Mini-Masterclass Episode 10: The Most Expensive Digitalization Myth!

Start now, clean later: Why perfect data slows down your digitalization

Do you know the most expensive myth of digitalization?

The idea that you have to clean up all master data before you can introduce new software.

So the big cleanup begins: asset lists are exported from various systems, Excel files are merged, and labels are standardized. At the desk, a project team tries to map the entire asset structure of the company completely and without errors.

This takes weeks, months, or even years.

And as soon as the list is finally finished, it no longer matches reality.

Because while data is being cleaned up at the desk, production continues. Motors are replaced, components are changed, and temporary solutions are installed. A technician sets up a bypass, but the change doesn't end up in any documentation.

Anyone trying to theoretically freeze the current state loses against everyday production.

The eternal master data project

Clean master data is important. Without clear assets, components, and assignments, information is difficult to evaluate.

But that doesn't mean your data has to be complete and perfect before you start.

It is precisely this assumption that leads to a standstill in many companies before the actual project. Digitalization doesn't start in the workshop but with endless preliminary work:

• Standardizing asset labels,
• removing duplicate records,
• revising hierarchies,
• adding missing components,
• clarifying responsibilities,
• and matching Excel lists with each other.

The planned digitalization project initially becomes a master data project. The operational benefit is postponed further and further.

Employees do not yet receive better support. The documentation does not improve. And new knowledge continues to flow into notes, emails, or personal files.

The company waits for perfect conditions while data quality continues to decline during ongoing operations.

Reality changes faster than your Excel list

A production facility is not a static object.

Components are replaced, parts are rebuilt, and technical solutions are adapted to new requirements. Not every change is immediately updated in the Enterprise Resource Planning system, or ERP system, or in the technical documentation.

Therefore, a growing gap arises between the documented and the actual state.

In SAP, a specific motor may be recorded. On-site, however, this motor has long been replaced by another model. The technician immediately recognizes the discrepancy because he is standing right in front of the facility.

The project team at the desk, on the other hand, can only determine this information with considerable effort.

Therefore, the best place to improve asset data is often not the office.

It is the facility itself.

Start now, clean later

The pragmatic counter-proposal is: Start now, clean later.

Start with the data that is available today. Accept that it may be incomplete, inconsistent, or partially incorrect.

The cleanup then takes place gradually in the ongoing work process.

This initially sounds risky. In fact, it is often closer to operational reality than trying to create a perfect data set in advance.

Because modern systems should not only be able to work with clean data. They should help to continuously improve data quality.

The software thus does not become the endpoint of the cleanup. It becomes its tool.

Correct data where errors become visible

Imagine a technician standing in front of a facility and noticing that the documentation does not match the actually installed component.

In the traditional process, he would have to note the discrepancy, inform a responsible colleague later, and hope that the data record will be adjusted at some point.

In everyday life, this information is easily lost.

In a user-friendly digital process, the technician corrects the information directly on-site:

He photographs the nameplate and describes the change with a short voice message.

For example:

"The documented motor is no longer installed. A motor of type XY was installed. The replacement took place during the last unplanned downtime."

The system structures the information, recognizes relevant master data, and processes the change in the background.

The employee does not have to open an extensive Excel list or operate a complicated input mask. He documents the discrepancy the moment he discovers it.

Thus, data quality does not arise from a one-time action, but from many small improvements in everyday work.

The software must clean up for you

Many digitalization projects initially demand additional work from employees.

They are supposed to clean up data, transfer information, and restructure existing files before they can benefit from the new solution themselves.

This is the wrong order.

A modern application should support the employee in his actual task while simultaneously improving the data base.

The technician provides what he can reliably assess directly on-site:

• a photo,
• an observation,
• a short description,
• or information about a replaced component.

The system then takes over as much of the further work as possible:

• reading out content,
• structuring information,
• assigning assets and components,
• comparing existing records,
• and preparing corrections.

The software cleans up for you while you work – not the other way around.

Why the work context is crucial

A cleanup at the desk usually considers data in isolation.

There is an asset label, a material number, or a technical description. Whether these details still correspond to reality is often not recognizable from the list alone.

The technician on-site, on the other hand, has the crucial context.

He sees which component is actually installed. He recognizes modifications and temporary solutions. He may also know why the original documentation was deviated from.

This knowledge must be able to flow back into the system as easily as possible.

The higher the input hurdle, the more likely the discrepancy remains undocumented. That's why photos and natural language are so valuable: they fit the specific work situation and do not require extensive post-processing by the technician.

Perfection is not a starting point

Of course, "Start now, clean later" should not mean that data quality is unimportant.

It also does not mean that every unchecked piece of information should automatically be adopted as a binding master data record.

The central insight is rather: Perfection is not a realistic starting point.

You need a controlled process that can start with incomplete data and then continuously increase quality.

Clear rules can help:

1. Critical data first Not every piece of information has the same significance. Start with the assets and master data that are particularly important for safety, availability, and repair processes.
2. Capture corrections in the work process Employees should be able to report discrepancies exactly when they discover them. The process must be quick and simple.
3. Review changes traceably Relevant corrections can be reviewed and approved by responsible persons before final adoption.
4. Gradually expand the data base Every repair, every replacement, and every feedback improves the digital asset inventory.

Thus, quality grows with the use of the system.

The journey is the destination

Digitalization is not a state that you suddenly reach after a long preparation project.

It is an ongoing process.

Your assets change. Employees gain new experiences. Components are replaced and processes are adjusted. A data base that appears complete today already requires the next update tomorrow.

Therefore, a learning system is more valuable than a supposedly perfect starting list.

The decisive progress does not come from the one-time cleanup of all data. It comes from a work process in which new information is easily captured, existing details are corrected, and the data is continuously improved.

The most important insights of the episode

The central message is: Don't wait for perfect master data. Use digitalization to improve it while working.

The most important key takeaways:

• Perfect data is not a prerequisite for starting. Waiting for it postpones the operational benefit further and further.
• Excel lists become outdated faster than production stops. While data is being cleaned up at the desk, the assets are already changing again.
• The actual state is visible on-site. Technicians recognize discrepancies directly at the machine.
• Data should be corrected in the work context. A photo and a short voice message can suffice for this.
• Modern software must support the cleanup. It structures information and processes master data in the background.
• Data quality grows with usage. Every repair and every feedback can improve the asset inventory.
• Digitalization is a continuous process. The journey is the destination, not the perfect starting file.

Are you still stuck in the eternal master data project?

Then stop trying to freeze reality at the desk.

Start with what is available. Let your employees capture discrepancies where they become visible. And use software that cleans up for you while working.

Start now, clean later.

• Prioritize critical assets instead of all data
• Capture corrections directly on-site via photo and voice
• Continuously improve master data during ongoing operations