ThoughtSpot’s offering is very distinctive. They hang their hat on one aspect; natural-language searching. ThoughtSpot is a search-as-query Business Intelligence (BI) tool, meaning users type into a search box (think Google) the question they have, and ThoughtSpot, with the help of Artificial Intelligence (AI), picks what it perceives to be the best data for the job, and the best ways to display it. It’s perhaps no surprise that ThoughtSpot was created by ex-Google employees.
As a concept it is hugely exciting and feels like the future direction of BI, certainly from the aspect of end users. They have recently announced version 6 of the core platform and are frequently adding and updating features. This is very much a platform on the move.
If you know us at DSCallards, or have read any of our blogs before, you’ll know we’re big fans of Qlik. It’s one of our benchmarks and standards against which we like to compare other BI offerings. So, how does ThoughtSpot stack up?
The Power of AI
ThoughtSpot uses its AI capabilities to translate the natural-language queries input by users, into SQL and then displays a chart or other visualisation with data related to that query. The AI helps ThoughtSpot understand the intent behind the question and therefore, exactly which data is best suited to answering it.
All this should mean it is incredibly easy for end users to pick the platform up and start gaining insights straight away. Indeed, one of ThoughtSpot’s strengths is how quickly data is visualised following a query. In theory, anyone should be able to use ThoughtSpot. In practice however, it’s not quite as simple as that.
Out of Control
There are a couple of problems with giving control of which data and which visualisations to use to AI. Firstly, although AI is incredibly powerful there will, inevitably, be a significant and complex set up process which is bound to have teething problems. Although ThoughtSpot is sold as a ‘plug-and-play’ solution for end users, for IT departments it will be much more difficult.
Secondly, the SpotIQ AI Engine used in ThoughtSpot does not always understand queries. Herein lies the problem with natural-language technology; it isn’t quite there yet. Several of the biggest criticisms levelled at ThoughtSpot focus on a steep learning curve for users, especially when queries go beyond the simple and obvious. To get the most out of ThoughtSpot users need to learn advanced phrases and understand just how the system works.
This is not to say that this isn’t a problem for other BI platforms, including Qlik, which don’t place such an emphasis on natural-language queries. Training on the use of ThoughtSpot would be essential in any implementation.
Choosing the right BI platform for your organisation is always horses-for-courses. We like Qlik so much because it gives you such a vast and powerful toolbox. There is capability for almost every need you can think of.
In terms of pure BI capabilities, Qlik trumps ThoughtSpot at pretty much every metric you can think of, from data connection to analysis through to visualisation. That probably isn’t why you’d choose a solution like ThoughtSpot though. There is one major selling point and that is the AI-powered natural-language searching. If you think your people would benefit from this then ThoughtSpot is certainly worth exploring.
Go in with your eyes open though and know that the set-up process will be complex and potentially incredibly lengthy, and despite everything, sometimes natural-language queries aren’t as simple as they seem. Fundamentally, you are trading off analysis and visualisation capabilities for the natural-language search function. That is the key factor in the decision between Qlik and ThoughtSpot.