Enterprise Reporting in a Changing World
Gartner’s decision to separate their magic quadrant BI report highlights the changing uses of Business Intelligence in organisations. With modern user-led software is the role of the BI professional going to disappear?
The world of Business Intelligence and Business Analytics has been evolving over the last few years, an evolution which reached another peak recently with Gartner separating their magic quadrant BI report – one report for modern, user-led software and another for the classic enterprise, IT-led tools.
It looks like the user-led tools such as Tableau, Qlikview, Watson Analytics and MS Power BI are enjoying increasing popularity and for good reason. They allow users to govern their data, free of IT dependency and therefore contribute to faster decision making and more efficient report authoring.
When users are empowered to freely access their data in almost any way they see fit, they familiarise themselves with the data and make themselves a part of the process for data-based-decision-making. That’s a very good thing and any organisation would benefit from having their users closely engaged with the data they need to make decisions.
This is especially true as users become increasingly savvy. Analysts these days can be as computer literate as IT professionals and they understand not only their data but also their metadata and the way the data flows into the organisation. Modern-day tools are built to use that savviness to gain an advantage.
However, the world of user-centric BI has some major pitfalls and an organisation which will not be aware of these might find that their BI solution is not working out as well as expected.
1. First, there’s the issue of standardised reporting. These require expertise that is still beyond that of an analyst or an end-user.
There’s more than being a good cook to being the chef at a restaurant. You have to also know how to run a kitchen, deal with suppliers, take reservations, direct and manage the kitchen and service staff and so on.
Perhaps the clearest example is the topic of data visualisation: Most end users do not specialise in the field of data visualisation and might, therefore, create reports and dashboards which mask, rather than reveal, insights. Altis partner Stephen Few once wrote:
“Data sensemaking requires skill augmented by good technologies. Even though data sensemaking skills can be developed by almost anyone, they can only be acquired through sustained effort to learn the relevant concepts, principles, and practices, which takes time. This hard fact isn’t as appealing as the fantasy that technologies can turn us into data analysts overnight.”
This is true. BI professionals who have studied the issue of data visualisation will still produce more effective reports in terms of “data sensemaking”. Their reports are also likely to be more efficient in terms of resources consumed.
2. Also, BI professionals are agnostic to the data. A BI expert will extract, model, load and report the data if it showed revenue to be millions or billions just the same. When a BI professional approaches the task of sorting, modelling and reporting data, they don’t expect the data to do anything.
End users and analysts, however, are more closely involved with their particular data. They often expect it to show this figure and not that. Sometimes it’s flat out wrong and the reports might be distorted and manipulated – even in good faith.
Other times, it means the organisation’s different functions understand different things when a certain measure is discussed. This is a problem inherent to most user-led BI software because the majority of user-led software encourages silos of data.
As an example, you might get a situation where a sales manager says revenue and means actual and pipeline, while their CFO counts actual only. They both say “Revenue” and their reports say “Revenue” and there’s no indicator that they mean two different things.
3. Another problem is that end-users and analysts aren’t always aware of IT wide considerations: data security and obscurity, the efficiency of models and query performance etc.
This problem is intensified by IT departments sometimes not being fully cooperative with the business user-led vendor selection, as they would rather have an IT-led solution which gives better governance (however at the cost of taking away user capabilities).
Knowing these risks is the first step in mitigating them. Mitigation could be in many forms, from end-user training to organisation-wide governance policies.
In conclusion, user-centric BI tools are definitely very advantageous and can drive more agile BI and allow organisations to identify opportunities and act on them faster than ever before, but the need for enterprise BI professionals and skills is still very much present for a BI project to succeed.