Rent the Runway is a business that’s dressed as a fashion company but has the bones of a tech startup. Their service allows customers to rent clothes for special occasions or everyday wear. Rent the Runway aspires to give customers the freedom to create a ‘closet in the cloud’ — receiving a rotating wardrobe of outfits that suit their style.  

Data drives all of the company’s operations — from personalizing recommendations, to efficiently dry cleaning and shipping clothes to customers. On this episode, Rent the Runway’s Chief Analytics Officer Vijay Subramanian shares:

  • Why Rent the Runway made its first C-Suite hire in analytics.
  • How to stop doing laundry by keeping your ‘closet in the cloud’
  • How Vijay has baked data into the DNA of Rent the Runway.
  • What you can do to take your company to the next level of being data-driven.

Want to make your own company more data-driven? Learn how Indicative can help.

rent clothes

Rent the Runway Chief Analytics Officer Vijay Subramanian

Full Transcript

Lauren Feiner: This data-driven conversation is brought to you by Indicative, the leading behavioral analytics platform that empowers marketing and product teams to understand their customers’ full journeys. Start deciding by data with 1 billion free user actions per month, and unlimited user seats.

For a limited time, Deciding by Data listeners will get early access to the Indicative platform. To sign up, go to indicative.com/DecidingbyData

~

Lauren Feiner: You’re listening to Deciding by Data, the podcast that brings you into the C-Suite to learn how data drives successful businesses.

Today on the show, we’ll give you a look inside Rent the Runway — a business that’s dressed as a fashion company, but has the bones of a tech startup.

Rent the Runway makes designer styles accessible to a wide array of customers. It does this by renting out clothing and accessories from its enormous inventory. Customers can choose from different subscription levels based on their needs.

While some choose to rent a single dress for a special occasion, others opt for the unlimited subscription, which lets them rotate their wardrobe every time they send back a rented item.

We talked to the guy who laid the foundation for Rent the Runway’s massive logistics operation — Chief Analytics Officer Vijay Subramanian.

What’s really unusual about Vijay’s role is that when he was hired in 2010, it was the first C-Suite hire the founders made in the company’s history. That’s how serious the founders were about making data a central pillar in the company’s culture.

We asked Vijay how data influences every part of Rent the Runway’s operations — from cleaning clothes to making the perfect recommendations for customers. 

This is your host, Jeremy Levy.

Jeremy Levy: If you’ve heard of Rent the Runway, you probably think of it as a fashion company. But what you probably don’t realize, is that the company is grounded in tech. Vijay explained how Rent the Runway’s infrastructure has rapidly advanced to keep up with growing demand.

Vijay Subramanian: … when we started, we were outsourcing the cleaning operation, right? So, we had a website. Consumers could come, and they could try inventory, they would drop it back. But we would send it to a cleaner down the street and we would go to that cleaner and have to sort through all the inventory that’s ours, make sure it’s actually really good condition, and then get it back to us, and then ship it out.

You know, when I joined and then sort of six months in, nine months in, we looked at the actual behavior of how inventory was moving in and out and the timing of all of this stuff, there was two things that sort of dawned us, right? Which is, the rental economics is predicated on, obviously, how frequently you can turn it because the more the inventory is sitting with us it’s not good. And B), how long can you keep this going effectively. And those are the two important variables.

And it became sort of- Now again, it’s obvious but it was not all that obvious then, is that the way to make it happen is to basically control the whole process. So, if you buy the inventory, you control how you take care of it, you control the timing of it. Not only can we make the back-end economics work because we can turn the inventory as fast as we can. We also can fulfill the customer promise better.

Jeremy Levy: That promise — to give women the flexibility to rent their clothes, rather than own them — was a promise that had never been made before.

The idea of renting DVDs wasn’t new when Netflix started, but, as Vijay pointed out, renting clothes meant gambling on an entirely new model.

