BI Trends

Why SaaS Business Intelligence (BI) Has Failed

By Steven Schneider
Share on LinkedIn Tweet about this on Twitter Share on Facebook

The promise of SaaS Business Intelligence (BI) has been talked about for almost 10 years now, but has yet to establish itself. I still remember back in 2006 reading articles that went on and on about how companies like Lucid Era (oops), SeaTab Software (anyone seen pivot link actually win a deal recently?) and Oco (gobbled up by a services company) were going to change the world.

They didn’t – and now we have a new batch of companies saying much the same thing. Most prolific is GoodData, led by serial successful entrepreneur Roman Stanek (someone whose passion and accomplishments I admire). To hear him talk you would think that on-premise BI is dead and that his company will be going public sometime this year (an exaggeration).

There were many problems in 2006 that kept SaaS BI from going mainstream, and while many of those challenges no longer exist (like cheap infrastructure – I love you Amazon), some still do.  The greatest of those challenges, in my humble opinion, is that of data. I refer to it as “The Data Problem.”

The “Data Problem” speaks to how difficult it is to:

  • Gain access to the data where it resides
  • Transform it to a format suitable for reporting
  • Sync it with the source system/location on a regular basis so that it stays up to date
  • Creating meaningful output, with aggregations, metrics, measures, and all of the things business people like to look at.

The problem today is that most BI solves these challenges with a set of tools and a lot of hours of technical people’s time.  Here is a typical conversation that I can envision with a SaaS BI provider:

  • Prospect: I want SaaS BI because it is inexpensive, fast, and I can pay for it over time. Do you have that?
  • SaaS Vendor: Absolutely. Let’s start by doing a two week scoping effort on your data, we can then present you a statement of work to build the ETL scripts, set up the synching, and build the front-end (the front end is the easy part by the way).  Oh, and you can’t change anything in the source data once we’ve done that because otherwise it will break.
  • Prospect: Uhh… huh? What happened to inexpensive, fast, and easy?

The most common way for SaaS Vendors to get over this hurdle is by choosing specific applications, like Salesforce, and building connectors that handle a lot of this for the prospect.  This is a great approach, and really the best way that I can see.  However, this really limits the scope of what people can do.  Now these become “analytic applications” that address specific needs for specific people in organizations.  It’s really hard to build a big company doing just one of these analytic applications, so you have to build lots of them (or partner) and develop expertise in a whole range of areas.

I’m personally a big fan of SaaS BI, and do think its time will come. However, for most use-cases that I see today and for broad-based needs, that time is not today.



Originally published February 13, 2013; updated on November 10th, 2017

About the Author

Steven Schneider is the CEO of Logi Analytics, where he brings more than 15 years of technology leadership experience. Steven has previously served as both Chief Operating Officer and Chief Product Officer at Logi, where he led the sales, product, engineering, marketing, and customer success teams. Prior to Logi, he was a founding partner of OnDemandIQ, a Hosted Business Intelligence solution, and a practice manager at leading web technology company Proxicom. Steven holds a BS in Computer Science from Virginia Tech and an MBA from the University of Southern California.