So, how good is Like.com?

Here, let’s try a few things and see how Like.com works.

Let’s look for these things:

Watch with square face. Worked pretty good. I loved the square dancing face.

Steel tipped shoe. Found only three results. I’d expect more.

Black dress shoe. Lots of nice results.

Diamond and garnet ring. Maryam would like this.

Silver watch. This brought back a lot of silver watches.

Pretty good, although I want a lot more choices than what they’ve given us.

The problem is it’s taken 250 servers with four processors each to do just this small subset.

Comments

  1. Pretty interesting. I was looking for ‘Shoes with Wheels’. Impressive results.

    But as u had pointed out there is a very restrictive subset of products. For eg. I was not able to search for a ‘black slim mobile’. Think it is initially only for Jewelry, shoes , watches & handbags !

    Or am I missing something ?

  2. Pretty interesting. I was looking for ‘Shoes with Wheels’. Impressive results.

    But as u had pointed out there is a very restrictive subset of products. For eg. I was not able to search for a ‘black slim mobile’. Think it is initially only for Jewelry, shoes , watches & handbags !

    Or am I missing something ?

  3. Cross product matching would be nice, for example show handbags that “go with”* these shoes.

    * The concept of items “going with” other items is lost on me but my wife often uses it as an excuse to buy a whole new wardrobe to go with her new shoes.

  4. Cross product matching would be nice, for example show handbags that “go with”* these shoes.

    * The concept of items “going with” other items is lost on me but my wife often uses it as an excuse to buy a whole new wardrobe to go with her new shoes.

  5. Sorry, this is NOT the least bit exciting or impressive. I tried over two dozen search terms for clothing items that I’ve searched for before. Eight had results – that were not even close. The rest returned no results. I finally chose a generic watch so I could do the likeness search. Color match worked well. Shape match average. Pattern not at all.

    You want to impress me, get some selection in there – men stuff – maybe some hand tools, then let me sketch an outline or submit a picture. If they can find things that way… I’ll be impressed.

  6. Sorry, this is NOT the least bit exciting or impressive. I tried over two dozen search terms for clothing items that I’ve searched for before. Eight had results – that were not even close. The rest returned no results. I finally chose a generic watch so I could do the likeness search. Color match worked well. Shape match average. Pattern not at all.

    You want to impress me, get some selection in there – men stuff – maybe some hand tools, then let me sketch an outline or submit a picture. If they can find things that way… I’ll be impressed.

  7. Nice article. Riya have some competition on this one – London based PIXSTA launched http://www.chezimelda.com recently – a huge online shoe store browsable by image. Comparing the two, Like.com’s site is more polished and great fun (I like the celebrity section), but for a PURE shopping experience http://www.chezimelda.com is much more intuitive. It will be interesting to which one consumers go for. The speed is also more Google like!

  8. Nice article. Riya have some competition on this one – London based PIXSTA launched http://www.chezimelda.com recently – a huge online shoe store browsable by image. Comparing the two, Like.com’s site is more polished and great fun (I like the celebrity section), but for a PURE shopping experience http://www.chezimelda.com is much more intuitive. It will be interesting to which one consumers go for. The speed is also more Google like!

  9. [...] I can already picture some of the ladies going crazy over this one. If it actually works. So far Riya’s new product Like.com, a visual search engine seems pretty nice (its still alpha). Robert Scoble ran some tests to see how effective it is. Robert says that it works, but he would expect more results. I concur. I ran some of my own tests, and even though they produced results, the number of results wasn’t enough. I don’t know if they don’t have enough products listed, or its their search algorithm which is not fine tuned enough. [...]

  10. Like.com seems to be biased towards amazon for results. This is quite lame from a consumer’s perspective since amazon is not nearly the best place for shoes. Maybe it will help Munjal and Co. to cash out to amazon since they failed in their bid to sellout to Google. Like.com is clearly positioned as a company that desperately wants to sell out to Google or amazon. Note the lack of biz model. It looks like they aren’t worried about making $ since they just want to sell out.

    Most people who buy shoes online know that Zappos.com offers the best selection (brands, styles, colors, widths – you name it) and customer service, *bar none*, and yet Zappos results are hardly represented. Zappos definitely has more matches to any kind of shoe – way more than amazon or any of the other copycat shoe sites – and yet Like.com mostly steers consumers to worse stores. Lame, lame, lame from a shoe consumer’s perspective!

    Well, good luck selling out (the second time), Munjal!

  11. Like.com seems to be biased towards amazon for results. This is quite lame from a consumer’s perspective since amazon is not nearly the best place for shoes. Maybe it will help Munjal and Co. to cash out to amazon since they failed in their bid to sellout to Google. Like.com is clearly positioned as a company that desperately wants to sell out to Google or amazon. Note the lack of biz model. It looks like they aren’t worried about making $ since they just want to sell out.

    Most people who buy shoes online know that Zappos.com offers the best selection (brands, styles, colors, widths – you name it) and customer service, *bar none*, and yet Zappos results are hardly represented. Zappos definitely has more matches to any kind of shoe – way more than amazon or any of the other copycat shoe sites – and yet Like.com mostly steers consumers to worse stores. Lame, lame, lame from a shoe consumer’s perspective!

