Savvy users, the final frontier

If big data and analytics apps are to achieve widespread success, then business users will have to pick up some of the high-level skills needed to get the best out of them.
Written by Phil Wainewright, Contributor

The problem with putting increasingly powerful software tools into the hands of more and more people is that, very often, most of them have no idea what to do with it. This is not a new problem in computing, but it's commonplace today thanks to the ease and low cost of delivering awesome capabilities to the masses using cloud platforms.

I see many SaaS start-ups in the business market struggling with this conundrum. They develop impressive functionality and then deliver it for online sign-up in a low-touch commercial model that makes no provision for intensive one-on-one consultancy. Few prospects sign up, and of those that do, most drift away after failing to produce any value from using the app. This is more than just naive 'build-it-and-they-will-come' thinking at fault (though there's usually an element of that). The apps in question often deliver exactly the same functionality as far more expensive and complex alternatives that already have a proven track record in the enterprise market. It's down to a lack of specialist skills in the wider business user population to be able to produce results when using the apps.

Nowhere is this more of a problem than in the fast-growing field of analytics and big data. As Constellation Research analyst Neil Raden wrote earlier this week, "There will be very few 'data scientists' in commercial organizations ... Finding people to fill this role is difficult ... They don't grow on trees." IBM's VP of emerging technologies, Rod Smith, thus hits the nail on the head in his comments this week that vendors must focus on producing applications that produce actionable results from big data rather than raw analytics.

Such thoughts were front-of-mind for me after a recent meeting with Jim Burleigh, CEO of sales analytics vendor Cloud9, who is highly critical of earlier generations of cloud analytics. Putting conventional business intelligence solutions in the cloud "is a wonderful half-step," he says, "but that's it. It eliminates that back-end, but it still necessitates someone that understands cubes, etc, to set the stuff up." And when someone wants to change or do something different, it means going back to IT to get it done. "You haven't put the user in control, that's why it's a half step," he concludes. In his view, business intelligence should add value to specific applications by doing a lot of the analysis automatically and straight away delivering meaningful results.

In Cloud9's case, the target application is sales performance management, and its latest product aims to help sales managers identify which deals in the pipeline they should focus most resources on. Rather than simply looking at aggregate data, the application also performs detailed change analytics, looking for patterns that indicate deals that could be borderline — for example, evaluating when key activities have taken place, patterns of time spent in each phase of the sales cycle, and revisions to important milestones such as predicted close dates. The aim is to highlight areas of risk and thus improve forecast accuracy and win rates. For sales managers, "getting that early warning system is huge," says Burleigh.

This is a useful step in the right direction by the application vendor, but it still requires some investment of effort and commitment from the sales manager to make sure the right risk factors are being evaluated. "We're talking to sales management about how do they determine risk," says Burleigh. "What would the bellwether look like to them that a deal is going off the rails?" Getting the right answer (and thus the most accurate analysis) depends on those sales managers being savvy enough to convert their experience and knowledge into metrics the application can work with.

To some extent, bridging that gap depends on developers building in even more automation so that the app can prompt the user through that process. But it also depends on the users gaining skills that they haven't previously had to have. They don't need to become data scientists, any more than anyone these days has to be a design professional to produce a presentable Word document. But users do need to be able to recognise and understand some of the high-level principles to be able to interact effectively with these apps. This is the final frontier for a range of up-and-coming app categories such as workflow automation, inbound marketing but most particularly analytics. Application vendors not only have to build great apps, they also have to find a way to create savvy users.

Editorial standards