In-Depth
Report Debunks BI Usage Myth
Real-world BI usage typically hovers at around 8 percent of all users in an organization -- with just over 11 percent in extremely aggressive BI shops -- a figure much lower than BI vendors would have you believe.
The latest installment of the BI Survey (formerly OLAP Survey) from the Business Application Research Center (BARC) serves up more than a few surprises. It also explodes a few myths.
Take BI adoption, for example. Business intelligence vendors like to talk up a 20/80 split -- i.e., in any given organization, only 20 percent of users are actually consuming BI technologies; the remaining 80 percent are disenfranchised. According to BI Survey 8, however, most shops clock in at far below the 20 percent rate. In any given BI-using organization, notes Nigel Pendse, a principal with BARC and the primary architect of BI Survey, just over 8 percent of employees are actually using BI tools. Even in industries that have aggressively adopted BI tools (e.g., wholesale, banking, and retail), usage barely exceeds 11 percent.
That's a far cry from 20 percent figure, yet the 20/80 split remains a mainstay of BI vendor rhetoric. Many tout their own (typically self-serving) spins on the issue, using figures as usage figures as high as 30 percent. No matter what figure is quoted, most BI products rarely reach one-fifth of a user population, yet vendors will claim that their product breaks through that barrier, penetrating to the remaining 80 percent.
To their credit, some vendors concede that the 20/80 "industry split" is a largely arbitrary figure. "We in the BI industry seem to have arrived at this magic figure, this artificial 20 percent threshold, where we say that penetration [of BI tools in any organization] only reaches about 20 percent [of users]," concedes Guy Weismantel, director of marketing with Microsoft Business Intelligence.
This isn't to detract from the basic message of such marketing, Weismantel stresses: in most shops, BI tools are chronically under-used. Vendors (along with integrators, consultants, and other stakeholders) have to do a better job of driving adoption. At the same time, he acknowledges, the 20/80 split does have a strange -- seemingly inexplicable -- kind of currency. "I'm not sure how we arrived at this [number], but that seems to be the magic number."
It's a wildly misleading magic number, according to Pendse and BARC, but it has stuck. What's more, Pendse says, it even has currency (of a sort) beyond vendor marketing: both vendors and BI consultants are the more deceived when it comes to BI tools adoption.
"Vendors and consultants think that BI deployments are significantly wider than users actually report. For example, users report a typical (median) deployment rate of eight percent, but consultants report a rate of 9.5 percent and vendors, 14 percent," Pendse writes, noting that "36.4 percent of users report deployments to less than five percent of their organization's employees, while only 17.2 percent of vendors think that their products are used by less than five percent of employees. It seems that vendors believe too much of their own propaganda about mass deployment of their products."
In other words, Pendse and BARC argue, BI vendors are falling far short of the "BI for the masses" pitch that they've been trumpeting (in many cases) for half a decade or more. In fact, Pendse indicates, BI adoption rates actually dropped between 2007 and 2008.
"Despite the vendors' optimism, the results were disappointingly low … and dropped even further in 2008. Only 7.8 percent (down from 8.6 percent in 2007) of those surveyed said that more than 50 percent of their employer's employees were regular users of BI applications," he writes. "The mean is just 13.23 percent and the median an even lower 8.16 (down from 8.74) percent."
When pressed, Pendse reports, most vendors will at least admit to inflating their adoption numbers. Vendors, for example, believe that median usage for their tools clocks in at around 14 percent -- 75 percent higher than the user-reported median. It's also a far cry from the 20 percent or more that vendors like to trumpet in their "BI for the masses" baseline figures.
"[A]lthough vendors do realize that BI deployments are narrower than they publicly claim, the actual figures are much lower than even they think," he writes.
Why do BI deployments lag? Pendse suggests a number of reasons, including -- especially in larger shops -- "security limitations, user scalability, and slow query performance as problems." Not surprisingly, he says, internal politics and (more precisely) internal power struggles are also to blame: sites with both administrative and political issues reported the narrowest overall deployments.
"When looking at deterrents to wider deployment, it is not surprising that hardware cost is the most common problem in sites with wide deployments; however, in typical (smaller) BI deployments, this is one of the least common deterrents," Pendse adds. "Data availability and software cost are also seen as significant deterrents to wider deployment. Sites that reported that software was too hard to use again had the smallest deployments."
A free sample chapter and report preview are available at www.bi-survey.com (short registration is required).