Overcoming resistance to BI in the cloud
“For business intelligence, the cloud used to feel like a headwind. Only a small minority of people were seeking it out. Now, it’s more of a tail breeze.”
Brad Peters, CEO of cloud-based business intelligence provider Birst, has never been under any illusion that the shift of enterprise analytics to the cloud was going to happen overnight. For one, the move was always naturally going to trail the adoption of cloud applications by a few years at least. But he’s now seeing many of the perceived barriers to cloud BI falling away, and the “slower-than-anticipated build up” turning into healthy demand. There’s even talk — by Gartner, no less — of the market reaching some kind of tipping point of adoption within the next one to two years.
Certainly, the numbers are only just appearing to support such a conviction. Barely a year ago, Gartner’s annual business intelligence decision-maker survey showed only 33% of respondents expressing an interest in deploying BI in the cloud. The latest figures (not yet publicly disclosed) show a notably rise in that, says Peters. Though unlikely to cross the 50% threshold, the upswing is consistent with the performance at Birst.
For the past two years, the Silicon Valley start-up has seen triple digit revenue growth, as bluechip companies such as American Express, InterContinental Hotels Group, New South Wales Government, Nippon Life, Samsung, Schneider Electric and Sunny Delight have signed up or renewed as customers. Interestingly, not all have signed up as users of pure cloud-based BI. Two years ago, Birst concluded the move to public cloud was too daunting for some prospective customers and that it needed to offer them the option of running its software on premise. Its Birst Appliance now accounts for 30% of revenues.
For Gartner that was a wise move in a market that still needs some persuasion: “Birst has achieved momentum not because of its cloud BI credentials, but rather despite them, given relatively low cloud BI investment intentions,” says the analyst group.
That highlights how the decade-old company’s successful growth is not yet rooted in selling cloud for its own sake but stems from offering functionality that can rival that found in many, much more established BI products such as SAP’s Business Objects or IBM’s Cognos.
As good as on-premise?
“Our whole message to customers has been, ‘there are a lot of advantages to cloud, but it is a distribution model — a way to consume software that’s easier to set up, easier to maintain, easier to manage for geographically-dispersed workforces.’ But if Birst applications aren’t competitive with the on-premise alternatives, [we] are not going to win anyway. And if there’s no cloud strategy and everything is on-premise, we are going to go were the data lies.”
Gartner agrees that Birst’s five-year, pre-launch effort to build a core BI environment, has provided it with a comprehensive BI stack. It’s assessment: “Birst has been successfully competing for and winning deals because of its functional breadth, depth and strength, ease of use and low cost of ownership value proposition. Birst offers a broad and highly rated and integrated set of model-centric, enterprise BI and managed business-user-oriented data discovery capabilities and predictive analytics applications that can be deployed in the cloud or in an on-premises appliance for companies that want to take advantage of Birst’s product functionality, but are unwilling to put their BI deployments in the cloud.”
As Peters emphasizes: “Cloud or no cloud, we have to beat the on-premise players regardless of distribution model. I think cloud is in our favor, but in a competitive set it’s still ‘can you beat the on-premise guys at their own game or not?’”
The question is also, can you live with the on-premise guys?
Other cloud BI vendors in this space have built proprietary back ends that do not readily integrate with on-premise databases, he says. In contrast, “we assumed our software was going to have to work well with on-premise as well as cloud sources.”
It does so by siting BI tools and customer applications on top of its own relational OLAP engine and in-memory columnar database, as well as supporting a range of third-party databases, including Oracle, SQL Server, ParAccel and Infobright.
Triggers for adoption
Notwithstanding that dual approach, he is seeing a definite change in the wind direction. The argument by customers used to go like this, says Peters: ‘I have one or two apps in the cloud, maybe a sales app or an HR app, but still the critical mass and the epicenter of my data is on prem, so obviously I need to have my BI where my data is.’
“Well, that is now moving as companies deploy Salesforce, NetSuite, Workday, Marketo, Zuora and all these different other cloud applications. Now there’s a split that forces people to think, “well, why would I bring my data back on premise, and lose the cost and agility advantages of cloud.”
Meanwhile, one major catalyst for growth in cloud BI has come from a source that would have seemed inconceivable when Birst started up in 2004: Amazon. The launch of its Redshift data analytics platform has redrawn the economics of BI, says Peters.
“It’s not so much that people are today moving their warehouses whole hog to the cloud. They’ll get there, but they are not there right now. But from a price/performance perspective the contrast between what they have on premise and Redshift is so stark that it’s changing the mindset of what people believe is possible in the cloud,” says Peters.
“They used to think of cloud tools as toys — you could load small data sets and you could get a little bit of performance but you couldn’t really do much other than with pre-packaged stuff. Now, when they see they have an MPP database in the cloud that can store up to half a petabyte of data, with up to a hundred nodes, power that can scale up and scale down — and it’s a thousand bucks a terabyte a year — then all of a sudden something like a Teradata or Netezza data warehouse looks very expensive.”
However, there are legitimate concerns that are still dampening enthusiasm — “real issues for cloud that cause people to take pause,” Peters says. One relates to internal politics and the growth of ‘shadow IT’ buying patterns.
“Cloud has tended to be more championed by business than IT. If elements of BI go cloud, internal control within an organization may not be quite so clear. For IT, there is a fear that the cloud is going to be a way to disintermediate them. For the business, cloud is just a way to go round IT. So, it’s true there is some disintermediation that can happen unless IT takes the lead on cloud strategy for their company.”
The most legitimate reason people have for resisting a move to cloud BI, however, is that, once their data is outside of their data center, “there is a perception – and a little bit of a reality too,” he acknowledges, “it is harder to work with.”
“The notion is if you have data sitting in your own database, you can do whatever you want with it. But if you load it into a multi-tenant BI solution, you can’t do as much with it because of the system in place for security or for ‘noisy neighbors or other reasons.’”
This is something he says Birst has had an intense focus for many years, and addressed with a set tools that provides users a high degree of control over the data itself from a browser interface, Peters says, adding: “With systems like Redshift we can also now give you naked access to the database itself, under the covers, so you have every bit of control you’d ever have on prem.”
Birst’s Brad Peters is not short of metaphors that describe the current state of the cloud BI market. And they all betray a certain relief that the market is about to accelerate: “It may have been slower to build than most people thought but it does feel that like the snowball is getting bigger.”
There are fundamental reasons why cloud BI has been slow to climb the adoption curve, but as a larger proportion of business applications shift to the cloud it is inconceivable that BI will remain a largely on-premise activity.
Rather, for some time to come, the challenge will be to blend BI across both environments, which puts Birst in a stronger position than many of its pure-play cloud rivals. It also means the company is likely to shoot for an IPO in 2015 (assuming it does not first fall into the arms of one of the long list of “legacy BI vendors” — Oracle, SAP, IBM, MicroStrategy…).
Featured image: iStockphoto