In What Order Should You Digitize Master Data?
Key takeaways
- Most digitalization projects start with all the data and invest months and millions in master data before any value appears.
- Circonomit reverses the order: first the company's most important decision.
- Then the five to ten most important drivers of that decision.
- Only then the master data that matters for exactly those drivers. Anything missing starts as an estimate and is gradually replaced by real data.
- This creates transparency from day one, and master data quality grows where it is needed.
In this episode we talk about a well-known approach: generating insights from data to make better business decisions. But what if data quality is insufficient, and it also works without a large upfront investment in master data and implementation? We share how we start with the central business questions, use them as anchors and digitize along these questions instead of staying blocked by missing master data.
Contents:
00:34 Why most digitalization projects fail
01:57 People as an important source of information
03:37 How digitalization for good decisions can work in a targeted way without a huge investment
Why digitalization projects often fail
Welcome to our first topic series. I brought my whiteboard because one question has come up again and again over the last few months: how do we build good decision models when the data is not there?
There is a picture I once drew old school on a sheet of paper during a talk. Most digitalization projects work like this: you collect data, extract insights, and afterwards you still don't know what would happen if, or what caused it. Behind this is a pyramid that has followed me since my time in mechanical engineering and that probably every second person knows: data at the bottom, insights above, and finally knowledge.
Almost every digitalization project is built this way: we look at the data and then invest millions of euros and twelve months in master data quality to eventually gain insights and, from them, knowledge. The problem: most of it already fails at the bottom. You spend a lot of money on digitalization and in the end realize it didn't work and only half of it is digitized.
When we have to make a decision, we need different people in the information chain and different data points. It starts trivially: one person has to tell another something, someone is sick, an email isn't forwarded. And it goes all the way to questions like: how do we keep our on-time delivery? How do we calculate the cost of a product correctly? Having 70% of the bill of materials digitized very well doesn't help much if the allocation keys for total costs never flow back to the products. Then we still don't know which product makes or loses money.
Flip the pyramid: start with the decision
A Circonomit product principle since day one: good decisions in a company must not depend directly on master data quality. Otherwise we are always blocked, run into huge investment budgets and come out with an unsatisfying result. Of course there are digitalization projects that work very well, including S/4HANA implementations. But we only see that about once in 50 times.
That's why we flip the pyramid in our thinking. This is less a data topic than a different way of thinking about digitalization, especially as the world becomes more complex and more and more data sources are added. We focus on what matters: that we, as entrepreneurs, heads of controlling, plant managers or executives, can make good decisions. We first ask every customer and partner: what is your most important and most urgent decision right now? The decision is our starting point.
This puts the focus on the outcome, namely steering the company well, instead of getting stuck at the hurdle of data collection. Then we ask: what are the most important drivers of this decision? These can be bottlenecks, everyday disruptions such as a sick employee who causes one order and all follow-up orders to shift, measures, or inventory. From many conversations we now have a catalogue of 50 to 100 such influencing factors.
Build master data along the most important decisions
From the decision we get to the drivers and focus on the essentials: the five to ten most important information chains that influence the decision. Then we ask: which of these can we already find in data sources, and which still live in employees' gut feeling? That is exactly the crux.
This way we take care of the master data that really matters and digitize along the most important decisions in the company. The benefit: we create value from day one. Our product looks a bit like a whiteboard that can calculate. We add the decision as a building block, for example the contribution margin of products or the on-time delivery of orders, add the five to ten most important drivers and, together with the team, create transparency from day one. Whatever is not yet available as data we add as an estimate and improve master data quality step by step.
This is the first step of problem solving, more on that in a separate video. The cause close to our hearts behind this: that Europe becomes able to steer again and that, as an industry, we understand how to steer well in complexity.