Cloud Analytics Trends: New Data, A Few Anecdotes, and a Couple Damned Lies

The hype around cloud analytics continues unabated, and upcoming announcements at Dreamforce will no doubt add to the noise. So I thought it would be useful to take a moment to share some actual data around what’s going on in cloud analytics, combined with some anecdotal observations. And to note a few points where survey statistics are more akin to “damned lies” than anything else.

Data points

  • Analytics investment is up. From Bluewolf’s annual “The State of Salesforce” report, 68% of companies are increasing their analytics investments, 20% of them significantly.

    Cloud Analytics Trends
    Cloud Analytics Trends
  • Why? Business demands are increasing. Based on preliminary results of an Informatica-IDG survey on cloud analytics (N=216 so far; the final results will be released in a few weeks), IT organizations are experiencing increased demands for new analytics and data management capabilities. The top four are: improving data quality (75%), improving ways to explore data visually (72%), analyzing data in real time (67%), and supporting big data sources such as the internet of things (62%). No wonder Salesforce is continuing to invest in Analytics Cloud, Big Data and the Internet of Things. They’re following the money.
  • Cloud analytics adoption is accelerating. The early results from the Informatica-IDG survey show that just 17% report that cloud analytics solutions are not on their radar over the next 12 months. In fact, 11% of organizations have already deployed cloud-based analytics solutions, 18% are actively planning to deploy cloud-based analytics solutions and 21% are evaluating solutions. More than one-quarter are in the information gathering stage (28%).
  • Implementations of cloud application integration and cloud database technologies pave the way for cloud analytics and data warehousing. Amongst early respondents who have cloud analytics on their radar screen, we asked which technologies they have deployed or are planning to deploy in the next year. Cloud application integration was the most widely cited at 53%, followed by cloud databases at 49%, visualization tools at 48%, cloud data warehousing at 46%, and cloud analytics platforms at 44%. This makes sense since you need cloud application integration and database technologies as foundations to support cloud analytics.

Damned lies

  • Who says bad data is a problem? In the Informatica-IDG survey, the preliminary results state that the top two challenges with traditional BI/analytics solutions are cost/budget (44%) and time-consuming configuration/maintenance (41%). Inconsistent, incomplete or incorrect data ranked well behind at #7, with 33% of respondents. But this is where a statistic starts looking like a lie. When you segment the respondents between IT and business roles, 51% of business people say bad data is a challenge, whereas only 29% of IT folks say bad data is a problem. Holy disconnect—after years of talking about business/IT alignment, doesn’t seem like these two groups are on the same planet. In fact, this is the one topic where the gap between business and IT respondents is widest. On this front, I weigh the business users’ feedback more heavily—if the data in the report or chart doesn’t look right to them, it probably isn’t. And bad data suddenly ranks as a top problem.
  • Architectural purity—often sought, almost never achieved. 25% of respondents in the Informatica-IDG study so far have stated that they plan to migrate completely off on-premise analytics to a pure cloud analytics environment; 17% plan to stick to an on-premise architecture (40% plan a hybrid approach). I think both numbers indicating a pure cloud or on-prem approach will prove to be overstated—time will tell. The folks forecasting a migration from a current on-premise to a pure cloud environment are likely to find that there are some patterns where their on-premise solution makes the most sense, and some deployments will linger. The people sticking to their on-premise architecture are very likely to find hidden cloud-based analytics deployments scattered around their organizations already—or very soon—they just don’t know about them.


  • Cloud analytics—tantalizingly easy? It’s harder than it looks. I’ve seen this happen numerous times. Business users get pulled into cloud analytics because almost all the tools are highly visual and seemingly intuitive. No one worries how to make the data and the rest of the underpinning work—because the demo is just way cool. Until you find out you need to code in R to get the data into the shape required to make the cool swizzly visual stuff work. Oops. After the initial projects get stuck, either some hard work begins or they are abandoned.
  • No one-size-fits-all. Most large companies have not been able to standardize on a single analytics platform, cloud or otherwise. The few that I have observed trying have failed. Why? Too many disparate business needs for different types of analytics and different data sets, and no single platform can meet all needs well. And little business patience to wait for IT to make the torturous customizations necessary to get a generic platform to

We’re looking forward to analyzing the full results of the cloud analytics study we’re conducting with IDG over the next few weeks, so stay tuned. And we’ll see over the next few years how accurately the respondents predict their own future.