Tuesday 18 August 2009

Why is it not like Google

Recently, one of our members posted the following on our KIN-only discussion forum. With his kind permission, I am re-posting it here.

Google Appliance as shown at RSA Expo 2008 in ...Image via Wikipedia


Here's one for a Friday afternoon. I was having a drink with some friends the other day and one was asking me why can't enterprise search be like Google. I started rattling off a bunch of reasons and thought that perhaps I ought to capture these thoughts so other's asking the question (and I am sure there are some) could see the answers. When I wrote them down, it seemed like a bunch of excuses so I have tried to add some learning points so that anyone reading it might feel a little more informed how to do things better.

I don't suppose that this is uncommon in our organisations and thought that together we might be able to contribute more/better points and then have them available to all.

So, here is my starter for 10 (well a bit more, I have already had a few comments and some is prompted from similar articles - of couse found through Google ).

Of course, what I would really like to be talking about is findability, no one wants to search, just find - that is a whole other debate.

Regards,
Andy


1. Search Engine Optimisation
Many millions, if not billions of pounds are spent by marketing departments ensuring that their service or product appears in the first few results. In the enterprise, the attitude of publishers is that once they have 'put it on the intranet' the job is done. There is little thought on how the user might then find that information that has probably had quite a lot of effort spent in creating it. If only a fraction of the effort spent on the creation was put to the task of ensuring the content was findable (through many, many routes, not simply search) then the intranet would work better.
Learning point: Publish the rules by which the search engine works (link to go here). Encourage people to 'game' the results. Teach them that the value is in people using their content, not simply creating it. There is perhaps also a role for social or enterprise bookmarking tools - although not strictly search, these tools provide users with another means of finding material, in a way that isn't dependent on content publishers. As I sometimes say when delivering our portal maintenance courses, "don't put up with being a putter upperer" make sure you are given the opprotunity to ensure your site and content does the job not simply exists.

2. Advertising.
In all the years that I have been involved in enterprise search no-one has offered me even a penny to ensure that their content is top of the search results list :-(. Compare that to the millions/billions spent on advertising with Google. Remember that the advertised results are not just the ones on the Right Hand side, in many cases the first few 'normal' results are also there through advertising.
Learning Point: Consider how you can put in place a process (and promote it) to enable people to have 'best bets' so that their important content is very explicitly linked to prominent results.

3. Enterprise search is based around document silos and a number of fairly independent web sites.
Google has got to the lead in the internet search engine space partly due to its patented PageRank algorithm. This algorithm uses the fact the web sites link to other web sites and those that are the most useful have most links to them. Google is effectively harnessing the power of us, the users, and our intellect to determine the best links. Although this is a little simplified it is mainly true and does not reflect the typical intranet in an enterprise.
Learning point: ????Perhaps some things are just different.

4. THE one or ANY one?
In many cases when looking for content in Google the user just needs a web site for their holiday, DVD, opinion etc. In the intranet it is one single specific instance of a thing that they have in mind and nothing else will do.
Learning point: ?????Again, perhaps some things are just different

5. Missing Content.
Once our intranet search got past the stage of people feedback little niggles about features, the single largest reason for people clicking the 'Dissatisfied - help us improve' link was for content that did not exist. Even the best search engine in the world cannot create links to content that is not there.
Learning Point: After all else, a real person with a very good knowledge of the organisation is best placed to put the requester in touch with the likely sources of information. In this regard an enterprise is in a much better place to react than 'the whole world'. Again, there may be a place here for some of the more "Web 2.0" tools, allowing users to find not just content but also individuals with relevant expertise, or communities with related interests - as always this would require both good tools and a significant change in behaviours!

No comments: