From data to decisions: why data-driven performance management starts long before the analysis

There is a lot of talk about how the finance function should become more strategic, work in a more data-driven way and use AI for analysis, forecasting and decision support. But during our roundtables at FAR''s Verksamhetsstyrning, another question turned out to be just as central:

Have we actually created the conditions for it?

Exopen brought together controllers, heads of finance and other finance leaders from organisations of different sizes and industries to discuss the road from data to decisions. Despite large differences between the businesses, the same fundamental challenge kept returning: there is rarely a shortage of data. The problem is that it sits in too many places, in different structures and with different definitions.

Want to know where your own organisation stands? Take our quick assessment – it takes a few minutes and gives you a picture of your conditions.

The finance function does not have a data problem because it has too little data

ERP systems, CRM, project systems, HR systems, sales systems and operational business systems together create an ever-growing amount of information.

Each system may work well for its own purpose. The problem appears when the information has to be used together.

In the conversations, participants described organisations with several ERP systems, different charts of accounts and dimensions, or even the same system where information is used and defined in different ways. To understand the whole picture, data therefore has to be collected, mapped, checked and consolidated. Often manually.

That creates a paradox.

The more data the business gains access to, the more time the finance function risks spending on pulling it together.

And when Excel becomes the place where information from different systems is finally consolidated, it easily creates further dependencies on individuals. Someone knows which file is the right one, how the figures should be mapped, or why two systems show different results.

It works. Until it no longer does.

The question is not only whether the number is correct

Data quality was discussed a great deal during the day, but the conversations also showed that quality is broader than a single figure being correct.

For information to work as decision support, the organisation needs to know what the figure means, where it comes from and whether it can be compared with the equivalent information from other parts of the business.

One participant described how the same product could have a large number of different names across the organisation. Others described how the same chart of accounts was used differently in different companies.

That illustrates an important difference:

Having data is not the same as having shared information.

Only when data is harmonised and placed in a common structure can it begin to work as a shared language for the business.

AI makes the data foundation even more important

AI was naturally part of the discussions. But perhaps not in the way you might first expect.

AI ended up at the top of the chain rather than at the bottom.

If data is spread across different systems, the same terms mean different things and important context lives with individual people, then AI will also struggle to understand the business.

A model can process large volumes of information. But it still needs to understand what that information represents.

This means that investing in AI also puts new light on an old problem: the organisation''s data has to be structured, harmonised and reliable.

Raw data is not enough. The context has to be there too.

Should every system really be replaced?

Another interesting discussion concerned the system landscape.

One route is to standardise the business around a small number of shared systems. During one of the conversations, a participant described an extensive transformation from around 200 systems towards two. At the same time, it was a multi-year change programme with significant investment.

For many organisations, replacing every system is neither realistic nor necessary.

The ERP system may work excellently for accounting. The CRM system may be right for sales. A specialised business system may be crucial for operations.

The more important question then becomes:

How do we create a shared picture of the business without everyone having to work in the same system?

This is where integration, harmonisation and a common data model become central. People can keep working in the tools that suit their job, while the information can be used jointly for reporting, analysis, planning and follow-up.

The goal is not better data. The goal is better performance management.

It is easy to turn data projects into IT projects. But the discussions at FAR showed why the question is really much bigger than that.

The finance function does not want to collect data for the sake of data.

It wants to understand what is happening in the business, spot deviations earlier, analyse why they occur, update forecasts and give the business better support.

When a large share of the time goes into finding, consolidating and checking information, less time is left for exactly that. Several participants also described more time for business partnering and business development as the real goal.

That is why the discussion about data-driven performance management needs to start further down than the dashboard.

It starts with questions such as:

• Where is our data?
• Do we have shared definitions?
• Can information from different systems be compared?
• Can we trace where the figures come from?
• Can the right people access the right information?
• Is the information available when the decision actually has to be made?

When those questions are resolved, the role of the finance function changes as well.

Less time goes into producing the decision support. More time can be spent on the decision itself.

That is perhaps the most important lesson from our roundtables: data-driven performance management does not start with the report, the dashboard or the AI analysis.

It starts with the data beneath the surface.

Want to know where you stand?

Take our quick assessment and get a picture of your conditions for data-driven performance management – or talk to an expert and we will guide you directly.