    Well, good luck selling out (the second time), Munjal!

  12. In my previous post ,where I liked the concept, but hate the home page (because of the lame celebrities pictured), I made a mistake.

    It’s not K fed/Brittney, it’s Brittney and Fed -Ex.

  13. In my previous post ,where I liked the concept, but hate the home page (because of the lame celebrities pictured), I made a mistake.

    It’s not K fed/Brittney, it’s Brittney and Fed -Ex.

  14. I think the site would do much better if you did a ‘likeness search’ on a steel tipped shoe rather than a text search for a steel tipped shoe.

  15. I think the site would do much better if you did a ‘likeness search’ on a steel tipped shoe rather than a text search for a steel tipped shoe.

  16. Oh I Sould say I have a couple female clients with a ton’e of products targeted at a mostly female audiance how do i get their products on there is it just a US thing?

  17. Oh I Sould say I have a couple female clients with a ton’e of products targeted at a mostly female audiance how do i get their products on there is it just a US thing?

  18. I’ve just started a wordpress hosted blog site and have been looking for some facial recognition software. The http://www.polarrose.com/ website also seems to hold promise. Right now it seems to be a powerpoint being developed into a beta. Difficult software to develop. Now that I’m aware of this

    I have a project under development that could really use visual recognition (of badly scanned personal photos) and tie it into a semantic analysis/mapping/social networks timeline based around MIT’s SIMILE (http://simile.mit.edu/timeline/docs/) with lots of AJAX-ATLAS-LINQ-PYTHON-NAME-YOUR-OOP action going on to encourage visitor interaction and participation.

    If I may be so bold as to add that I’ve officially dumped/blogged my girlfriend for Dee Dee Walsh. I know you danced on tables with her at GeekFest. We have a deeper relationship where jealousy doesn’t exist. She just tells me what to do and I do it. Simple.

    I also shamelessly name-dropped. Chosing Robert Scoble, Dee Dee Walsh, Jonathan Zuck and Tom Button to link together in my “first” blog is either a shameless commercial plug or hanging out way too long with the developer marketing teams.

    No link included out of total respect. If I haven’t grabbed your interest yet…

  19. I’ve just started a wordpress hosted blog site and have been looking for some facial recognition software. The http://www.polarrose.com/ website also seems to hold promise. Right now it seems to be a powerpoint being developed into a beta. Difficult software to develop. Now that I’m aware of this

    I have a project under development that could really use visual recognition (of badly scanned personal photos) and tie it into a semantic analysis/mapping/social networks timeline based around MIT’s SIMILE (http://simile.mit.edu/timeline/docs/) with lots of AJAX-ATLAS-LINQ-PYTHON-NAME-YOUR-OOP action going on to encourage visitor interaction and participation.

    If I may be so bold as to add that I’ve officially dumped/blogged my girlfriend for Dee Dee Walsh. I know you danced on tables with her at GeekFest. We have a deeper relationship where jealousy doesn’t exist. She just tells me what to do and I do it. Simple.

    I also shamelessly name-dropped. Chosing Robert Scoble, Dee Dee Walsh, Jonathan Zuck and Tom Button to link together in my “first” blog is either a shameless commercial plug or hanging out way too long with the developer marketing teams.

    No link included out of total respect. If I haven’t grabbed your interest yet…

  20. [...] Like.com (from Riya) is a visual shopping system.  There are a few hands-on reviews and commentary you can read (pick your favorite: ZDNet, GigaOm, TechCrunch, BusinessWeek, Robert Scoble, Thomas Hawk).  I’m going to skip the usual rigamarole of anaylizing how good their algorithms work, go check out one of the others’ for that.  Whether it’s perfect or not today, I generally have enough faith in engineers and entrepreneurs that they’ll continue to improve it to the point where it works as promised.  And if they don’t, well, someone else will come along instead.  Instead, I’m a bit more interested in the launch itself. [...]

  21. @ Jojo

    Note the lack of biz model. It looks like they aren’t worried about making $ since they just want to sell out.

    Like.com actually oozes financial model. All the products listed earn revenue. It takes time to add additional resources for increased product range and price comparrison, so it seems they have opted for the easiest to integrate (and the ones that are guaranteed to pay on time).

    My own opinion of the value of the service was enhanced when I showed my wife it.

    Now while she didn’t buy anything (buying from the US and shipping to Europe is not ideal), she did spend a couple of hours just shopping for ideas.

    Like.com makes an ideal “window shopping” environment and the amount of products is always likely to increase in both depth and width.

  22. @ Jojo

    Note the lack of biz model. It looks like they aren’t worried about making $ since they just want to sell out.

    Like.com actually oozes financial model. All the products listed earn revenue. It takes time to add additional resources for increased product range and price comparrison, so it seems they have opted for the easiest to integrate (and the ones that are guaranteed to pay on time).

    My own opinion of the value of the service was enhanced when I showed my wife it.

    Now while she didn’t buy anything (buying from the US and shipping to Europe is not ideal), she did spend a couple of hours just shopping for ideas.

    Like.com makes an ideal “window shopping” environment and the amount of products is always likely to increase in both depth and width.