Vijay Subramanian: The way I think about it is that the rental economics, and how you actually build a business, how do you acquire customers, how do you make money, none of this was known. You can argue even Netflix, when they built it, they were kind of leveraging off of a history of people renting movies. Renting movies was not a new thing when Netflix launched.

Jeremy Levy: This entire model is new for fashion, yeah.

Vijay Subramanian: Exactly. So, Blockbuster existed. So like, we had companies before that where people were used to the idea of renting versus owning. So, we were creating a completely new category where the customer itself is changing their behavior while we’re figuring out how to make the back-end logistics and the equations actually work. Like how do you actually make money in this model?

And you’re dealing with highly capital-intensive inventory, right? So, we have inventory that is not a DVD. It’s actually a beautiful gown. It’s not as cheap as a DVD is. It needs a lot more care. When the inventory gets back to us, before we give you somebody else, we have to make sure it actually, it’s like new. So, there is a lot of elements to it that … kind of implied that we need to A) be incredibly good at building a brand so consumers would try it and be proud of it. And B) we had to be really good at making the economics and the back-end infrastructure work. And that’s where tech and data really came into play.

“We were creating a completely new category where the customer itself is changing their behavior while we’re figuring out how to make the back-end logistics and the equations actually work.”

Jeremy Levy: That’s why Rent the Runway’s co-founders, Jennifer Fleiss and Jennifer Hyman, decided to bring Vijay on as their first C-Suite hire. At the time, Vijay was working for the retail analytics startup ProfitLogic, which was later acquired by Oracle.

ProfitLogic’s co-founder Scott Friend made the introduction to Rent the Runway’s founders.

Rent the Runway’s decision to make their first executive hire in analytics was a deliberate one, and spoke to the core values of the company — to make data-driven decisions.  

Vijay Subramanian: … I mean, if you look back, it seems like an obvious decision to make but it was definitely not obvious. So, it really took, I think both Scott and Jen, who made the hire- who actually made the introduction to me and then made the hire, it was a very bold move on their part.

I think they saw something that, this company is doing something completely new, first of all. Like rental didn’t exist in any kind of scale. So we had to figure out a lot of stuff.

Jeremy Levy: Vijay’s introduction to data science was not an obvious decision either. Before ProfitLogic, Vijay had landed a prestigious role at Proctor and Gamble.

After a year, he couldn’t shake the feeling that this was not what he had been searching for.

Vijay Subramanian: I knew I wanted to get into data and computing, and I was building these models for Proctor and Gamble’s production facilities. I mean they have all these facilities around the world making shampoo and dog food, and all of this stuff. It’s pretty exciting stuff, actually. And there my job was to model those plans and see if we can drive better throughput and optimize the flow of the inventory through the facility. And there, I mean, I learned a lot, but I felt like a very small, sort of isolated cell in this giant organism and I didn’t feel like I had the impact I wanted to have, I felt restless, I mean you can call it ambition — like I felt like I could do a lot more. And that’s when I actually, for the first time, dug into this world of startups. There actually is a concept where you can go work in a company that is smaller and is like a small group of people who are really motivated to work together and build something from scratch.

Which was, by the way, an alien concept because I grew up in India in the 80’s, and there if you’re growing up in India, if you’re growing up lower to lower middle class. The middle class in India, the pinnacle of success is to get a stable job at a large company.

Jeremy Levy: Right, sure.

Vijay Subramanian: That was, your- you made it, right? So, when I came to the U.S. for grad school and I got my master’s and my Ph.D, and I took a job at Procter and Gamble, my parents thought I had made it. I have like finally arrived, and now I have this job with this incredible company- which is all good is all true and all good- except I still felt very restless and anxious about whether I’m actually fulfilling my potential.

Jeremy Levy: Did you study data or math back in college and university? When did, sort of, data become the focal point of your career?

Vijay Subramanian: That is a very interesting question actually. So my undergrad is in chemical engineering which is nothing really to do with data per se. But when I came to the grad school I started to think about what kind of work do I want to do …  within chemical engineering.

