Snap Rethinks Search

As I've pointed out many times, Bill Gross, the man behind Snap (and Goto/Overture and about 25 other search related companies) is not one to take lightly. When Snap launched, I watched closely, and while many of its features were admirable (the transparency, for one, and later, the CPA…

Snap NewAs I’ve pointed out many times, Bill Gross, the man behind Snap (and Goto/Overture and about 25 other search related companies) is not one to take lightly. When Snap launched, I watched closely, and while many of its features were admirable (the transparency, for one, and later, the CPA model, for another), it never quite got enough lift under its wings, at least in its first year.

Today Snap is relaunching as a “broadband search engine.” That means, it’s heavy on Ajax features, clustering, and related results, among other things. It certainly is a new look. The results include large thumbnails of prospective pages, for example, and a suggested terms autocomplete feature (not unlike Google Suggest). In fact, there are tons of features that have been tried in various other places, but never have so many been implemented in one place at one time. It’s an attempt to fight one’s way out of the single search box interface, and whether it works or not, it’s worth a look. The theory is sound – which is usually the case with Gross’s companies – but often he’s ahead of the market.

My quick take is that I’m so used to Google’s dominant interface, I initially got lost using Snap. It takes some time to get the hang of it. One thing that I want to do is click through directly to the site, but instead, I’m in a window on the right. There’s much to think about here – it feels more like search inside a multi-column RSS reader like Shrook, oddly. You can ask Snap to show you a new window, which is good.

Snap is also rethinking relevance by licensing ISP clickstream data and feeding it back into its relevance engine. This to me is where the really interesting stuff lies. It’s a way to fight spam – folks tend not to spend a lot of time visiting spammy sites – and, in an ideal world, provides a potentially better set of results than simple link analysis.

The Snap model incorporates paid inclusion and pay per action, and I think this may be where it falls down. While this is certainly an innovation in the affiliate/adsense spam market, it’s also open to charges of blurring the lines, which was exactly the problem with Goto when it launched. We’ll see.

Snap is promoting its new engine and its launch on Searchblog, among many other sites. As part of that promotion, I’ve agreed to head over to Snap’s blog and include my own ideas in their “Other Way to Launch” contest. I’ll be doing that soon, and will add the link here so you can see what I wrote.

Author: John Battelle

A founder of NewCo (current CEO), sovrn (Chair), Federated Media, Web 2 Summit, The Industry Standard, Wired. Author, investor, board member (Acxiom, Sovrn, NewCo), bike rider, yoga practitioner.

15 thoughts on “Snap Rethinks Search”

  1. glad to see they upgraded from gigablast to Ask for results, but there are ways to do what they are doing with the interface without making it so clunky. Using the scroll on your mouse for the results just plain sucks. Having to click on a search result once here, and then once over on the right to open it up is horrific! As far as their cpa model goes, that is a chicken and egg problem there. Gross should have created a separate company for CPA ads, and then syndicated. Snap doesn’t get good enough traffic for them to figure out if the CPA is going to work on a wide scale.

    John, why don’t you check out Dumbfind for someone trying something truly different in search? And we have our own index (although that is a drawback now because it is small, it shows we are serious about making something new, and also our methodology requires it). Oh yeah, looks like snap borrowed our little + button for displaying more results within a domain. I could be wrong, but I think we were the first to do that. Not that I blame them, it is pretty sweet functionality.

  2. Interesting comments. Here’s my first reaction when I went to – how is this going to work with mobile smartphones?

    My guess is – not very well. AJAX is not well supported on mobile phones and performance has a way to go. Why they don’t compress the search results is beyond me. It would speed things up considerably.

    Mobile is the future – anyone designing a search engine and omitting this market is behind the curve.

  3. The trend now for newbies is to focus on design and interactive extras – because it is vitually impossible for any company to invest in RELEVANCY that would in any way equal Google Yahoo and MSN.

    Every Algo combination, probably has already been thought of by GYM. The only option would be to become a META Search Engine and a PPC and add interactive AJAX Web 2.0, and Social Networking extras and keep adding them as they evolve

  4. SE Web: I think you’re a bit off base here. GYM themselves have said that they’ve barely scratched the surface of search.

