Why self-service BI isn’t the be-all and end-all

Widening access to business intelligence (BI) tools beyond the bounds of the IT department through self-service was a big step forward in the democratisation of data. But is ‘free for all’ access really good for our organisations?

According to Australian BI developer, Yellowfin –  not necessarily… especially when it comes to ensuring data can be trusted and shared.

Yellowfin CEO Glen Rabie likens this scenario to the self-service checkout at the supermarket – while you can scan and bag your own groceries to avoid long queues at other checkouts, you’d really rather someone else would do it for you.

Speaking at Yellowfin’s recent ‘Making BI Easy’ roadshow, Rabie says the company has undertaken research on what proportion of its customers actually use its tools to conduct their own analysis or write their own reports, and how many consume what is available, and the take-up rate for the former category is still less than 10 per cent.

“Masses of people consume but only a handful of people ever want to become the experts to do this well and do it properly. So I think BI vendors themselves are part of this hype cycle that everyone needs self-service, when really there are a handful of people in the business that really want to do it, and most don’t,” Rabie says.

The question then, is how to find the balance between being able to provide self-service BI to those who want it, and how to make it easier for information creators – IT and data experts – to be able to provide ready-made data-based reports and analysis to those who want to consume them.

With organisations now requiring cross-functional collaboration to make the fast decisions that will keep them competitive, being able to easily share data is another important factor, according to Rabie.

“Organisations need to be able to quickly and easily share data with the people who need it at the right time in the right place. The more people have access to data, the more innovative solutions they can build. The more they understand their business, the more proactive they themselves can be, so they have insights to make the visions that otherwise they wouldn’t have,” Rabie says.

He contends that while sharing information broadly in this way, organisations must also maintain a level of governance and control, not only to ensure data isn’t shared in unauthorised ways, but also to enable users of the information can trust the validity of data and data analysis.

It’s a problem that analyst firm Gartner has already identified, saying that through 2016, less than 10 per cent of self-service BI initiatives will be governed sufficiently to prevent inconsistencies that adversely affect the business.

“Part of the governance issue is really about trust, and people truly trusting the data they have. So who has built it for them? How has it been published? Through what process have they gone through so that consumers know this data is validated?,” Rabie says. “My view is most BI projects fail because ultimately the business loses trust in the platform they are using, and then they turn to other tools because they believe they can do it quicker and easier by building it themselves.”

So for organisations looking to make data insights more freely available, how can they ensure data consumers trust what they are seeing?

Rabie points to a simple mechanism Yellowfin developed in conjunction with Macquarie University for watermarking content and having an approval cycle in place to affirm that the content in a particular report has been validated by a data steward.

“They say, your calculations are correct, your filtering is correct for what you are trying to say, therefore yes, we will approve it, and apply a watermark that says it has been validated. Anyone can see that watermark and feel that it has been vetted, as opposed to someone just downloading data, pushing it through Excel, creating a whole lot of charts, missing some really important conditions of the data, and creating a reports that tells a very different story from what the organisation believes based on its own data,” Rabie says.

He says the benefit of a ‘pervasive approach’ to BI – that is, sharing intelligence throughout the enterprise – is in the transparency of information and giving all staff the information they need to do their job well.

“To enable that, it comes down to can you build a platform that delivers that information to people on multiple devices and in easily consumed ways? Is it affordable – can organisations take this platform and scale it out for the enterprise? And is it not so complex that it’s achievable? That’s our vision and we continue to work towards that.”