I met with Andrew Holt, co-founder Dontbuyjunk, after Search Engine Strategies (SES) in August. Finally caught up with him again for an interview. Andrew was a computer science major at Stanford. He left Yahoo! Shopping last August after almost 2 years at the company. He is currently working on Dontbuyjunk with his partner, Rishi Khaitan.

History…
“Rishi and I are product enthusiasts. Friends would come to us all the time and ask us what product to buy. We wanted to provide a clear, concise, and accurate technology solution to answer that question.”

“We spent significant time developing our technology, TotalRank. We modeled it after a knowledgeable friend. The goal is to form accurate recommendations personalized through technology. There are three parts of TotalRank:

“Ratings. We’ve tried to make this very comprehensive with user reviews and over 200 sources of web reviews. Products are rated on a list of attributes.”

“The confidence level we have in a rating. In terms of confidence, we look at review quantity, age, consistency, and source quality - all things that people would weigh when determining the relevance of reviews. TotalRank lists products ranked by how they would be recommended. The products at top are there because we’re confident they have a good rating.”

“Personalization. We dynamically re-rank the product list based on attributes that the user finds important. If you’re looking for portable mp3 player, you go to friend and he personalizes rankings for exactly what you need. We’ve found that it’s very accurate.”

“We’ve received a lot of feedback from enthusiasts saying that the TotalRank results match what they would recommend. In the end, the goal is to answer the question of what to buy. We don’t sort by price or popularity, as these things that aren’t as relevant if you’re trying to figure out best product for you. If you select everything you feel is important, you could buy the first product [listed] and that would be the best product for you.”

What about competition…
“Editorial review sites. CNET and Consumer Reports are examples of this, but the problem is that CNET and Consumer Reports can only cover a small number of products and it’s just the opinion of one editor. Also, sometimes there’s discrepancy between CNET and Consumer Reports. There are flaws if you use them individually.”

“Shopping comparison engines. The problem is that you have to sift through reviews to figure out what to buy. You can sort by ranking, but the highest ranked product might only have 1 review. Also there may be inconsistent reviews.”

“Vertical product search companies. Right now there’s only Become.com. It’s a great utility, but it’s for people who have narrowed down the choice of products. If you haven’t narrowed down your search, though, there’s nothing to help you. Become.com is like requesting a research report and only getting a bibliography.”

“In the end, the best way to figure out what to get is through a friend. However, if people come to me and ask what TV to buy, I’d have to go and catch up [on the current product offerings]. The great thing about our site is that it’s very comprehensive. Any review about a product will be aggregated for you. It’s like having a friend who knows everything about products.”

Tell me more about TotalRank…
“It was totally built in house. Before we started the site, we sat down for 2 months in front of a whiteboard and figured it out. It’s a lot more than a weighted average. We’ve changed it many times (right now we’re on about the 5th revision) to make sure it’s really accurate. Right now it’s patent pending.”

Why haven’t the shopping comparison engines gotten into this? Are any of them coming close?
“Epinions seems to work well enough and unless a product is an absolute piece of junk, people are pretty satisfied with shopping comparison engines. It’s a good business model from their perspective. If consumers want to click around before they buy, the shopping comparison engines are getting revenue.”

“If we can answer this question, what should I buy, though, there is no reason [for a consumer] to go anywhere else. That’s our goal. For the shopping comparison engines, it’s a peripheral goal and they aren’t going to switch their business model.”

Your business model…
“Right now we’re using the Shopping.com merchant links. We can expand beyond this, and we’re thinking of building our own way to help people purchase. We don’t have advertising anywhere [right now], but when a consumer gets down to where to buy, we can put up a lot of stuff beyond advertising links like those found on shopping comparison engines. Snap.com is doing a lot of interesting things with advertising. We could even have a fully automated system. When you have to hand hold merchants, the scalability really gets hurt.”

“At this point, we’re still thinking about answering the question what to buy. Because we help the users figure out the answer, the conversion for merchants will be very high. When the user gets to the where to buy link, they are really targeted. We can offer the merchants a great way to advertise.”

How do you market the site?
“Currently, it’s just word of mouth, and that’s working fairly well. Because [the service] is so unique, we’re finding that people tell others. We will use AdWords, the search engines, and various forms of internet advertising, but since we’re self funded, we can’t do a lot. However, we are starting to look for funding.”