ERP process mapping: what to map before the implementation starts
Somewhere early in every ERP project, the business is asked to document its processes. Done well, it shortens the workshops and gives the integrator something real to work from. Done badly, it produces hundreds of pages describing the old system in detail, which the project then spends months trying to recreate. Here is what is worth mapping, at what level, and what to skip.
Why map before the project starts
The implementation partner will run their own design or fit-to-standard workshops. Your maps are not a replacement for those; they are what makes them productive. A team that arrives with its main processes already drawn spends the workshop deciding things. A team that arrives without them spends the first two sessions explaining how the business works.
There is a second benefit that is easy to miss. Mapping your own processes forces the internal disagreements into the open before the consultants are billing by the day: who actually approves a purchase above the limit, whether credit holds are released by sales or finance, which of the three spreadsheets is the real price list.
What to map
Map the end-to-end processes that carry the most money and the most handovers. For most companies that means:
- Procure-to-pay: requisition to supplier payment, including how you handle invoices without a purchase order.
- Order-to-cash: customer order to cash collected, including credit holds, returns and disputes.
- Record-to-report: the month-end close, with the real timetable.
- Plan-to-produce or inventory processes, if you make or hold stock.
- Any process that is unusual for your industry and that you suspect the standard system will not handle.
How much detail is enough
ERP projects often talk about process levels: level 1 is the end-to-end process, level 2 its main phases, level 3 the individual steps, and level 4 the screens and fields. Before implementation, stop at level 3. A swimlane map with twenty to thirty steps per end-to-end process is enough.
Below that, you are describing how the old system works, which is exactly what you are replacing. The integrator does not need to know which tab of the legacy screen holds the delivery date; they need to know that a delivery date is promised at order entry and who is allowed to change it.
- Do capture: who does each step, the decisions and their real rules, handovers between teams, exceptions that happen every week, approximate volumes.
- Do not capture: screen layouts, field lists, legacy report names, workarounds you already plan to stop.
Map the exceptions, not just the happy path
The happy path is what every ERP does out of the box. What differs between companies, and what drives most of the cost of an implementation, is the exceptions: partial deliveries, price differences on invoices, rush orders, customer deductions, intercompany mismatches. Draw the ones that happen every week as real branches with their own steps. Leave rare ones as a note on the step where they occur.
Add rough volumes as you go: invoices per month, share of orders on credit hold, number of manual journal entries at month end. Volumes tell the integrator which exceptions are worth automating and which are worth a manual workaround.
Use a spreadsheet as the source
Process owners are rarely diagram tool experts, and they should not need to be. A spreadsheet with one row per step (step, owner, next step, decision label) is something every key user can fill in, and it turns into a swimlane map without anyone drawing boxes. It is also the format the implementation team can import into their own tools later.
Start from a template rather than a blank page. Even if your process differs, editing a realistic map is much faster than inventing the structure, and it shows key users the level of detail you are asking for.
Hand over to the implementation
Once the integrator starts, your maps become the input to fit-gap: each step is compared with how the new ERP works, and each difference is classified as a fit, a configuration, a process change or an extension. That stage needs version control and sign-off by process owners, because its output sets the scope of the build. QueryChart, our sister product, is built for that stage and has system-specific flows for the major ERPs:
Frequently asked questions
Should we map the as-is or the to-be process before an ERP implementation?
Mostly the as-is, but lightly: enough to understand who does what, the real decision rules, the frequent exceptions and the pain points. The to-be process is best designed during the implementation, starting from how the new ERP works by default, rather than designed in advance and then forced onto the system.
Who should map the processes?
The key users who do the work every day, with the process owner reviewing. They know the exceptions and the workarounds. An external consultant can facilitate, but maps drawn without the people who do the work describe the process as it is supposed to run, not as it does.
How long does pre-implementation process mapping take?
For a mid-size company, two to six weeks of part-time work for the main end-to-end processes, typically one or two workshops per process plus time for key users to fill in and correct the maps. Starting from templates shortens it considerably.
Try it on your own process
The editor is free and opens with a starter template. One row per step — the swimlane diagram renders as you type. No signup required.
Start Process Mapping Free