In tax firms, the client base often grows over decades. New mandates come in, firms merge, legal forms change, clients get created more than once. As long as operations run locally, this usually stays manageable, even if it is tedious.
With the upcoming move to DATEV cloud applications, this changes fundamentally. What used to „somehow work" now has to be unambiguous, consistent and machine-readable.
That turns the client base from a background topic into the decisive success factor of the migration.
Why client consolidation is now a strategic topic
1. Historically grown data is the norm
In practice, we regularly see the following in larger firms:
- the same client created more than once
- different client numbers for different legal forms
- parallel master data from earlier firm structures
- inconsistent spellings, addresses and assignments
From a professional point of view, much of this is understandable. Technically, however, it means that DATEV „sees" the same people or companies more than once.
As long as data is maintained locally, you can deal with it. In the cloud, however, this leads to: faulty transfers, duplicate or contradictory master data, high manual rework effort and risks in reporting.
2. DATEV migrates data, but not firm logic
DATEV provides tools and migration paths for the move to the cloud. What DATEV cannot do:
- professionally merging historical clients
- assessing duplicates or similar clients
- deciding which client number is „leading"
- dissolving legacy structures from mergers
These decisions always rest with the firm.
The typical mistake: „we will clean that up later"
Many firms keep putting the topic off because:
- day-to-day operations take priority
- no one knows exactly how big the problem is
- nobody wants to make risky changes in DATEV
The problem with that: after the migration, cleaning up does not get easier, it gets significantly more expensive. By then the data is already in the cloud, processes are built on it and corrections come with a higher coordination effort.
The right approach: structured preparation instead of a permanent manual building site
A proven path is to clean up the client structure in a targeted way before the DATEV cloud migration, without disrupting day-to-day operations.
The basic idea is simple: the machine does the legwork, the firm makes the professional decisions.
The technical process, explained clearly
1. Reviewing the client master data
First, relevant client master data is exported (e.g. name, address, tax number). There is no access to ongoing bookkeeping and no intervention in DATEV production systems.
Goal: all clients are available in a comparable, neutral structure.
2. Automatic detection of possible duplicates
On this basis, clients are identified that:
- have identical or very similar master data
- use the same tax number or VAT ID
- probably belong together despite different spellings
This is not about decisions, but about flagging anomalies.
3. Reasoned preliminary assessment
Each detected case is rated: „very likely the same client", „possible, please check", or „probably not the same client".
Important: every assessment comes with a traceable justification (e.g. same numbers, same address, high similarity).
4. Grouping clients that belong together
Often it does not concern just two records, but whole groups: several client numbers, several legal forms, different time periods. These are presented as client groups that belong together.
5. Professional decision by the firm
The firm receives a clear review list:
- Which clients should be merged?
- Which should not?
- Where is clarification needed?
The professional decision remains entirely with the firm.
Conclusion
The DATEV cloud migration is more than an IT project. It is the moment when historically grown structures meet modern requirements.
Those who systematically clean up their client base before the migration save:
- time during the migration
- costs for rework
- risks to compliance and data quality
The best preparation is structured groundwork.

Author
Joyce Marvin Rafflenbeul
Founder & AI Engineer
Joyce has been building production systems for the enterprise space for over 5 years. As founder of QUIKK Software, he focuses on RAG architectures & AI agents.
LinkedInAbout QUIKK Software
AI engineering studio from Minden
We build production-ready AI systems with a focus on RAG, for the German-speaking Mittelstand.
Book a consultation