Legacy systems are not the main obstacle to DX

Legacy systems get named as the reason digital transformation is not progressing. It is a long standing explanation.

IPA's DX Trends 2025 does not support it.

In Japan, the United States and Germany alike, fewer than 20% of companies say legacy is a major obstacle. Legacy modernisation activity has also plateaued.

What under 20% means

If fewer than one in five companies calls it a major obstacle, then for more than 80% legacy is not the binding constraint.

Legacy still exists in these companies. It is simply not the centre of why the work has stopped.

Modernisation activity flattening points the same way. Effort is not increasing, which most plausibly means it has been deprioritised.

What is actually blocking

The same IPA survey contains numbers with more explanatory power.

3.8% of Japanese companies say their DX talent is adequate by quality, against 52.9% in the US and 25.1% in Germany. And the most lacking role in Japan is business architect, where the US and Germany both name data scientist.

A business architect decides what gets built and designs where it enters the existing operation.

Set that against the 2026 white paper. 86.4% of Japanese companies use generative AI. 27.0% report no organisational effort. Between 6.0% and 11.9% have redesigned a process around it.

The technology is in the building. Nobody is deciding.

Why the conversation defaults to legacy

The legacy explanation is convenient in three ways.

Responsibility does not attach to a person. A system being old is not anyone's failed judgement.

The remedy looks clear. Replace it. Budget and timeline can be estimated.

And executing it changes no structure. Running the same process on newer technology alters nobody's role.

The business architect explanation is harder to work with. It raises the question of who decides, which means touching the existing allocation of authority.

Questions that settle it

You can test whether legacy is really your constraint.

If every system were replaced tomorrow, what would change. Write it out concretely. If the sequence of work stays the same, the constraint is not the system.

Name three stalled initiatives and write down why each is stalled. Technically impossible, or undecided. Usually the second.

Check why recently cancelled initiatives were cancelled. What share ended for technical reasons.

Those three locate the constraint fairly reliably.

Legacy is real. But treating something that fewer than 20% of companies consider a major obstacle as a company wide priority needs a better justification than habit.

A question of order

This is not an argument against modernisation. Ageing systems carry genuine cost and risk.

But replacing a system without first redesigning the process gives you the same operation running on newer infrastructure. A few years later the same discussion returns in the same words.

Deciding comes first, rebuilding comes second. What the IPA numbers show is that Japanese companies are short of the first one, not the second.