There were a handful of schools, it’s very small by the way, Carnegie Mellon, Purdue… schools where they had programs where they were really focused on math programming, statistics, computing to chemical engineering problems, they were applying these methods, to chemical engineering problems like manufacturing facilities and that kind of stuff. And the rest of chemical engineering was like this broad swath of experimental work, like how fluids flow, how solids reacts to like chemical reactions and like it’s pretty heavy duty experimental stuff.

And when I saw the 95 percent and the five percent, I wanted to do the five percent. I really wanted to get into computing and I can’t explain why, necessarily, but I wanted to be able to model things and understand things through the lens of like computers and numbers much more so than sitting in a lab and running experiments.

Jeremy Levy: Even as data scientist roles have expanded, the role of Chief Analytics Officer is still pretty new. I wanted to get a sense of how Vijay has carved out a space for himself in this role, and understand the significance of the position at a company like Rent the Runway.

“I think you need, at this point, if you are at scale, if you have a CTO, if you have a CFO, if you have a CPO, head of product, or a COO of operations, you better have a head of data analytics.”

Vijay Subramanian: I don’t know if I’m the average sort of representation of a chief data officer or a chief analytics officer, right?

Jeremy Levy: That’s my point, I don’t think there is a template.

Vijay Subramanian: I say it only because I think I’ve been there since the very beginning and you have to wear so many different hats in order to build a company. I mean, you can’t just be focused on the data alone. You have to think about how the data marries with the software you’re building. So, I played a very big role in building a lot of the software that drove the logistics.

So, when inventory shows up from a customer, we scan it, we need a triaging system to figure out which units do you push through in what sequence? Like a ER if you will, a system for the, like in the hospital, right? Because certain units more demand, they have to go faster, some need more cleaning time. So, you have to build an algorithm just to like sequence the inventory through the facility, right? You cannot do it just through thinking about data, you have to build software and you have to build a product.

… To answer your question, though, because I’m not trying to get out of that question which is, I think when you get to scale, there are functional domains that need functional leaders who run that domain. There is a reason why we have CTOs, right? If you wind the clock back ten, maybe 20 years or so, you will find companies where engineers are working under finance. I know it sounds crazy. Engineers will work under marketing and their job was to build software products for marketing, for something else.

But at some point, we realized that that function is a unique function. It has tremendous leverage and value, and there’s benefits for the function to be centralized under a leader who can drive not just the day-to-day of the function, but the strategy of the function and have a voice in the C-suite. Right? So, that’s where the CTO role came into play. Right? Even the CFO role, by the way, is not like that ancient. It’s actually a fairly recent evolution over time.

So, when you think about these, I think of the data in the same lens, which is… there should be some function that is centralized that is together, that builds a craft around it, that’s thinking about what to collect, how to collect it…

And then, most importantly, run experiments. Have a culture of experimentation. That only trumps all of the stuff. You have to be able to say, “Okay, here’s the data I’ve look at. Here’s my hypothesis. Okay. What do I need to know in order to prove the hypothesis?” Go run the experiment, look at the data, prove this hypothesis, rinse and repeat. It’s the good old scientific method that’s led to the glory of all of our civilization. So like, let’s use that process in building companies. And to do that, I think you need, at this point, if you are at scale, if you have a CTO, if you have a CFO, if you have a CPO, head of product, or a COO of operations, you better have a head of data analytics.

“It’s the good old scientific method that’s led to the glory of all of our civilization. So like, let’s use that process in building companies.”

Jeremy Levy: Vijay talked me through all of the tedious and logistically challenging steps that make up a garment’s journey. His team analyzes each and every one of these steps to ensure that their inventory smoothly arrives at its destination. Vijay explained how the logistics engine got started.

