Don’t let the laid-back Pacific Northwest demeanor fool you. Luke Friang is a self-described “speed freak.” Whether bombing down icy slopes on his snowboard or maneuvering through hairpin turns in a mini-racecar, Friang simply loves to move fast.
And that’s why the 42-year-old college drop-out and self-taught technologist has found such a good home at Zulily.
No company (at least that we know of) is moving quite as fast as Zulily. Already on its fourth headquarters and fresh off of a massive $43 million venture round last summer, the daily deal site for baby and kids’ products now employs more than 300 people in Seattle alone (not counting staff at its two distribution centers in Reno, Nevada and Columbus, Ohio). Its daily email blast goes out to more than five million subscribers each day, with thousands of products sold through the site and mobile apps each year.
As chief information officer, Friang oversees an expanding technology staff that’s encouraged to push the envelope and solve complex technical problems, some of which are only associated with a fast-growing e-retailer expanding at hyperspeed. It’s certainly not an easy chore. And Friang — the former CIO at drugstore.com and senior director of e-commerce technologies at Costco — admits that there have been some speed bumps along the way.
“Literally, we had a situation where we were building things and replacing them in a matter of months,” said Friang in a recent interview at the company’s headquarters in Seattle’s Sodo neighborhood. “We have touched everything, multiple times, and I think that’s the challenge from the technology perspective, and from the company perspective, is how do you keep up with ‘Zulily Time’ and the speed at which we are moving.”
Since he joined the company 16 months ago, Friang said that they’ve implemented an entirely new Web store platform and order management and fulfillment system. Following an approach taken by Amazon.com, they’ve chosen to build versus buy in most cases.
That’s often more challenging, but Friang said it is an important distinction that allows Zulily to be more in tune with its customers. Not everything has gone as planned. Shortly after his arrival, Friang said that they attempted to implement a dramatic upgrade in the open source platform that was being used to power the Zulily site.
They had clearly outgrown the system, and so Friang led efforts to do what he dubbed a “big upgrade.”
“We put a lot of effort in to it, got it ready, released it, and it didn’t go well,” said Friang, adding that the site was frequently crashing or moving slow under the weight of users. Some of the problems arose from inadequate testing of the new software, but Friang learned more valuable lessons. Most importantly, they changed the way software was developed and deployed.
“In this world of move ultra quick, I think the big lesson is stay away from platform upgrades, meaning iterate, and break it apart,” he said. “We basically started rolling it ourselves, still using the open source toolkit, but it allowed us to … break it apart, roll bits and pieces out, and kind of control the positive and/or negative that might come out of a release versus the first time when we said: ‘Hey, let’s just get it out there, and we think it is going to work.’ When you are changing out the engine when you are going 150 miles per hour, maybe you should just change the carburetor or change the fan belt? But, with packaged software, that gets challenging.”
First-time visitors to the Zulily headquarters will quickly realize that this is not your typical startup. After all, how many startup companies do you know that have a kid’s playroom and waiting area stocked with toys and books. On my recent visit, baby and pre-teen models passed by with their moms, arriving for appointments to be photographed in the company’s expansive photo studio.
“Nothing traditional works here,” admits Friang, who dropped out of college to join a startup company before landing a gig at Eddie Bauer. At Zulily, Friang said that he’s trying not to put a “gate” in front of developers, creating a culture of experimentation and, of course, one which thrives on moving fast.
“You are never going for perfection,” he says. “It is really about doing what you think is right, getting it out there, and then you know it is going to change, so we are going to learn, we are going to evolve and so we really emphasize the continuous development life cycle.”
Here’s more from our interview with Friang in which he discusses the company’s approach to mobile development; how they are analyzing customer data and what he’s learned over the years watching Amazon.com.
On whether Zulily is a technology company or a retailer: “Definitely, we are a retailer. We are an online, e-commerce retailer. I think what is a little unique is the technology challenges that we face on a day-in, day-out basis because of the type of online business we are operating, and this notion of the daily deals. Whereas if you take an example of myself, drugstore.com, Costco.com, EddieBauer.com, and the real difference for Zulily, and I think what puts a different spin on the technology aspect of it is this notion of launching the newspaper every day. And, instead of going through a seasonal type flow, say that an Eddie Bauer would go through where you are going to have a fall line and a spring line and a summer line …. here we are working with thousands of vendors, we are launching 20, 30, upwards of 40 events a day, thousands of styles of product and we are doing that every single day, seven days a week…. Those things put a stress on your technology infrastructure and they put interesting problems in front of you. And one of the reasons I am here, and why really, really good engineers are attracted to Zulily is the problem space of this type of online business…. We are not here because of the technology, but what I would also say is that we wouldn’t be here without the technology. You can’t be an online retailer without amazing technology, and be ultra successful. You can be average and make due and plug along, but if you really want to be the best in your space, if you want to be out in front and innovating and bringing amazing new features to your customers every week, you better have a pretty talented technology staff, and you better focus on it.”
