Moving on from the Google appliance
The cessation of the Google ESA search appliance came as a surprise, but is just one more in a stream of acquisitions and disappearances over the last few years.It started in 2005 with the acquisition of Verity and Ultraseek by Autonomy. Then in 2008 Microsoft acquired FAST Search & Transfer to fast-track the development of better search functionality in SharePoint 2010. For a while FAST ESP was available as a stand-alone application but main-stream support ceased in 2013. Between 2010 and 2012 Exalead was acquired by Dassault, Endeca by Oracle, Vivisimo by IBM, Isys by Lexmark and Autonomy by HP. When a search vendor is acquired search managers get no advance notice. Whatever the acquiring company states at the outset the independent existence of the software soon vanishes, but usually the company provides support for the migration process.
If you read the pronouncements from search vendors you would be forgiven for thinking that all you have to do is unplug Google and plug in A.N.Other software and carry on as before. I was very disappointed to read a column in CIODive which totally failed to understand the situation. As Edwin Stauthamer has pointed out in some detail, search is not that simple! There are two important flaws in the ‘replug and play’ approach. The first is that any other software application will not be able to mimic the Google UI, APIs, security permissions, connectors and ranking algorithms and much else ‘out of the box’. It may get close but there is a lot of careful work to do, not just on the server side but above all else with users. This is where standard test/query collections are so important. Enterprise users repeat queries and expect to find (for example) the procedures for setting up a new project on the first page of results. If the result is further down then users start to have concerns.
The second flaw is that a like-for-like replacement may not be in the best interests of the organisation. Now might be a very good time to look at user and business requirements, especially given the range of appliance, cloud, hybrid, commercial and open-source products available. I have a list of 60 vendors on the Enterprise Search Book site and I’m sure that it is not comprehensive. Only last week I was contacted by a new enterprise search vendor in Romania.
Although the recent Findwise survey indicated that there is an increase in the number of organisations with enterprise search strategies my own experience suggests that this outcome is biased by the survey respondents being from organisations with reasonably mature search implementations. I wonder how many organisations with a Google appliance have a fully documented and agreed search strategy just in case Google exited the business? My new Search Migration Assessment service is designed to help organisations (and not just those with a Google appliance) understand where the gaps are in their search strategy. There will probably be a lot of ‘unknown unknowns’, to quote Donald Rumsfeld’s axiom.
No-one knows how many appliance replacement opportunities there are but my guess would have to be thousands worldwide. Although there will be many organisations who are now in a difficult position over the Google exit I think that it is a superb market opportunity for all search vendors, implementers and developers.