Organisational design for analytics function needs governance

Scant are the absolute rules for the organizational design of the analytics function. It can be outsourced or in-house; sit within a business unit or run from the centre. But it needs governance.

When the whole point of business intelligence (BI) and analytics is to bring disparate information together, it can be difficult to know which business unit or department should start the process.

Yet the two business functions most often given the task to lead an analytics mission could well be the least appropriate to perform it, said Howard McMinn, an organisational and change management consultant at Deloitte’s UK operations. By default, organisations often single out finance as the starting point for analytics programmes, because it is numerate and can see how data relates to the bottom line.

For more on the organisational design of the analytics function

Learn how the data analytics team needs a business home

Advice on structuring data warehouse and business intelligence teams

Gartner on BI strategy and structure: aim for middle path between monarchy and anarchy

Another common starting point for BI and analytics, not surprisingly, is the IT department. Although IT and finance both have their strengths, businesses should look beyond them when setting out the organisational structure around analytics, according to McMinn.

“Look at the natural resistance of any organisation going into transition,” he said. “It is not their fault necessarily, but finance departments bear the pejorative of ‘You just understand the numbers, not the business.’ Meanwhile, IT is seen as pushing new technology. People ask, ‘Is this taking us forwards?’”

Those starting out on a formal business analytics strategy, which could bring together all sorts of enterprise reporting, querying and advanced analytics applications as well as analysis of unstructured data types, should identify the department that could benefit the most and recognises the potential gains to be had, McMinn said.

“Our strong belief,” he advised, “is that you start in the area of least resistance: Which bit of the organisation already gets this? That may depend on the industry. It could be marketing, finance, IT or HR, but who has the yen to build up its own capability? That’s the place to start; then you let it permeate the rest of the organisation.”

A sign of a mature approach can be that a distinct analytics function almost disappears, McMinn added. “If you have got an organisation that is truly analytics-driven, there is no analytics department. There may be one or two true specialists that develop algorithms or models, but everyone has that mindset imbued in them.”

Stacy Blanchard, global head of Accenture’s organisational effectiveness consulting service for analytics deployments and its human capital analytics practice, agreed with McMinn that technology-led initiatives may be short-lived and fail to fulfill their early promise. And she pointed out that because analytics projects can create such competitive advantage, companies may not enjoy the luxury of building their own capabilities.

“Going to managed services, whether outsourcing or some sort of hybrid, is a trend we’re seeing,” she said. “Analytics is one of those key levers in becoming a high performer and sustaining that position. What time will it take you to build that mature capability, and what is the cost? If the gap is so massive, companies are going to ask themselves: ‘Do we have the time to build this from the ground up? Or is the way to stay ahead in this market to outsource it?’”

Even so, organisations thinking of outsourcing some or all of the analytics function must ensure that the results feed directly into top-level business executives. “That goes back to best practices around outsourcing,” Blanchard said.

Not only to interpret, but to change
No matter what the structure for a BI and analytics operation may be, it must be able to effect changes in business processes to become a credible entity within an organisation, said Andreas Bitterer, a Gartner research vice president specialising in business intelligence.

“BI is there to let you act,” he said. “The whole point of BI is change. There are a few exceptions, such as compliance requirements, but in terms of finding trends and patterns, that’s when you want to act.”

As such, businesses need a governance centre to approve and oversee BI projects, with enough muscle to ensure that employees act on the results of BI and analytics applications, Bitterer said.

“There are countless situations where people are pumping out reports to lots of users, and they hope that it’s going to be good for something,” he said. “You need senior people, at least director or VP level, who understand what people are doing in sales or marketing.

“And whatever comes out of this team should be understood as a policy or a mandate. Otherwise, it is more like BI anarchy than a BI strategy.”

There are few rules to the organisational structure around BI and analytics. It can be managed in-house or outsourced, and it can sit within an individual business unit or be run centrally. But analysts pick out effective governance as a common factor to ensure the technology architecture is coherent, projects are based on business needs and senior executives drive a better-informed organisation.

Read more on Business intelligence and analytics

Data Center
Data Management