DATEV Simulation: Running Scenarios on German Accounting Data (2026)

A DATEV simulation means running what-if scenarios on the bookkeeping data that German companies already produce in the DATEV ecosystem. In practice it starts with an export of the trial balance (Summen- und Saldenliste), from which an integrated projection of P&L, balance sheet and cash flow can be derived. Drivers and measures are then layered onto that base model, for example a 10 percent drop in revenue or a two point rise in interest rates, and their effect is traced through to liquidity. What separates this from a conventional planning round is not the arithmetic but the repeatability: once the model sits on the DATEV data, changing a single assumption costs minutes rather than days.

Some context for readers outside Germany: DATEV is a cooperative owned by tax advisors, auditors and lawyers, and it is the de facto accounting standard for the German Mittelstand. As of 30 June 2026 it reported around 1.01 million customers and 40,176 cooperative members on revenue of 881.0 million euros. Whoever wants to plan on German actuals is, in most cases, planning on DATEV data.

That is also where the gap sits. Bookkeeping runs reliably, often through the tax advisor, and produces clean actuals. But as soon as a material assumption changes, because a key account cuts volume, material prices rise or the bank asks about debt service capacity, there is no place to recalculate that change consistently. What emerges instead is a spreadsheet with a handful of assumptions: quick to build, hard to defend when management asks how the number was derived.

The bottleneck is the calculation, not the bookkeeping

In most Mittelstand companies, account-level actuals are available every month, reconciled and reliable. What is missing is the layer above them: a model that turns those actuals into an integrated projection and lets a single assumption be changed without rebuilding the plan.

That is not a gap so much as a division of labour. DATEV provides the reliable data foundation and, with DATEV Analyse und Planung on the advisory side, analysis and planning built on top of it, explicitly as a basis for the bank conversation. That is what those tools are for, and they do it. Repeated simulation is a different job on a different cadence: it sits on top of that foundation rather than replacing it.

The need arises between two planning dates. Not during the annual planning run, but when an assumption breaks in March and management needs a number it can stand behind by Friday.

DATEV demonstrates what that additional layer looks like in its own house. To steer its shift from an on-premise product into the cloud, DATEV’s controlling function maps several perspectives onto the same data cube and runs the drivers behind costs, prices and revenue as scenarios, as described in the DATEV customer success story. The trigger differs from that of a smaller company, but the principle is identical: reliable actuals carry a driver logic, and variants are calculated on top of it.

The building blocks available around DATEV

This is not a ranking. A company with 8 million euros in revenue and no controlling function is often well served by its tax advisor plus a planning application. The need shifts from “plan once” to “recalculate repeatedly” only when new assumptions arrive several times a year.

The starting point: the trial balance

The Summen- und Saldenliste is the practical data foundation because it is account-level and available monthly. There are two routes to it:

Two details decide whether the file is usable: the export should cover the account range 1 to 99999, and it should be CSV rather than PDF. For a first model build, pull at least two closed fiscal years, then add monthly exports for the current year. If the accounting reports area is not visible, the tax advisor has to enable it first.

How does a simulation on DATEV data actually work?

A defensible simulation comes together in four steps:

1. Structure the actuals. Trial balance accounts are mapped onto a planning structure: revenue categories, material costs, personnel costs, other operating expenses, depreciation, interest. This mapping is one-off work and is the real substance of the model.
2. Add driver logic. Line items are calculated rather than extrapolated: revenue as volume times price, material cost as a material ratio, personnel cost as headcount times average cost. The result is a driver-based model instead of a collection of isolated assumptions.
3. Project on an integrated basis. P&L, balance sheet and cash flow are linked: revenue flows through payment terms into receivables, capex flows through depreciation into the P&L and through disbursements into liquidity. Only this integrated financial planning turns an earnings projection into a statement about solvency.
4. Run scenarios. Variants are layered onto the base model. What matters is that a changed assumption propagates automatically to every affected line item and year, instead of being reconciled across several worksheets.