On balancing the need to move fast with innovation: “It is always a balance, that balance of moving quickly and innovating quickly, but learning from your mistakes, and applying those the next time. But one thing we do balance is: ‘Hey, it’s OK to make mistakes.’ Be honest about it, and that’s right out to our customers, and then fix quickly, learn and continue to move on. It is something that we literally talk about every day.”
On how Zulily is analyzing customer data: “We track a ton of data about our customers….The great thing about the private sale piece is that you have that user name, you have that account, but there are all sorts of techniques we are using to capture the experience of the customer, and we are also trying to do more and more of asking the customer (via) email outreach, surveys, onsite, social, whether that is Facebook, Pinterest, Twitter. We listen a lot. And I think that is what gets us back to this customer space. They talk a lot. They give us great information…. We are bringing that click data in continuously and this is where things like Mongo and Cassandra and some proprietary software that we are building out allows us to mine that data and look at data, and then how do we leverage that in the experience that we are providing that customer, whether that be through the email they receive or the mobile experience if they’ve downloaded one of our apps, the mobile site, or coming back to the site itself, everything from what events do they they see, and what products are showing up and where.”
What’s the one feature you’d like to add to Zulily today? “One that we are working on and is out in tests right now is we’ve done quite a bit of work around the shopping cart, and the checkout process. I am a huge believer in that you’ve got to take friction out of process. So, for our customers, how are we taking friction out of the shopping process. One of things we’ve been doing lately is spending a decent amount of time …. on a new experience around their shopping cart and how they are able to interact with that shopping cart, and how do they go from that shopping cart to a quick purchase — because with timed events and daily deals, there’s that sense of urgency, and I need to be able to get through quick, make it easy for me, make it painless for me.”
What percentage of your customer base interacts with Zulily on mobile? “I don’t think I can give you specifics, but what I can tell you is that it is meaningful, and it would be more than a typical e-commerce, online apparel retailer…. What we see with our customers, especially the time of day our events go live, many of them are not going and cracking open their laptop or cracking open their PC. They are picking up their iPhone, picking up their iPad, picking up a tablet or device…. It is a big part of our strategy, and it will continue to be a big part.”
What’s your biggest challenge right now as CIO? “Speed. It really is. It is the speed at which the company is moving.”
What are the questions you ask potential software developers that you are looking to hire? “We literally give people problems to have them solve. We stay away from the more traditional puzzle questions, but we do ask the real math questions, and we try to get people to solve them. Most of the interviews are spent at the whiteboard. And it is not just the candidate at the whiteboard. It is the candidate and the person doing the interview, and they are literally collaborating together. What we are trying to figure out is, can you work in a collaborative environment? We want people who can think on their feet, and think about the solution as much as the technology that might be applied to it.”
On whether they’ve looked at deploying robots at fulfillment centers: “It really gets back to the speed thing. We opened these two facilities in less than six months, so they are still pretty manual in how we are doing it. They are effective. They are efficient. Drugstore.com was starting to roll out Kiva (robots) when I was leaving, and the challenge that we face is our SKUs and products are turning over every 72 hours whereas drugstore.com might have this bottle of water and they sell it in a certain pack, and they are going carry it for the next year-and-a-half. It doesn’t change. I know the size of it, I know how much it weighs, and so I know how I can slot it on this robot…. It doesn’t necessarily work for us, but does automation work for us? Absolutely. And it will be a thing that we will continue to look at, and it will evolve.”
How closely do you watch what Amazon.com is doing? “You do watch them. Obviously, when they went and decided to buy Kiva, that was an interesting thing to see. On one hand it was a surprise, on the other hand it wasn’t a surprise at all. It makes sense, considering who they are and what they are trying to do. You have to watch them, and you have to learn.”
Any lessons from Amazon? “The thing I’ve really learned from Amazon and apply it everywhere — what they’ve done is that they are big believers in building. Build it. Build it. Build it. Own it. Own it. Own it…. And that’s what we are doing here. We are making a major investment in technology, but that investment is really in software engineers. We are not going and trying to figure out what the right merchandising platform that we can buy to run Zulily’s daily newspaper process, because there isn’t one out there….What is cool about Amazon — and I think it is the way to look at it — is build strategic, buy commodity. So, don’t become completely dependent on some software company to make sure that you can continue to innovate and grow and move fast. And that’s one thing that Amazon has done well over the years, and continues to do.”
How did you get into the software business? “Oh, geez. It was interesting. I’ve been in this area all of my life, and dropped out of college. I went to just about every junior college in the area. I actually wanted to be a ski and snowboard buyer, because I love to ski and I love to snowboard. I was living in a house with like four or five of us, and we all decided to go out and buy PCs. And this would have been the late 80s or early 90s, and we started hacking around, doing bulletin board stuff and so forth. And I just found myself really enjoying it. I went back to school and said I was going to bone up with a computer science degree. I got about half way through it, and then a buddy was starting a company, a collegiate marketing company …. and I said, ‘Ok, I will come help.’ Everybody has their path, and this one worked for me.”