Review: Google Co-op

I’ve had some time to mull over what I saw at Google Press Day. Let’s review the products that were made available:

Google Co-op. See below for the review.
Google Trends. Pretty much everyone can see the appeal of this in a few seconds, and it’s fun to play with. For example, Marc Grobman noticed that people are more interested in “breakfast” on Saturday/Sunday, but “brunch” is more of a Sunday thing.
Google Desktop 4. I updated my ancient, crufty version of Google Desktop last night, and it’s nice. My favorite is the “ctrl-ctrl” shortcut to bring up a search box. I wanted to make it so that “shift-shift” toggles the sidebar and gadgets on and off, but couldn’t get that to work. The newest version is clearly more polished than my older version though.

Of the products, Google Co-op is one that I’m getting more excited about over time. A few webmasters had an initial reaction of “My head just exploded!” But several folks are digging into this more now. Barry and Detlev had an interesting conversation in the May 12th SearchCast about Barry’s implementation of a subscribe link. Barry had some really good comments about how this is self-regulating for spam. If people have to subscribe, it’s less likely to attract spam attempts. And if you overreach and try to stick useless labels on every search under the sun, it’s very easy for a subscriber to click “remove” to get rid of your labels.

Should Google Co-op make your head explode? I don’t think so. Personally, I believe that the core concept is pretty smart: it gives experts a way to augment and annotate Google’s search results. The simplest level of interaction with Google Co-op is to annotate pages, whether your field of expertise is TiVos or Kingdom of Loathing. Most people are an expert on something. 🙂

I was talking to Ramanathan Guha about Google Co-op the other day, and I realized that Co-op is much deeper than that though. For example, I talked to Marti Hearst last week at a dinner, and I was reminded of her group’s research about different ways to slice and dice data. One example is that a recipe could be categorized by ingredients, type of dish (appetizer vs. dessert), or occasion. See http://www.epicurious.com/recipes/find/browse for an example of that Hearst-ian way to access recipes in lots of ways instead of a strict hierarchy. Or this Nobel Prize winner page is another example. In Co-op, the ways to slice the data are called facets, so users aren’t constrained to a strict hierarchy to find information.

I’m still poking into Co-op, but I’m excited about the ability to merge different people’s views that share the same label. I also noticed this intriguing entry in the Topics Developer Guide:

ELIMINATE: Results with this label will not appear among the results.

I checked with Guha and the functionality sounds cool: if I had a “webspam” label with the ELIMINATE mode, then I could share a blacklist. Anyone could subscribe to me to share that blacklist. And then anyone could make their own file and just use the same label with the same semantics. Eventually, you might be able to subscribe to 9-10 people who kept a list of domains that they wanted to remove from Google’s results. Shame on me for going straight to webspam when most people will want to augment and annotate Google search rather than remove results, but it’s neat that Co-op has this functionality built in. 🙂

I’m still digging into Co-op, but I’m liking what I see so far. It gives a richer vocabulary to talk about things than just links. So a site such as Quackwatch could produce some really interesting ways of labeling sites. Subscribing is easy and unsubscribing is lightweight. So there’s not too much hassle to subscribing to quite a few people. I’m looking forward to digging into Co-op more over time. 🙂

45 Responses to Review: Google Co-op (Leave a comment)

  1. Co-op is Google + Wikipedia. You guys should allow more than experts in though; Wisdom of Masses is key. I don’t get the subscription model thing in the previous comment.

  2. Somebody looking for some link love should put together ‘google co-op for dummies’ with some really basic easy to follow instructions with multiple examples. No offense to the implementation team but after looking at the instructions it’s something only a programmer could love.

  3. graywolf is sooo right here, but do not expect geeks to notice that the pencil holders in their top pockets are very uncool.

    look a froogle and it’s primative interface, how many people are not using it because of that?

    wiki is cool because anyone who looks at it get’s it, period.

    and not to be too negative but “trends” are right hear in our brains, do I need “trends” to tell me that “people are more interested in “breakfast” on Saturday/Sunday, but “brunch” is more of a Sunday thing”? I do if I am autistic like the majority of the folks in silicon valley!

    sorry, but all this change is bothering me and the lack of clear guidelines is really getting beyond annoying.