    And I think you’re mixing up “relevancy” (or, as it is more properly known, “relevance”) with the current GYM focus on ad hoc, known item “lookup” search.

    There is a whole heckuva lot more out there than what the majors are doing right now, at it is far from just the newbies that are looking into it.

  5. While I think that one should review and comment on a number of sites, I do think that you get into a grey area when you start to review sites that you are paid contributor/advisor to. Fine, if you talk about MSFT, and they have ads here, but if Snap is paying you to post on their site and promote their site, then it gets into the question: Was the above post because you found Snap interesting or because they are paying you? Would you read a review of a movie by someone who was a paid advisor to that movie? I would not.

    I personally would rather you let advertisers buy whatever ads they want on this site, but not get into the game of reviewing here sites you are being paid to promote. They can place ads here, but that does not mean you are going to post or comment on them. Otherwise, this makes for a grey area in my view. What about this: have someone that you respect who is not a part of their promotion campaign review their site here? That to me would be a better approach — and I think we are more likely to get a clean review. I think people want a “messenger” to “say it like it is” and if you are being paid by a company to make comments on their site, I am not sure how good you can do this. WIthout this, where does one draw the line between reporting (saying it like you see it) and promotion (saying it in the best way for a particular comnpany)?

  6. Thanks Soren, for your comments. While it’s true that Snap is paying for ads on the site, and I agreed to comment on their site as part of their promotion, I hope that you’ll understand that it in no way colors my thoughts about the company. I have written about every single thing Bill Gross has done. This is no different. I only take advertising from companies I find interesting in the first place. That’s the beauty of FM’s model.
    However, if this crosses a line for you, I will take that very seriously and consider it in the future. Thanks.

  7. The interface is hard to get used to. Google uses text results, but to a very effective degree. More results are displayed and you scan through them quickly with your eye. With Snap, you see less results on one page. And even though the images I assume are cached, they still take that split second too long to load.

  8. A9 is a meta search engine, so is snap. . . crawling the web requires too much investment in resources due to increawe in size and speed of change of the web in the last 3 years. . . classic economies of scale if there was ever one. . .

  9. I salute Snap for “blurring the lines,” as you say John. For commercial search, is there really such a thing as a pure organic result anymore? There aren’t many, in my opinion. Most of the results are dominated by the SEO industry. This isn’t said to bash SEO; they are playing by the rules that have been created. But the simple fact is commercial search results at GYM are dominated by the companies that have paid someone lots of $ to be there (either the search engine for paid links or an SEO for “organic” results). I think this current model is weakening the relevance of commercial search and I salute Snap for realizing this and trying to change it.

  10. Thanks John. It seems that I may be in the minority with this concern. Certainly only picking groups/sites you are interested in helps this quite a bit.

    I love my bloggers and want to make sure they make a good living and can continue to do what they do — and successful advertising campaigns no doubt help them do this. Likewise, one thing bloggers offer to me is honest takes and the freedom to “say it like they see it” and I want to make sure they keep doing this.

    I am not sure the answer to this issue, or what FM’s policy is. Techcrunch, for example, just gave “ho-hum” review to Google Notebook. Can they still do this, even if Google was paying them big money to help them in an ad campaign or as a sponsor? Does the fact that someone is paying you or is a sponsor ever even slightly impact your view of a site? Can you still remain impartial? I have never been asked to do something like this, so I do not have the diorect experience with it. Maybe it has no impact. Again, I do not have the answer, but it is an issue I feel compelled to raise. Thanks for considering my thoughts . . .

  11. I am use to using Google or Yahoo interface that at first snap appeared complicated due to the features. However, once on it for a while It’s pretty cool. Well, it might just be me.

  12. Well 6 months after this post snap still sucks.

    The preview javascript stuff actually takes longer to load a tiny preview (~4 seconds) so its actually quicker to open the remote link and ACTUALLY VISIT THE SITE.

    Presumably their 6 months worth or however long worth of caching has been pointless? Am not talking of previews of small unworthy sites either – huge ones like the BBC.

  13. How exactly does one promote a search engine? Did Snap just advertise on a bunch of blogs to become known — or did they try something else? Their keyboard search is kind of like uGuX.

Leave a Reply