What matters is how the effort is distributed: steps 1 to 3 happen once, step 4 happens continuously. Rebuild the model every time and you do not have a simulation tool, you have a planning file. If the structure and the integrated projection arrive already prepared, the work shifts entirely to step 4, and the recurring “what happens if” question comes down to changing one value.

The difference between step 3 and a P&L-only calculation, incidentally, is exactly the difference that counts in a bank meeting. A decline in earnings is uncomfortable; a breach of debt service capacity is a different conversation.

Three simulation questions that actually get asked

These questions rarely arrive on the planning calendar. They arrive between planning dates, which is why speed decides the outcome: an answer that takes three days gets replaced by a guess. How that turns into a full-year projection is covered in the article on building a year-end forecast.

Why Excel runs out of road

Excel is the obvious place for a first attempt, and for a single scenario it works. According to the BARC Planning Survey, the largest global user survey on corporate planning, 90 percent of companies say they plan with Excel. 47 percent use scenario simulation and around 40 percent use driver-based approaches. The gap between “we calculate in Excel” and “we simulate in a structured way” is substantial.

The failure modes are well known: scenarios exist as file copies and drift apart, the balance sheet is dropped because linking it is laborious, and after three months nobody remembers which assumption sits in which cell. We have covered how far spreadsheets carry and where they stop for scenario analysis and the rolling forecast.

What separates a defensible simulation from a gut number

Three criteria that hold regardless of tooling:

• Traceability. For every figure, you can name the driver it comes from and the assumption behind it. When management asks how you got there, the answer fits in one sentence.
• Integrity. P&L, balance sheet and cash flow stay internally consistent after every change. A simulation whose balance sheet does not close is not one.
• Repeatability. The next change costs minutes, not days. Only then does anyone actually simulate rather than plan once a year.

Outlook

The preconditions are widely in place: for a large share of German companies, structured account-level actuals already exist every month, with no additional data capture required. The interesting development is therefore not data integration but preparation. The more that chart-of-accounts mapping, driver logic and integrated projection come standardised out of the box, the shorter the distance between a DATEV export and a first defensible simulation. A planning project then becomes a question of hours, and every further question a matter of minutes. The prerequisite stays the same: explicit driver logic, because even an AI assistant can only answer what a model is able to calculate.

Frequently asked questions about DATEV simulation

Can you simulate scenarios inside DATEV itself?

DATEV Analyse und Planung produces an integrated P&L, balance sheet and cash flow plan and is designed around the tax advisory firm’s process. For repeated what-if calculations with varying drivers, companies in practice supplement that foundation with Excel or a simulation platform. The two layers are not in competition: one supplies the agreed planning view, the other the variants on top of it.

Which DATEV report works as a data foundation?

The trial balance (Summen- und Saldenliste), because it is account-level and available monthly. It can be exported as CSV from the accounting reports area of DATEV Unternehmen online, or provided by the tax advisor from DATEV Kanzlei-Rechnungswesen. The BWA is useful for reporting but too aggregated for building a model.

How current does the DATEV data have to be?

Closed fiscal years are enough to build the model structure. For an ongoing forecast, the last closed month is the relevant position, because the actuals of the elapsed months form the fixed part of the full-year projection. A monthly export is therefore sufficient; weekly currency is only needed for short-term liquidity monitoring.

How long does it take to build a simulation model on DATEV data?

The effort sits in mapping accounts onto a planning structure, not in the calculation. For a company with a manageable chart of accounts, that is a task of hours to a few days. After that, an additional scenario costs minutes, because it is only a change to a driver value.

Do I need a controlling department for this?

No, but you need someone who can name the drivers of the business. The substantive question of what actually causes revenue and the main cost blocks cannot be delegated to a tool. The arithmetic part, the integrated projection of P&L, balance sheet and cash flow, can largely be automated.