  4. Supplemental Challenged

    Co-op unfortunately is basically a disaster in regards to its stated aims. The instructions are at best incoherent.

    Bottom line, the only people who will devote themselves to finish doing the Co-op tasks are spammers. Period. There is no chance whatsoever that any “normal” person will be able to understand garbled, incomprehensiblely babbling text.

  5. Could someone explain to me more what is that Google Co op?

    I didn’t understand

  6. Hey Matt, you mention the Kingdom of Loathing in you blog??? So an adventurer is you? Ore you in a clan? Level 12 Spaghetti Spellbinder slarty_de would be proud to share the same clan with you… 😉

  7. coop looks very interesting but ime not sure if the implications have been thought through – and i need to properly look at it.

    whats to stop me atacking my competitors or a politcal group trying to manipulate the serps to remove liberal/republcan content – this of course would be done under the guise of “protecting da kids from pron”.

    You do know wikipedia is used as part of a SEO’s tool kit sometimes with ligitimate uses and soetimes not.

    Rgds

    ps Can’t remeber if we put google.coop on the reserved list when .coop was launced

  8. Google trend is so nice! I like it, but as SEARCH ENGINES WEB i think it misses approximate amount of searches.

    It works quite well with French searches but somehow, when searches are too qualified (eg. key phrase with 3 words, that are actually used by internet users), I can find any results. Does it means there is a minimum amount of searches necessary to appear in Google trend?

  9. Not really had the chance to look too closley at Google Co-op but the idea seems spookily similar to Squidoo started by Seth Godin among others. As is stands http://www.squidoo.com is doing a pretty good job of the expert idea with it’s lenses, hey it even uses Adwords as a plug in.

    The difference at the moment is Squidoo is run by live people can Google mange the same trick with machines instead?

  10. Matt, maybe i’m missing it in the spec… but does it offer me the ability to hit an external script to tell it it’s data?

    That is.. instead of having an external file with all my tags in it, I’d like to have it call a php file that just creates 1 dataobject tag on the fly, generated from whatever was typed into the query box..

    I could do some really cool and useful things with that..

    Otherwise, it’s like giving somebody an xml copy of my site’s database… and that’s not something I’m willing to do just yet.

    If i missed it in the spec, could somebody point me towards it?

  11. Ryan, Barry had done some neat things like that, but I think he did a cron job to generate the fresh XML file and then re-uploads that.

    10080:BTG174, the biggest thing that keeps that from happening is that I have to trust you enough to subscribe to you before I see any of your proposed labels or changes.

    Franck, a simple example might be UCSF deciding to annoted thousands of urls that they consider helpful in the medical field. Check out http://www.google.com/coop/profile?user=015903972350817203414
    for example. Google Co-op isn’t for every individual user, but there are certainly a lot of organizations out there who may want to be able to label the web with data from their perspective.

    graywolf, if you think that other folks would need help to set up their annotations, that smells like an opportunity to me. 🙂

  12. Not that I have the best understanding of Google Co-op either, but here’s my take:

    To address ‘10080:BTG174’ (nice name), so called experts in various fields create lists of pages that people have to subscribe to in order to view the results on the Google SERPs. It’s not clear at this time weather those sites are effected in the organic results.

    A user would subscribe to a list created by an individual or organization that they trusted to provide quality links. In order for these subscriptions to display in the SERPs, the user must execute a query that matches the ‘tag’ describing the page added by the list provider. These results are them displayed above the organic results in a section marked ‘Subscribed Links’. Note the coloring of this section looks very similar to the ‘Sponsored Results’ section.

    Needless to say with Google getting their hands in to so many things these days, it’s hard to believe their not using this data on a higher level. If multiple trusted list creators are adding the same pages with similar descriptions, maybe that site should be moved up in the organic results? Just a though.

    Anyway, that’s my take on it.

  13. Yeah Matt, I was thinking of something more involved than a cron job..

    for example, 90% of my search traffic to noslang.com is for stuff like “define lol” or “acronym wtf” etc..

    so it’d be awesome if I could make something that pulled the definition they’re looking for out, defined it, and showed my link to where they can get more information.