Vijay Subramanian: A lot of our capital and mind-space was invested in- on the logistics. Now, why was that important? Because that ultimately led to the equation by which you know that you can buy a unit of inventory and you can get enviable unit of economics out of it, right? So ultimately, if your unit economics is incredibly profitable on a unit of inventory, then everything else is just about scaling the business. And many startups actually start up with a model where the unit economics is not profitable and they hope that scale gives you that. And that’s actually a very important distinction. We knew that we had to get the unit of economics to be profitable…

So, to do that, we needed to own the logistics, we needed on own the cleaning, we needed to own the sourcing, we need to own the control of inventory, the timing and the technology of all of this stuff. … [E]very unit of inventory that you as a customer may consume, when you get it back, it goes through giant fulfillment center.

Honestly, it’s kind of a wrong thing to call ours a fulfillment center. It’s actually a manufacturing facility. Because the units actually come back. These are like tens and thousands of units that come back every day, that we have to go through a process where we have to sort it. We have to figure out what needs to be processed how, because different fabrics have different needs, right? So, the eight years of knowledge really where data’s played a very critical role is figuring out how to take these units of inventory with different fabrications and different uses and figure out the optimal way to take care of them. That makes sense. So, by doing that, we’re able to prolong how long again actually it can last and then we’re able to get more value out of the unit of inventory.

Jeremy Levy: So, what do you mean by take care of them? Does that mean using some level of data in terms of the maintenance of those dresses? The ideal chemicals for stain removal? Or is that figuring out the optimal time that address can be used in terms of how many times can it be rented? Help me understand sort of what you mean by that.

Vijay Subramanian: It’s the full life-cycle. It start with even how you buy upfront. Right? You figure out overtime what kind of units of inventory are better for long-term use, basically. And you can over time figure out what kind of fabrics are more amenable to cleaning and processing over time… [F]or example, we have 60 odd fabrication codes for a unit of inventory and we can look at the data and say what- and depending on how it went through each turn, what happened to it, controlling for those factors, which fabrics had what kind of longevity and what kind of care?

And then it comes down to what care are you actually doing? We can test different types of cleaning methods. Right? So how do you how do you clean this? In what kind of machine with what kind of chemicals? And then you can test when something happens to it, how do you take, how do you like repair it? … [I]t goes all the way through the whole life-cycle till the end of it. Now, to do that you, you can’t do that if you don’t control the whole technology and the process in the data collection around it and that’s really what we did.

“So, the eight years of knowledge really where data’s played a very critical role is figuring out how to take these units of inventory with different fabrications and different uses and figure out the optimal way to take care of them… So, by doing that, we’re able to prolong how long again actually it can last and then we’re able to get more value out of the unit of inventory.”

Jeremy Levy: When it comes to the data that Rent the Runway collects and the data that traditional retail stores own, he drew a sharp distinction.

Vijay Subramanian: The two elements that I think are really interesting for us, that very few companies actually work on and invest in, is let’s go one way in this direction which is the signal for what is true demand. So, we know what you’re getting, but what you’re getting is often a function of what is also available at that point in time. How do you really figure out what do you truly want? So, we track a lot of that data from a user perspective to understand what are you searching for, what are you filtering on?

And then, we have this feature where you can like something, and you may like something even when you don’t even have open slots, right? So, I have four slots at home, it’s fully booked. I don’t really need to pick anything, but I could still like something. So that when next time when you come to pick, you actually may want to select that item to get it to your home, right? So, those signals are orthogonal, or independent of true availability at that point in time. What do you like and what are you filtering on, and all stuff. So, we use that to understand how miscued are we, by looking at customer segments, in terms of what people want and what the true demand is?

So, that really affects and corrects our buying process upfront, and that data is a very rich data set because we’re doing it at user level, and we’re then asking the question, for every user, what’s actually- what does she like in sort of the abstract sense, in terms of the ideal sense. What do we have in our catalog? How much in our catalog is relevant to her? How much is available at a point in time? There’s a whole funnel, right? Then, we then ask, how do we correct it at every user, and then we aggregate it to go do the buy.

