So I think from an operation standpoint, staying ahead of that power curve was always a big challenge.
55:22 Speaker 0 A couple of things that's a demand planning uh at Stitchfix reports uh into merchandising. Uh but we do it a little bit differently. Um it feels right to me, but it's because I went from church for Easter Sunday and really big peaky season, like uh holiday season, to at SitchFix. Uh we try to have quality be the foundation and the three constraints being quality of inventory, how many stylists can pick the five things after the algorithm runs and ops capacity, and that drives what we actually do. So essentially it's a little bit of a tail wagging the dog where we say buy day by size and kind of a really cool algorithm. This is how much we can do and still keep quality across all of those dimensions, and the customers come into that. Now, I just left a meeting where you know marketing is saying add more capacity, we got more demand, why can't you do more?
56:14 Speaker 0 And we have really tough conversations about sort of demand planning as it relates to keeping quality, because what's great about from an operator's perspective is I tell our two warehouses many weeks out, this is how many fixes you're doing. And they can operate that very well. And that's when when you run it like this, that's where I can see where we have process issues or cost issues. It's harder from my perspective to see how to get a really good process when one day you're shipping six thousand, the next day you ship twelve thousand, the next day you ship two thousand. I can see it when I get to ship six thousand three days in a row, because then I can see where we need to improve. So we manage that.
56:57 Speaker 4 Your models tends towards linear versus
57:00 Speaker 0 it has. We don't have a C yeah, we that we are benefit we benefit a lot by having a non pretty much a non-seasonal business. So
57:06 Speaker 4 like shut the shutter flies non-seasonal, right? Yeah, yeah.
57:10 Speaker 0 Yeah. I'm used to peaky too, and it's hard. I think it's hard to manage peak and be held accountable to costs and quality, because like Mike said, you have to build the church
57:21 Speaker 1 free. You either have to invest in technology or have flexible workforce.
57:27 Speaker 4 Or both. Love to thank our panelists for a very uh stimulating conversation. Thanks, everybody. Thank you.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated August 25, 2026 · transcript and summary are machine-generated