    I can do it now, but with thousands of terms in the DB, hard coding them in just doesn’t make sense. it a.) makes a big file that constantly needs updating and b.) says “psst.. hey here’s my data, come take it”

    how about a slang: operator that points to my site? come on, you know you wanna… 🙂

  14. I’ve been playing with the co-op today, and I reiterate what’s already been said – the instructions are *very* poor. But it’s not the first time that I’ve seen very poor instructions when Google has launched something new. The people there may be exceptionally good technologically, but they are no good as instructive writers.

    I created a topic to ‘play’ with, and I discovered a few things. One is that, unless a page makes it into the “normal” results set for the searchterm, it can’t be included in the topic’s results. This limits co-op somewhat.

    Another is that it is very easy to spam the hell out of a topic, and if a topic becomes popular, then the ‘owner’ is likely to find it a full time job despamming and maintaining it. I’ve based that assumption on the fact that anyone can add to anyone else’s topic, and that the additions are added to the annotations file, which can be downloaded by the topic owner, edited, and uploaded again. I’d need 2 accounts to test that, but if I’m right, then I can see popular topics being abandoned. Of course, subscribers will unsubscribe as the topic gets worse and unmaintained, so it means that people will be able to successfully attack any topic, and cause its demise. Mass spamming may be dealt with by the need for a Google account, but steady, persistant spamming may succeed. I suspect there will be an awful lot of abandoned topics.

  15. PhilC, I think that they’re doing a new push soon that improves the documentation listed.

  16. Hobby site worry wart

    I sincerely hope that Google Coop will include a statement of a “Code of Ethics”.

    I can see where a group of dedicated zealots could set about to use Coop as a method of destroying what they perceive to be their “competition”. What is to prevent such nonsense?

    BTW – If I have a modest hobby site that provides references (and links) to sites or pages I feel are worthwhile for my audience, how is this different than what generating a coop topic would achieve? (Other than the tremedous overhead coop seems to require).

    – Concerned, part time voluteer for a worth while hobby site

  17. You are free to make a topic with the same name as mine. Nobody subscribes to you, nobody cares. You are free to create annotations with the same labels as I use. Nobody subscribes to you, nobody cares.

    If you find a way to affect anything my subscribers see without first having to get them to subscribe to you, please point this out in detail.

    I think the poor documentation has led to people spending time imagining how the system might work instead of actually using it to see exactly how it does work. IMHO :-).

    Of course, the existence of “special” topics that Google is working on that get treated differently in some ways probably adds to the confusion.

  18. I have wasted so much time on trying to figure this out. Surely it would have been so much easier if the documentation was more specific?? Anyway I can confirm that you can label a site however you like it will not appear in a context you make to match the label. That’s if your site is in the sandbox. In other words if you are not in the mysterious exclusive “club” of unsandboxed sites DONT EVEN THINK ABOUT participating in Co-op. Wait 1 month, 2 months, 3 months ……….. 1 year, 9 months? Who knows what you must do – you are excluded until Google god decides you have joined the unsandboxed club……………… Google must just continue with its STALE content – or new content at the whim of unsandboxed club members. The sandbox is now the official barrier to entry – so much for my post on my site http://www.indigoguide.com/coop-user.htm

    I see that google adsense scrapper sites are added to the co-op directory – they don’t promote co-op at all with links to co-op – maybe Google has become a “look after your cronies” company or is fast becoming that way.

    Its sad to say but the collateral damage caused by Google’s fight on spam has stalled development on the internet by 18 months or more.

  19. Please, oh please do not invest the mythical “sandbox” with even more imaginary powers. Let’s construct a counter example to show that “sandboxed” sites are not singled out for any nefarious Google Co-op punishment.

    dmoz.org is not sandboxed. Here’s a PR6 resource it contains that clearly ranks well for lots of things: “/Computers/Software/Backup/”.

    Now, that page contains the word “abc”. Suppose I want to construct a Google Co-op refinement such that when you search for “abc”, that page appears at the top of the results. Can’t do it, even if I give that URL a label, and name that label in a element, and use a mode of “FILTER”.

    Here’s what’s really going on: labelling only lets you tinker with existing search results. It doesn’t let you add new URLs into search results. It just so happens that even though the URL I’m talking about has a PR6 and contains the word “abc”, it does not show up in the visible SERPs for the word “abc” (hey, that isn’t really the point of that resource’s text, so that’s an understandable ranking on Google’s part).