Jeremy Levy: So if I understand that, you essentially are measuring what people are interested in from a fashion, maybe size, perspective, and then that’s influencing the buying decisions and also the recommendation decisions independent of what’s currently in the inventory?

Vijay Subramanian: Exactly. Of course, those are somewhat correlated but able to control for that factor, and we control for the factor by getting signals of what she likes that are independent of availability.

Jeremy Levy: On an individual basis.

Vijay Subramanian: Exactly. So, you know what you get. We have our signals on what do you want, and we try and adjust what we buy to get you what you want. And then, there’s another thing which is even more interesting and more rare in commerce. Which is, once we ship the item to you, we get a feedback on every single unit that you actually got… How many people- how many commerce companies have that data, right?

… The reality is, unlike other companies, we get feedback from you on every single unit, which is a very rich tapestry of data. Did you wear it or not, right, first of all? If you did wear it, did you like it, did you love it, or it was just okay? If it was just okay, why was it just okay?

Now, did you not wear it? If you did not wear it, why? Was it a fit issue? Was it because it wasn’t flattering? If it was fit issue, then tell me where the fit issue was. Was it the chest, was it the waist, was it the height? If you can think about that, data points that we are collecting on every single item, and you actually cannot go pick your next item until you give feedback on the previous items that you actually selected…  

So, we do that and then we get this rich data point on what actually is being, not only consumed, but what actually people genuinely wearing and not wearing, and for what reasons. And that actually also feeds back into the buying process, right? Because we can start looking for signals on, okay, this thing is popular but it has certain issues with fit with certain types of people. So, maybe we should do something, we should get something else that is like this item that probably has a better fit profile. So, we can do the same calculation that we did for true demand.

So, true demand is like what do you like in terms of style and what is your work aesthetic? Like when you go to work, how do you want to dress up for work, essentially. And then, there is the actual consumption data and what do you wear, and why, and did you like it or not. So, both of these go into this machinery to figure out what you’re buying in the first place.

Lauren Feiner: We’re going to take a short break, but when we return, Vijay will explain Rent the Runway’s vision of a ‘Closet in the Cloud.’ Stay tuned.

Midroll

Lauren Feiner: Welcome back to Deciding by Data. We’re here with Vijay Subramanian, Chief Analytics Officer of Rent the Runway. We left off with Vijay explaining one of the key challenges his team is working on: understanding fit. Fit is just one of the considerations Rent the Runway takes into account when recommending clothes to its customers. Vijay outlined the future vision for their recommendation engine:

Vijay Subramanian: Look, ideally, what’s the holy grail of recommendations? You show up on the homepage, I literally could show you the next two items that you want to get. Let’s say I’m returning two items and I want two more items. You literally want to get to the next two items and you literally show that and I’m like, “That’s what I want, press the button.” I get it.

Jeremy Levy: No need to browse.

Vijay Subramanian: That’s the holy grail. So, I was giving you a metric of like, if I show you 10 items. You are on the app, right? So, you open the app and you scroll through it. If I can get to the two items out of the 10 that I show you, that would be fantastic. Are we there today? No. Because today people are looking at it, they may select one and then they may browse, they may apply other filters and they may still get to it, right?

So, but that, as our catalog expands and as we get more and more data, right? Because this program is about a couple of years old.  So, we have now millions of data points on the feedback right now. But as we get better and better at figuring out the customer segments more and more granular, and then we start to figure it out to fit and all of those elements of it, I do think you’re going to see higher and higher selection rates from the very first 10 items.

… And not only is it a selection thing, we want you to wear all of the items, actually, and have a feedback that actually says you loved it. Right? Because that’s ultimately is the holy grail. It’s not just saying, “I can show you things that you like.” That is somewhat of a solved problem, right? I mean, we’ve done that with music, we’ve done that with movies. So, we will apply similar techniques to figure out what styles you like, like what things you like in terms of your style DNA, if you will. But the thing that is definitely not solved in fashion is being able to maximize your wear rate of that item and being able to make sure that actually fits you. Because fit is an incredibly hard problem.

