StyleFeeder’s Execs on How to Do A Lot With A Little

(Page 2 of 2)

another bubble. We’ve taken in just $3.5 million, and you just met the entire staff of six people. There’s nobody else.

X: I thought it was unusually quiet when I came in, like everyone was out to lunch.

PJ: It’s always like this. The engineers don’t talk much. No, seriously, we have what I refer to as the “no-passengers rule” for hiring, which basically means there are no pure managers here. Shergul runs business development, which means he is on the phone all day long. Dina Pradel [the vice president of marketing] will sit down and type our e-mail newsletter, which goes out to our one million registered users. We also have a very technical team—there are only architect-level people here, not junior-level coders. We turn those people on as freelancers that we find on places like Authentic Jobs. That keeps the costs down and enables us to burst if we have a bunch of projects that suddenly need to get done.

X: How fast have you been growing—can you give me a few numbers?

Shergul Arshad: In the month of March, we had about 2.1 million unique visitors and 6.1 million page views. Those are both new highs for us. With the staff the way it is now, as Phil mentioned, we are profitable. I think the big focus for us has been not only getting to profitability, but also this type of partnership that we have with Elle. That opens up a very interesting path for StyleFeeder. There are a few sites that have focused just on the shopping piece, like ShopStyle and TheFind, and there are obviously all the price comparison engines like Shopzilla and But we can bring a lot of interesting community elements and the recommendation aspect, which no one else really has at scale. And we’re doing that with only six people, whereas some of those other companies might have $25 million in funding and 30 employees.

X: Well, six people plus all the freelancers you mentioned writing the actual code.

PJ: We all code, we do development work, but the grunt work—the HTML stuff that is time consuming but is not high-value—we outsource to a network all over the world, from Brazil to Romania to Colorado to Vancouver. To save money, we also rely heavily on Amazon Web Services and on our hosting partner, Contegix. We don’t own a single machine—everything is in St. Louis [at Contegix’s data center] or in the Amazon cloud. We are one of the highest-volume users of SQS [the Amazon Simple Queue Service, which passes messages between computers]. We push tens of millions of messages through the Amazon servers every day, and we have tens of millions of images and other digital assets stored up there on the cloud.

SA: We have over a million registered users, so we have shopping preferences expressed by over a million users. Not only that, but we also have about 14 million products imported from retailer’s data feeds, from thousands of retailers.

We mainly focus on soft goods, things that people shop for based on preferences—especially clothes, furniture, stuff for the home, shoes. We’ve got all the TVs and camera equipment that you want, and the site works quite well for that, but we’re at our best when things get fuzzier—“It’s a shirt and I don’t know why I like it, it’s my style.”

SA: A shopping site isn’t supposed to be a place you come and spend all day. It’s not a portal, it’s not MySpace or Facebook. But at the same time there is a subset of young female shoppers who love to shop and will spend all day doing it. So we need to cater both to the light user who will be one the site for a minute as well as someone who wants to be on there all day, and we’ve struck a very good balance. For the user who wants to come in and spend all day, the more items you rate and the more inputs you give, the better your personal shopping experience will be.

We’ll get messages from users saying, “I spent all day in class rating products,” which may not be the best thing from an educational standpoint, but we may actually educate them more than their class. The point is that there are heavy users who want to spend all day interacting with the site, but for a Google visitor who comes straight to one of our pages, we offer quite a bit of content right there on the page: price comparisons, product information we’ve extracted from the retail catalogs, user comments, related items. We will also do a thing where, right off the bat, we’ll ask them if this is the exact item they were looking for, and if the answer is yes, we’ll take them straight to the retailer, which accelerates our path to revenue.

PJ: Say you’re on Google searching for Reebok men’s Turbo running shoes. You click on a link that goes to Stylefeeder, and you have the opportunity to say, “That’s what I want, now get out of my way,” or, “Not really, but something here seems close.” Two-thirds of the time, people are clicking the “Not really” button, which to me is an indictment of Google’s inability to help people with product search, which creates an opportunity for us. It says that people are talking to the search engine with the best terms they can come up with but two-thirds of the time they are still not finding it.

We give a lot more value to the user because they are then shown a bunch of things that are in the zone of what they want. And the hit rate for us on that is really good. We also have product buying guides. If you’re looking for an espresso maker, we’ll show you an espresso maker buying guide. This is a new initiative for us, but there are buying guides for a range of products from dress shoes for men to how to dress for your new job.

SA: If we can get someone off our site to a retail site by matching them up with the right product, that will increase conversion and end up monetizing the visit for us. To do that, we need to be as accurate and informative on that first page as possible. But there is a lot of original content that we are also offering up, which is critical to ensuring that Google visitors will come back to us next time. Right now about 14 percent of people who visit us from Google return within a month.

PJ: Which is just a gobsmackingly high number.

Continue to Part 2, in which Jacobs and Arshad expound on Facebook, machine learning, affiliate marketing, and how men and women shop differently.

Single PageCurrently on Page: 1 2 previous page

Wade Roush is a freelance science and technology journalist and the producer and host of the podcast Soonish. Follow @soonishpodcast

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

Comments are closed.