Living a Google Life

p. Several weeks back I attended the Google Seminar in a box. I was not the target demo, being a one man shop with a savvy I/T department but it was interesting nonetheless. p. One of the big take-aways was the insight that everything Google designs is around search. I know, Duh!. But seriously: everything they do is architected, designed, reliant on search and reliant on the assumption that you'll enter the right search terms, and that Google or the Google mini device you've implanted in your intranet will return the right results. p. Coming from a background where we spent days, weeks, years agonizing over site navigation and taxonomy for our internet and intranet sites, seeing an organization rely entirely on the quality of search as the way to access information (not just documents: applications, non-traditional-web documents, CRM crap, etc) just was sort of mind-blowing. p. It assumes two things: * First: that the documents relevant documents are written well enough, written for SEO basically, and have SEO relevant URLs. * Second: it assumes that your users are search savvy and know how to enter the right terms and know how to refine a search. p. Now, I've been out of enterprise web operations for nearly six years and imagine (and hope) that there's been much improvement in the technology and that people have become more skilled, more knowledgeable in their use of search, but I'm guessing that there's been far more advancements in the former than the latter. p. At IBM we had a terrible series of search engines we had to use. The software itself was not necessarily bad, but the process of pulling together and determining who designed, who funded (always important) and who ran the search system was just a complete nightmare. We were never happy with the results at www.ibm.com but we were also pretty powerless in the great corporate scheme of things to actually pull together something we could be proud of. p. One of the activities we did after yet another high-profile failure of the search system was to go through and analyze the search queries. My colleagues Alex Wright and Jean Mountford determined that most people were searching on very simple terms (aptiva,thinkpad,os/2 drivers) and eventually that people just wanted to get to the relevant home pages. p. They didn't want to sift through our many, many pages of results. p. They also signaled that they weren't finding these pages through our site navigation, which we fed into planning for the so-called Bullseye redesign in 1999. p. But back to search: search was a dismal failure and the corporate bureaucracy at IBM pretty much ensured that it would not be resolved soon. Alex and Jean set about building a sort of pre-search engine. p. We were required (under various penalties, usually endless conference calls) to use the search engine we were told to use, but due to its frequent failures had managed to set up a proxy service. So you'd enter your terms into www.ibm.com/search and a script on the servers would proxy out a call to the search engine d'jour. p. The optimization we added was this: we knew that a vast majority of the search terms were simple and that the users just wanted to get the damn site for the term. We built a simple keyword driven database (eventually adding even handy thumbnail graphics of the relevant products) so that if you searched on thinkpad you got back as the first result, always as the first result an entry for the Thinkpad homepage including links to product, services, support and sales (if I recall correctly, it's been a long time). p. For those who wanted more information, who wanted complex queries and the resulting complex results, the proxy returned the results from the real search engine. But for most people, the keyword engine was completely satisfactory. Our search related complaints dropped dramatically. p. There were two lessons here: first, people thought they’d have a better chance at finding the information they were looking for by going directly to the search engine and clicking on the first result. Second: these same people don’t really know anything about using search engines (hence the whole field of search engine optimization). p. So, back to Google’s approach: I question whether this is a smart thing for an organization to do, relying entirely on search to retrieve documents and information necessary for day to day business. It may work within Google where everyone has at least ten Ph.D.s (I jest, I know it’s dropped to three to get in the door), but for a small-medium organization to switch to a Google Mini and rely on it entirely for information retrieval (today search, tomorrow Google Apps for your domain)? I just wonder if that’s realistic, has the collective intelligence about searching improved that dramatically in the past few years? Can you safely rely on Google’s ranking algorithms to return the information you need within your business? p. I don’t know. I don’t really have the resources to even try to know, but for the next month I’m going to try a bit of an experiment on myself. p. I’m a one man shop, I have a few active clients, and a relatively simple schedule for the next few months. I don’t have SEC or other regulations to worry about. I’m not presently under litigation (I hope). p. So, for the next month, into January, I’m going all-Google. I’m forwarding the mess of email addresses I have to a Gmail account. I’ve stopped “logging out” of Google and I’m letting the Search engine track my every move. I’m copying calendar entries into Google Calendar and I’m toying with Google Reader (though I find it frustrating that I can’t just punt and mark everything read). p. I don’t know that I will learn anything of real value or not, my needs are not so complex that I need to rely on search regularly, but I do find myself searching my mail regularly and Outlook is a bit of a nightmare (I made what I though was a smart move to archive mail into separate files by year, however Outlook XP “improved” search by removing the feature to search across multiple PST files).

Posted in Metadiscourse