Google and Enterprise Search: The Eichner Vision
June 7, 2012
Google has a new head of enterprise search, Matt Eichner, Yale, Harvard, and Endeca. Computerworld UK ran an interesting article on on June 5, 2012 by Derek du Preez. “Matt Eichner: Bringing Google.com to the Enterprise” walks through what appears to be the game plan for the enterprise search unit for the next three or four months, maybe longer if Google generates more traction than it has in the previous year or two.
The article reports that Google “commands over 90 percent of the UK’s online search market.” Mr. Eichner allegedly said:
If you look at Google in the search space, we are taking that consumer expectation that we developed on Google.com and packaged both the user interface and the algorithms behind it into an enterprise appliance.
The GSA as the Google Search Appliance is presented has been available for about a decade. Based on chatter at conferences and opinions floated by assorted search experts, Google has placed upwards of 55,000 GSAs in organizations worldwide. Autonomy, by contrast, is alleged to have about 20,000 licensees of its search and content processing systems. Microsoft SharePoint, which includes a search system, is rumored to have more than 100 million licenses. It is difficult to know which enterprise search vendor has the most customers. The numbers are not audited, and each vendor in the enterprise search market tiptoes around how customers many customers are signed up, how many customers are paying their bills, how many customers are dropping licenses, and how much revenue flows to the vendor from enterprise search service and support. In short, it may be difficult to know how big any one vendor’s share of the enterprise search market is or if there is even a market for enterprise search in today’s mash up and fluid business environment.
A block diagram showing a GSA in an enterprise installation. Note the presence of “OneBox” units. Authorized Google partners may be needed to get this type of implementation up and running. If this is accurate, then Mr. Eichner’s assertion about an “out of the box” solution may require some fine tuning. Image source: DevX at http://www.devx.com/enterprise/Article/33372/1954
Google believes there is a market, however.
The pointy end of the spear for Google is its search appliance. The idea is that a customer can order an appliance and get it up and running quickly. The GSA can scale by plugging in more GSAs. The GSA understands “enterprise context”.
According to Computerworld’s write up, Mr. Eichner asserted:
At Google we have billions of queries from Google.com coming in every day that we are able to analyze and deliver an enterprise tool that balances human behavior and search relevance.
Google’s enterprise services are cognizant of big data, which most vendors suggest can be managed by their search system. Google is no exception. Mr. Eichner, according to Computerworld, observed:
Big data is in the eye of the beholder. If I gave you 500,000 documents, which doesn’t sound like a lot, and I said to you find something in there – you would look at me and say, ‘can I use a search engine?’ From your perspective, 500,000 would be big data. We often lose sight of that. Insight needs to be delivered when you have more data than you can process. This can come in the form of 500,000 documents or hundreds of millions of documents. The real mandate in the world today is to get up the competitive stack by being more knowledgeable about what you are doing more quickly – that’s the nature of the information economy. The imperative is to get better at assimilating the knowledge you have and acting on it. The inverse of this is if you have big data and you don’t have insight. That’s the equivalent of saying ‘I’ll take a guess, I won’t use the information and I’ll take a guess.”’
Computerworld’s story revealed that Google acknowledges the importance of open source search in the enterprise. The Computerworld story Mr. Eichner as saying:
Open source does have a role within enterprise search. If you have a lot of engineers and your goal is to adapt search to one specific application, then it works well. The Original Equipment Manufacturer (OEM) market is a classic example where companies have the internal engineering departments and competencies to manipulate the search outcome to be very specific to a single system – open source is a great for that. However, when you like the idea of universal search and searching across multiple systems, much like Google on the web, I think it falls down. It doesn’t have the search relevance capabilities and it requires expensive PhD expertise to establish how that relevance should be calculated.
Google’s view is that the Google Search Appliance arrives in one box, tuned, and ready to deploy immediately. Open source search has to be installed, configured, set up, tuned, and managed. The GSA does not require this type of baby sitting.
My take on this view of the GSA is that on premises search appliances are appropriate for some organizations. In other situations, solutions which do not impose taxi meter pricing and permit customization may be more appropriate. After years of marketing the GSA, cloud solutions, search based applications, and business intelligence systems which include a search function may be of more interest to some organizations.
With major search vendors selling out to large enterprise vendors, it is not clear what will happen to the market. Microsoft, for example, may be increasingly aggressive in making the Fast Search system available as a component in SharePoint or as a freebie when an organization signs up for Microsoft Office client access licenses. Oracle now owns Endeca (which Mr. Eichner worked as a senior vice president until becoming a Googler), TripleHop (remember that technology), SES11g, InQuira, RightNow’s Q Go system, and InQuira’s natural language processing system. These complement the widely available Oracle Text product. When it comes to competing with Google, Oracle may become increasingly aggressive. IBM offers its Lucene / Solr based systems. Hewlett Packard’s integration of its $10 billion Autonomy purchase is a wild card as well. (I discuss one facet of this deal in my forthcoming KMWorld column, which will appear in six or seven weeks and then on the Information Today Web site.) Add to these shifts in search the strong showings of Lucid Imagination and Polyspot with their respective open source search solutions. Google may not be able to dominate enterprise search in the way it dominates online advertising.
I learned that at one recent presentation Mr. Eichner was captured by two videographers and one sound technician. My hunch is that a YouTube video about the GSA and Mr. Eichner’s view of the enterprise search market will be available in the near future.
Stephen E Arnold, June 7, 2012
Sponsored by Polyspot
Comments
2 Responses to “Google and Enterprise Search: The Eichner Vision”
Mr Eichner’s comments on open source are rather misguided and pure FUD in my opinion. You don’t need a PhD to use an open source search engine, nor do you need lots of internal engineering talent. We’ve spent years implementing open source search for a myriad of applications and clients and the idea that it is only for a narrow set of targets is plainly wrong. It’s also important to realise that Google’s appliance gets *very* expensive at scale.
Google and Enterprise Search: The Eichner Vision : Beyond Search…
This article has been submitted to Intranet Lounge – Trackback from Intranet Lounge…