There are a few areas that startups can’t resist trying to fix, despite the fact that almost every attempt almost inevitably ends up crashing and burning. Take online music, where dozens of failures in previous generations haven’t stopped the likes of Rdio, MOG and Spotify trying to make it work this time around. Or virtual currencies — a graveyard more heavily-populated than Arlington’s.
Another one of those perennial problems that never seems to quite get solved is product search. Plenty of clever people have come up with ways of making finding the perfect purchases online easier and better, but only a handful have ever managed to actually compete with the likes of Google, (s goog) Amazon (s amzn) and the rest.
None of this is stopping the team behind Ineze (pronounced “in-ez-ay”; it’s Hebrew for “here it is”). This small gang of Israeli technologists think their way of crunching data can help make finding the right product easier and better than ever. The idea is fairly simple: that lots of people search for information about products, but they find it hard to uncover what they’re looking for.
“Lots of people look through the long tail of product information through Google — sometimes they are people crazy enough to care about a specific feature, such as the weight or thickness of a phone,” says Chief Executive Gil Rozenberg. “But that’s not how shopping sites list information — in fact, doing that is impossible for lots of shopping sites because they’re just doing plain text search.”
Rozenberg’s argument is essentially that existing product searches fail because they require somebody to organize information into taxonomies that are inflexible and not necessarily accurate. Instead, Ineze collects product information wherever it can get it, and breaks it down into chunks that can be more easily compared, contrasted and updated. That means that if you want a specific feature, it can look at the data and pull out the options. Equally, if you want to compare the market over a set of criteria, it can run the numbers and present the best option.
Here’s an example: Let’s say I’m a speed freak, searching for the fastest high-end phone on the market. On most shopping sites, all I can really do is look under the general “cell phone” category and start sifting through popular models to find the information. There are, maybe a few pre-selected criteria I can sort by — price, brand, maybe operating system. Processor speed, most likely, isn’t going to be on there.
Over on Google’s product search, meanwhile, things may be a little better. If I type in “fastest smartphone”, I can get a list of results that include the phrase ”fastest smartphone”. That’s better than nothing, but it’s full of problems — not only is it clearly open to abuse (is a phone the fastest just because the blurb says so?) but in a rapidly changing market, being fastest is a moving target.
Rozenberg says something better is entirely possible, because the data is there — it’s just poorly interpreted. So when I type “fastest smartphone” into Ineze, it determines what “fastest” and “smartphone” means (processor speed and capability), compares that data across the possible options and offers me the highest-rated result. In this case, the Samsung Galaxy S with a 1.4Ghz processor. It also tells me that the Palm Pre 3 (s hpq) is the same speed, with a smaller display, while the Samsung I9100 and the BlackBerry Bold Touch (s rimm) are both 14 percent slower.
Right now, usage is small. The site has just launched properly, and Ineze provides product search across one vertical: phones. The product is simple, but not particularly user-friendly — all things that the team is working on improving.
But they’ve got a little bit of buzz and momentum, particularly since being named as one of the winners at the most recent Seedcamp, the well-regarded European accelerator program. As a result, they’re working on making things better and expanding to new areas, such as tablets, cameras and television sets. Rozenberg says that the site needs to expand and improve rapidly if it’s going to fulfill his ambitions.
The company says it may open its search functions as an API for other sites to use, but one thing Rozenberg doesn’t want to do is simple white-labelling for other commerce websites, since providing a broader, cross-industry search is where the opportunity lies.
“The bigger vision is to build a destination site for product information. If we can build the ultimate product pages and let people access them with hardly any knowledge that’s great. In the bigger vision, I believe that within 18 months we could have more than 100 verticals.”