    So, a user searches for “abc”, and my lovely dmoz.org URL is nowhere in the results: THAT MEANS NO LABELLING OF THAT URL IS GOING TO ADD IT TO THE RESULTS.

    Labels are for rearranging the order of results in a Google search query, and you can also block results in that query from appearing at all, but they don’t let you bring in new URLs that were never a part of the original results from that query.

    I could still use a to add to the original search terms until I get a query that DOES include my favorite dmoz.org URL. For example, I could use a to add the terms “Computers software backup” to the original query of “abc”. By gum, THAT produces a search whose results contain my favorite page — but of course the user was originally searching for “abc” and now I’ll be delivering a bunch of pages that seem to mainly be about a backup software product whose name includes the word “abc”.

    The fact is, Google disagrees with me that my favorite URL is relevant for the search term “abc”. And if Google doesn’t think the pages in your website are very relevant for your favorite search queries, then it’s true that you’re going to have a hard time using Google Co-op to construct search refinements that elevate your website into the results.

    Note that you can still FORCE your website to appear by using a stilted , or heck, you can use the element to just send them straight to your favorite page on your website. It’s just likely going to be obvious to your subscriber what you’re up to, and they are perhaps more likely to stop subscribing to your results.

    I’m going to go out on a limb and say this is probably a good feature of how Co-op works. If you believe you’re “sandboxed”, Co-op isn’t going to help you. Better to focus instead on the real problem, which is that Google has a low opinion of the relevance of your website. Life is hard but remember: this is just computer stuff — it ain’t cancer.

  20. Death to all webforms that mindlessly eat less than and greater than signs! To fix my previous post,

    “I could still use a ” [Rewrite element] “to add to the original search terms”

    “I could use a ” [Rewrite element] ” to add the terms”

    “Note that you can still FORCE your website to appear by using a stilted ” [Rewrite element] “, or heck, you can use the” [Redirect] ” element to just send them straight to your favorite page on your website.

  21. Found a solution – which allows good results which are all very relevant.

    I agree it was a bad idea to bring up the sandbox – just pure frustration on my part which is not a really good excuse I know. G is doing its best to tackle the situation as fairly as it is able – I respect them for this and would challenge anyone to be able to do a better job at it.

  22. Google Co-op needs a better front end. Context/annotation seems to provide functionality similar to del.icio.us, but where’s the “post to Google co-op” button? It should be that easy, but it isn’t.

  23. still use a to add to the original search terms until I get a query that DOES include my favorite dmoz.org URL. For example, I could use a to add the terms “Computers software backup” to the original query of “abc”. By gum, THAT produces a search whose results contain my favorite page — but of course the user was originally searching for “abc” and now I’ll be delivering a bunch of pages that seem to mainly be about a backup software product whose name includes the word “abc”.

  24. I’ve read that the BigDaddy will ban you from the Google listings if it see’s more than 50 or so new In-Bound Links (IBLs, In-Links, back-links) appear in a month even if they are links from articles.
    This doesn’t seem fair! I’m sure I remember reading that Google encourages the posting of articles, however, now I’m led to believe that if I write 10 really informative and well formed articles and use a service to distribute them, I will certainly get banned!
    Perhaps you would like to comment on this?

  25. I’ve read that lots of article those are saying that will ban you from the Google listings if it see’s more than 50 or so new In-Bound Links, but I am still finding the such kind or links pages with good ranking. what is that?

  26. Granted Google Co-op can not be easily understood but with the proper application your ROI is multiplied.

  27. What about Google Analytics? I registered months ago and haven’t seen it working yet. I use Urchin, but it’s a comercial version that my host provides – not the google version. I think Google should cut back on its new-feature-a-week strategy (or is it daily now) and get some the really cool things working, like Urchin and the Link command. Otherwise you get bad press. I author books, and my new SEO book verson 9 is bad-mouthing Google Analytics because I simple report what is.

  28. I had a look at Google co-op. It is a great concept but the documentation is not so user friendly. And definitely Google should do something to prevent it from being used by SEOs as their tool kit. It would otherwise defeat the purpose for which it was created.

  29. Matt

    If you ever get round to investingation on Co-Op this may help you
    http://www.indigoguide.com/implementcoop.htm
    I have used background labels to delete or promote search results which is really powerful stuff.

