Google Cloud Search – a seriously interesting search application
When the news broke back in February 2016 that Google was going to exit from the GSA search appliance business there was considerable concern that there did not seem to be a migration path to another Google application. A cloud search application was promised but when it did emerge it was restricted to Google G-Suite.
At the Google Cloud Next 18 event in San Francisco earlier this week an enhanced version of Cloud Search was announced that offers federated search across multiple repositories. The Group Product Manager is Jayanth Mysore, who worked on Google Maps and Google Analytics from 2007-2015 before a year working in India, coming back in January 2017 to head up the Cloud Search project. The video of the presentation is up on You Tube and is a very good way of spending 42 minutes. There are a couple of interesting comments at the outset. The first was that Google do not regard ‘knowledge workers’ as a defined cohort; instead they see all employees as knowledge workers. The second is that people want to search from inside the application they are using, not open up another application in another window. That is a very good insight into information search behaviour in the enterprise. There were also some references to ‘immersive people search’ but no details were given.
The initial impressions were very positive, and I had a sense that the presenters from Colgate-Palmolive and Whirlpool were genuinely impressed not just by the functionality/performance but also by the fact that real people from Google turned up on site to monitor progress and to listen to where improvements could be made. Google has certainly appreciated that for a product of this complexity there is a need to work with partners who understand how search is used within organisations. The statement was made that Google aim to be ecosystem-led in exploiting Cloud Search. Applause all round for that. All you had with the GSA was an email address. Already a good range of partners is in place, though mostly in the USA. In Europe there is Wabion (Germany/Switzerland), Accenture (who of course own Search Technologies) and Lumapps (France).
One of the interesting comments made by Whirlpool on the launch video is that with the GSA it took 6 weeks to index 12 million documents from 21 applications. With Google Cloud Search it took just a week. The short demonstrations also gave an indication that Google Cloud Search with present the results from a federated search integrated within a single results page. The demonstration by Lumapps was of an image search application, which seemed to work ridiculously quickly. Faceting seems to be dynamically built, but on short demos in front of a live audience it is not easy to work out what is real and what is smart promo.
There are two packages at present. One is for Google G-Suite customers with more than 5000 users. The other is a standalone package. This is especially interesting as it would seem that you could just by a standalone package without being a Google customer. That was certainly the indication given by Jayanth Mysore on the video. No price points were mentioned but it is clearly based on content volume, and not seats or servers.
The winners are clearly going to be the search integrators. A very high functionality together with very high performance is just want they need, and AWS, Azure and other cloud service vendors are going to have to consider whether they take Google head-on or just be extra-special nice to their existing enterprise customers. Much will depend on the pricing structure – many GSA customers never quite realised how much a GSA could cost as they had no idea of how many content items it would find on the initial index. Many organisations have a concern about cloud search and may well wish to stick with an on-prem solution, but that can surely only be a short-term strategy. Other main-stream search vendors are going to have to sharpen up their marketing. Elastic may want to rethink what it is going to do with Swiftype. If I had shares in MicroFocus (IDOL) it might be a good time to consider what to do with them
The main benefit is going to be that enterprise search gets fresh visibility. If I have a concern is that it doesn’t really matter how much AI and ML you are going to unlease and how easy the implementation is. You are still going to need a search centre of excellence to meet the very high levels of anticipation there will be for Google inside the enterprise. If they are achieved then search will become the primary integration platform for next-generation digital workplaces. I’m so pleased I’m in the search business!