Last week Google told its reseller partners that the Google Enterprise Search Appliance (ESA) was going to be withdrawn from the market

ESA three-year license terms will no longer be available and only one to two year renewal options will be offered to existing customers. From 2017 there will be no new hardware sales. Google will still offer support for bug fixes, but there will be no further technical development. 

No one knows what the installed base of ESA units is, but it must be in the thousands globally or Google would have pulled the plug before now.

These customers were likely attracted by the plug-and-search capability of a search appliance. 

At the time of its release in 2002, Google put a significant amount of effort into persuading organizations that all installation required was putting the servers into a rack, pointing them in the repositories' direction and providing employees a user interface with the Google logo. Without doubt, organizations could quickly install ESA and needed little in the way of technical support. The Google ESA provided organizations around the world with a search application that needed no staff to manage it. 

Now the ESA has no future and customers may well have no staff with the skills to make decisions in fairly short order. Many customers will be unaware of what commercial and open source applications are available and how best to match them to their probably poorly defined requirements.

Marketing Ahead of Engineering

The Real Story Group commented a few years ago that, in the case of the Google ESA, marketing was a couple of steps ahead of engineering. Managing security trimming and connecting to anything that did not have an HTML interface was very tricky and tuning the ESA to improve ranking and overall performance was challenging, to say the least. 

Although resellers did their best to support their customers, they had no direct line into Google and direct customer support was nonexistent. It also did not take long for customers to realize the difficulties in calculating the total cost of ownership with certainty due to Google's convoluted methods of counting "content items" for the purposes of setting the license cost.

After 13 years of providing search in a box, Google has made it clear that the future of enterprise search is in the cloud. Google stated that the forthcoming solution has been in limited beta for some time. Google asks anyone interested to provide an email address (of course) to be alerted to developments. 

“This definitely seals the end of the era of commoditized search. It also highlights the fact that web search and enterprise search are a completely different ball game. With this news, existing GSA customers, especially large global organizations ... have no other options than to look at an alternative solution,” said Laurent Fanichet, VP Marketing at enterprise and big data search provider Sinequa.

With any other company questions would be raised about why ESA was pulled before a clear transition was in place, but this is not unusual with Google. Do you still have your Google Glasses? Nevertheless, it's not going to be easy for the IT Director to explain to the Board that they have to invest in a new search application. 

Miles Kehoe, search practice director at Avalon Consulting LLC, outlines what you should be doing this week below.

Actions and Decisions

Remember that when the license runs out, the machine stops. Full stop. You own nothing apart from some nice Dell servers. 

The best place to start is by reviewing your search strategy, but sadly few Google customers thought it necessary to have one. Take a look at Martin White's A-Z Search Strategy Checklist for an indication of what needs to be in one.

Depending on when your current ESA license expires, you could have from 12 to 24 months before your search servers become bright yellow room warmers. That is not a lot of time, so get started now. I'd strongly recommend you work with a vendor-neutral integrator that has a strong search practice combined with a knowledge of the wide range of proprietary and open source platforms. Be aware that a large number of organizations are in the same position and they are all going to be hunting for a solution at the same time.

Review your requirements, talk with representative users, and determine whether you should add additional functionality. The Google ESA included a number of good features, but use this as an opportunity to review what capabilities you may have been missing in the "box" solution.

Once you’ve narrowed your potential replacement platform to one or two candidates, consider a side-by-side evaluation. Your partner should be able to help in both narrowing the platforms to evaluate, and then setting up and running the comparison. Test everything: indexing and index coverage, security and the user search experience.

Whatever the option, the clock is now ticking because it may well take six to nine months to define requirements, make a selection, find the team, install and then manage the delivery. And that would be fast track timing.

Finally, keep your users up to date on the process: why it's happening, and what they can expect when. Good luck!

Title image by Kai Oberhäuser