Jeremy Levy: I wanted to take a step back for a second to something you mentioned a moment ago which is sort of the holy grail of that recommendation engine …Correct me if I’m wrong, the holy grail would be really that I have almost nothing in my closet or I have like the basic essentials in my closet. That my entire inventory or my entire inventory as an individual, my closet, is essentially something I can rent on an ongoing basis. Is that the way you think about that recommendation algorithm? Is that go far that to say, “I don’t even need to browse 10 items to get to, you know, to find those two.” How far does that go?

Vijay Subramanian: … I think the holy grail for our company, for the market is that your closet is predominantly stuff that’s on rotation. Look, you will still buy your basics. You’re going to buy your jeans, you could still buy your T-shirts and stuff that you need to live your daily basic life. But I think for anything else, for any kind of dress you’re wearing to work, for any outfit you are wearing for an event, for a wedding, for whatever. Our hope is that for all consumers in the U.S., or in the world eventually, you are rotating it through a closet in the cloud. That really is the vision for us.

“But the thing that is definitely not solved in fashion is being able to maximize your wear rate of that item and being able to make sure that actually fits you. Because fit is an incredibly hard problem.”

Jeremy Levy: Hm, I like that, closet in the cloud.

Vijay Subramanian: So, we do think of it very much of it as just inventory that’s out there, and just shows up magically and you wear it, and it goes back, and it’s perfectly aligned with your life.

Jeremy Levy: Yeah.

Vijay Subramanian: Right? During the workday, during the workweek, it shows up for stuff that you want to wear for work, maybe there’s one dress that you want to wear for going out at night with your friends, and then, when you return that stuff, and then for the weekend, you are going to a wedding, so it shows up on a Friday for the weekend, a perfectly beautiful wedding dress that you can wear. So, that’s kind of what really the macro vision is…  

Jeremy Levy: … [G]etting back to this notion of ‘closet in the cloud,’ it’s fascinating, right? The idea that some system can take not just my fit but my calendar, where I’m going, what I’m doing, integrate that with my fashion sensibilities and essentially pick for me what I’m going to wear on any given day. How far away are we from something like that?

Vijay Subramanian: We’re seeing signals already, where we see in certain markets and certain companies, even, where one person tries it and she becomes obsessed with it and you monitor her behavior or a while and you find out that she’s spending less and less money owning things. Her closet has basically been transformed. So, what we’re seeing now in the last couple of years is, we’re seeing these pockets and we’re seeing these network effects emerge in these markets and these companies. That said, I do think we’re probably like, in the next decade, we’re going to be much more of a mass market product. I mean, we’re growing, we’re very happy with our growth rates. I mean, we’re crushing it in terms of how this is growing, but we’re not Amazon scale, right? So, I mean, our hope is that we can be sort of a scale where mass market consumers around the world will be able to do this.

“Our hope is that for all consumers in the U.S., or in the world eventually, you are rotating it through a closet in the cloud. That really is the vision for us.”

Jeremy Levy: I love this idea, though, that I don’t have to care about what I wear anymore. It makes me think about the way Steve Jobs had his turtleneck uniform, right? And Mark Zuckerberg always wears, you know, a hoodie. And they describe it as well, look, I have a certain amount of brain cycles that I can allocate to things in the day and choosing what I’m going to wear is not at the top of their list. I wonder what are the other implications of this in someone’s life if I no longer have to think about what I have to wear? Would people be that much more productive every day?

Vijay Subramanian: For sure, yeah. Well, except they’re not wearing the same thing in this model. You’re also getting the variety of it and you’re getting the diversity of it without you having to do the heavy mental lift you actually need to get.