  30. Matt,

    I’ve expressed some opinions about Google Co-op in a recent blog post. If you get a moment I’d like to know your thoughts and if I made any errors describing the Co-op:

    http://www.roboburp.com/search-blog/8/under-the-radar-google-social-search/

    Robert

  31. Well, Google has already been taking stringent measures in identify spam by the use of robots; so it should not be a surprise to have similar features built in Google Co-Op.

  32. About the co-op, in particular the health labels. Who decided that WEBMD’s labels to their own content should go to the top of the results? I understand the HON code and even the dot-orgs and not that I think dot-gov sites should be promoted (because we don’t want to be restricted to only what the FDA has approved), but it’s still better than a for BIG PROFIT site like WEBMD…really, WEBMD and Mayo Clinic — they are laughing all the way to the bank. It costs a fortune for the sponsors content to appear on WEBMD and Mayo …but that’s where the content comes from. Hope they’re paying a hefty price for the special treatment.

  33. I agree with comments from graywolf. Sometimes Google over complicating things, for instance installing the google sitemap. It’s so easy, but google does it opposite. I think google needs to create it’s own online tool that able to generate XML sitemaps. Do you agree?

  34. Hi Matt,
    I have been testing Coop now for 1 week, with my “SEO Search Engine”:

    http://www.seoking.eu/seo-search-engine/

    People seem to love it. The Engine has good traffic, and lots of requests. I use it myself and administer it as well as I can, to boost the quality of the SERPs.
    For just having this one set up for fun and testing…It´s quit good. I gonna keep and work with it. Thanks for this opportunity!

  35. Here’s a simple implementation of Google Co-op, http://www.gootube.net
    By Customizing Google Search, GooTube was able to provide a Search Engine exclusively for videos. The results can be refined by different sources or video hosting sites.

  36. google coop is superb. It will be the dominant search method in the future.

    I created a few of my own, and it works fine.
    There’s also a Directory for coop engines at http://www.coopdir.com

  37. So basically it’s Google’s encyclopedia……

  38. Google desktop is cool feature but it takes a long time to index the harddrive.

  39. Hi,

    Is this true that using Google Co-op pushes your results to Google’s supplemental index?

    I had launched my website on 19 Dec 2006. It performed well on Google, rather I would say the performance was increasing day by day. I needed the site search software, I came across Google C-op, I tested it and then after repetitive successful testing I incorporated it on my website. It worked fine for 4-5 days and just went off the 6th day. Only the home URL was displayed. After much research I discovered that with in a month all my site’s URL have been pushed to Google’s supplemental index. I need to depend much on Off-line adverts and should forget about Google’s main index listing.

    I found out from more resources that this happens when you use Google Co-op.

    Any suggestion would help. The URL is http://www.gurgaonshoppingmalls.com

    thank.

  40. When Google first launched Co-Op I thought it was pretty raw. Having to write one’s own XML by hand wasn’t my idea of great stuff for the masses. But I recently went back and saw that they had made the interface much more user friendly. They are still missing a large number of features I’d like to see, but I’ve been working on putting a large directory of mine (around 1600 links) onto it and it seems to be working pretty well thus far. You can see a rough sketchup of it at http://www.civilwarsearch.com if you want (I’m using DotNetNuke for the page and having some difficulties with the way it wants to render, so its a little choppy).

  41. Fantastic product. My head exploded as well 🙂 But Google Co-op doesn’t support national characters, and for a Norwegian developer like myself, Google Go-op instantly deteriorates from excellent to virtually useless.

  42. You can use the following link to search for Google co-op sites
    http://www.google.com/coop/cse?cx=004698572784507413576%3Aioxenquztmi

  43. This is a great feature. Does it generate income for the webmaster?

  44. I found it really easy to get into using Co-Op. I set up http://www.boatr.org.uk in a matter of hours to search 200+ boating websites.

    I’ve started using it on a website where I need to search across two domains, a phpBB discussion forum and a WordPress blog and didn’t want several different search boxes.

    It’s a really useful tool. Simple to use. Simple to incorporate in a website.

  45. Google had this coop but now they have integrated it with Adsense but there is one bug which I have noticed. When you submit the search the results page is not pre-populated with the search term now but it used to when it ws not with Adsense.

css.php