Yesterday, I wrote a piece about Like.com, a visual search engine company based in Silicon Valley that was reporting great progress: The young site has had rocketing visitor traffic over the past four months, and it’s headed toward profitability. People are using it to find the color or style of shoes they like, and then clicking through to merchants to buy them, and Like gets a cut, according to chief executive Munjal Shah, who showed me his internal data. The site today released its new, slicker look (see Shah’s post about it here).
In the piece, however, I also quoted Shah as saying Like.com is “killing” the social shopping engines, including Santa Monica, Calif.-based ThisNext. He argued that’s because shoppers don’t like sharing their shopping preferences with each other; shopping decisions are more solitary than most people think. Critics took me to task for not investigating Shah’s claims more deeply, so I decided to do a little more digging, even though I provided quite a bit of context and justification already in the original piece and in subsequent clarifying comments.
Namely, I talked with Gordon Gould (left), chief executive of ThisNext, and I’ve got a lot more to say: In short, ThisNext practically hasn’t gotten started yet, and it faces huge, perhaps insurmountable challenges. However, its vision — if it ever is realized — is more profound and much more powerful than Like.com’s. It’s true that Like.com is indeed killing ThisNext for now on a revenue basis. But it’s really not fair to use the “now” as our measure, because ThisNext is using venture capital to build out a bold, audacious “product graph” project and isn’t yet focused on making money, or even on maximizing traffic.
So let’s go back to the Like.com v. ThisNext debate, and I’ll tell you why I think that, in the end, this is an apples to oranges comparison that isn’t fair.
I initially told critic Fred Wilson his reaction centered too much around the traffic question and not about their respective revenues, which is what Like.com’s Shah says he’d intended to refer to. Both measures, I’d argue, should be looked at when judging success, because one can often be traded off against the other, especially in the early days of a company.
So first of all, let’s start with traffic. Fred showed an old Comscore graph through December, depicting how the traffic of both sites is comparable through most of last year. However, that graph showed Like.com’s traffic exploding upward in December, which Like.com’s Shah (left) explains was due to a number of factors, but none a primary driver. He admitted he was buying traffic (by placing ads next to Google search results; for example, next to a search for “red snappy shoes”), but he was doing a lot of other things too, which we noted here and here. His traffic continues to be robust through Jan and Feb, and is now at 3 million unique visitors. ThisNext, meanwhile, only got 1.4 million uniques in January, says Gould. So, no question Like is beating ThisNext, but ThisNext isn’t buying any traffic, Gould tells me.
Now let’s go to revenue. Like.com’s Shah says he’ll make more than $10 million this year, Gould says he’ll probably make about half that. Again, Like.com wins. But again, ThisNext’s Gould has an out: He’s not focused on making money at this stage.
Like.com is aggressively focused on an arbitrage business, buying traffic at a price and then making more money on the traffic it gets than what it pays for the traffic, Gould argues. ThisNext, by contrast, is in the “discovery” business, he says.
Like’s Shah is flat out wrong that people don’t consult with others when buying, Gould adds. If people are using Like.com, they’re likely to be further along in the shopping process, Gould concedes. But otherwise, people often go into the store and end up buying something they never planned to buy. “It happens all the time,” Gould said, both in offline and online shopping. If a cute saleswoman says you look terrific in those jeans you’re trying, you’re gonna buy those denims even if they’re $20 more expensive. If your friend likes them too, you’ll probably buy two pair. Online, people — especially women — send IMs and email back and forth with URLs and coupons of stores or products they’ve found, he says.
Some 70 percent of shopping is made on a discovery basis, and only 30 percent is made from search, Gould says, citing research. Research from Womma.org shows that 27 percent of social conversations have some meaningful discussion of products, suggesting that social interaction is powerful input into our decision making.
To help in the shopping discovery process, ThisNext is building what Gould calls a “product graph,” or essentially a compilation of the web of connections between people who buy trendy products and other people they influence, as well as a graph that tracks the products themselves. What you end up with is a dataset that lets you know who the influencers are and which products are popular in which populations. The challenge for ThisNext, then, is to grow large enough so it can become the leading “product recommendation” service across the Web.
For instance, if you are an influential flyfisher and recommend products to others from your MySpace page, ThisNext wants you to be able to access those same recommendations while on other sites, via social network platforms offered by Facebook and others. Right now, the site is trying to optimize its reach across the web, and so traffic to its home page isn’t the primary goal.
If it gets there — and this will be very, very hard, in my view — ThisNext could be a boon to the social sites such as Facebook and MySpace, which are looking to make money. By becoming a plug-in social shopping layer in each site (just like Photobucket offers the photo layer for MySpace, YouTube offers the video layer, and PayPal offers a purchasing layer, etc.), ThisNext would earn revenue each time a user buys something through it, via an affiliate payment. ThisNext would then share the revenue with these social sites. With an effective CPM advertising rate of about $30, ThisNext’s offering would be stronger than many other ways of making money, Gould says.
In the long run, eBay, which already has lots of information about its shoppers and preferences, could be headed in this direction and so poses ThisNext’s toughest competition going forward. Other players, however, have fallen by the wayside (Kaboodle was bought by Hearst and no longer looks like a threat, says Gould; and Stylehive and Stylefeeder appear too focused on fashion, he says).
To conclude, comparing ThisNext with Like.com seems premature. Has Like.com really killed social search? Yes, probably, in the short term, if you’re only looking at the numbers. But social shopping search and shopping technology is still in its infancy, and we haven’t given it a chance. It’s absurd to make any conclusion right now that social shopping search has lost to visual shopping search.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.