Jeremy Levy: And the fact that it sounds like in this model, I would never have to do laundry ever again, which would be awesome also. [laughs]

Vijay Subramanian: So, right now, the price point for the product for the one that’s unlimited is $159 a month. And you look at that and, for sure, I would definitely not claim that every single consumer in the US can afford that product on a monthly basis. But you look at what people spend on clothing today and look at what they spend on cleaning it and maintaining it, you’d be surprised how much it adds up. I think we’re at the beginning of a revolution. And it’s really a question of scale and how fast we get to scale.

Jeremy Levy: The influence of data extends far beyond Vijay’s desk. The company is structured so that data is an integral part of every job function.

Vijay Subramanian: So, you have these functional teams: you have technology, you have product, you have data, you have marketing, you have operations, you have finance, let’s say, right? I’m just being like simplistic here. We have those functional leadership and they all report up to somebody who’s got a voice in the C-suite. But the way we operate is cross-functional. We actually have a matrixed organization where we carve out these initiatives. This team is going to work on growth. Go acquire new users. This team works on, once users are in the program, work on retention, work on making them happy, work on improving satisfaction, basically. These units work on operational excellence. Build the software to make their operations more efficient, better throughput, higher quality. This team works on inventory sourcing, better inventory segmentation, right?

All of these teams are all cross-functional. They have engineers and data and products. They’re all embedded together. Data is like woven into the DNA of all of these teams.

“Data is like woven into the DNA of all of these teams.”

Jeremy Levy: Many companies like to think they are data-driven, but Vijay sees many levels along the spectrum. We closed out our conversation by outlining what these different levels look like.

First, there’s what Vijay calls “data-informed companies.”

Vijay Subramanian: … which I think most companies are, hopefully, by now, where they do look at data. They make decisions, but they’re largely reactive… There’s this data floating, everyone looks at it. They think about it, and then they move on. Right? A lot of these are very reactive conversations.

Then you have the data-driven, the term that you used, the data-driven companies. But it’s much more proactive, where you have healthy debates about not just like- the numbers are table stakes, everyone knows we have numbers. Right? We are data rich. But we’re discussing a lot about what it means. What’s the interpretation of the data? Was the data collected properly? Is there a bias in the way you’re looking at the numbers? And then, you’re thinking about experiments. Is this the most important thing. You’re thinking about, “Okay, I know this much. What do I need to know in order to make this or do that?” So, you’re running tests all the time. With customers internally, you’re running true scientific tests, like AB tests, and you’re saying, “Okay, are you done experiment and see what happens.” And then, you’re drawing conclusions again. So, it’s a constantly- it’s an iterative process where you’re figuring out the business together in this cross-functional team, and the data team is the one who’s helping build that culture. But it’s not just one person. If the entire team doesn’t think that way, it’s going to fail. So, the job of the data person is to imbue that culture in the way of thinking across product, across the engineers, across a marketer so that they’re all thinking the same language and same lengths.

“So, the job of the data person is to imbue that culture in the way of thinking across prduct, across the engineers, across a marketer so that they’re all thinking the same language and same lengths.”

And then, the pinnacle to me, which is a special case of data-driven company, is you’re just data native. Like you’re born that way. It’s woven into your DNA, it’s how you’re organized. No one even thinks of data as a secondary thing. It’s just part of how you operate. And when you do that, something special happens, which is again you’re building the company by applying the scientific method. It’s just like second nature to you.

Like any time you talk about things that that’s happening, you always ask about what do I not know? How do I collect that data? These are all native phrases you are using day in and day out. What experiment do I need to run to prove or disprove a hypothesis? When you hear people are talking like that, that is a data native company, right? And that’s really the holy grail, I think, of like a culture that it’s just part of how you think and how you operate.

“And then, the pinnacle to me, which is a special case of data-driven company, is you’re just data native. Like you’re born that way. It’s woven into your DNA, it’s how you’re organized. No one even thinks of data as a secondary thing.”

Comments are closed.

Deciding by Data © 2018