Lesson 6 of 6
From findings to decisions and KPIs
9 min
Listen to this lesson
Signavio is only worth its cost if findings change something: a fit-to-standard decision, a KPI with a target, a fixed control, an updated model, a better test scope. This lesson gives the process owner a routine for turning findings into decisions, connecting the suite to sap-pow-202 (fit-to-standard) and sap-pow-302 (KPIs).
Classify each finding
| Finding class | Example at Contoso (illustrative) | Typical decision |
|---|---|---|
| Master data | Rotor price differences from outdated purchasing info records | Data correction plus a data quality rule |
| Configuration | Tolerance for price differences tighter in 1710 than 1810 | Align configuration through change control |
| Behavior and training | Purchase orders created after the invoice | Training and a KPI on after-the-fact orders |
| Design gap | Approval skipped above threshold for one cost center | Fix workflow or authorization; add control test |
| Opportunity | Innovation recommendation for invoice automation | Evaluate in fit-to-standard; plan and test |
Each class of finding leads to a different kind of decision.
Findings in fit-to-standard
In fit-to-standard workshops, mining findings replace opinions with numbers. When a key user says "we always need a custom step here", the owner can ask how many cases actually take that path. For Wave 2, when Contoso brings Alternadores del Norte into company code 1910, Lucía's order-to-cash workshops face a gap: Alternadores runs a non-SAP legacy ERP, so Process Insights, which is built for SAP data, cannot analyze it. Process Intelligence could, if its event data is extracted, but that is a license and effort decision. The cheaper evidence is Contoso's own SAP S/4HANA order-to-cash variants, which show which standard paths already serve a similar business, plus the legacy system's own reports on order volumes and exceptions.
- Bring the top variants and their volumes to the workshop, not the full map.
- For each requested deviation from standard, ask for the share of cases and the value it affects.
- Record decisions with the evidence that supported them, so they can be revisited.
- Mark findings that become requirements, so they can be traced to tests in SAP Cloud ALM.
Findings also protect the standard. When a team asks for an extension to handle "a common exception", the owner can check whether the exception is common at all, and whether the standard path would handle most cases if a data or configuration problem were fixed first. Many requested extensions turn out to be workarounds for a finding that has a cheaper fix, which keeps the core clean, as sap-pow-202 recommends.
From findings to KPIs
| KPI | Baseline (illustrative) | Target | Measured in | Owner |
|---|---|---|---|---|
| Supplier invoices posted without manual intervention, 1710 | 58% | 70% within two quarters | Process Insights, then embedded analytics | Dan |
| Invoices paid after due date, 1710 | 9% | Below 5% | Process Insights or SAP S/4HANA analytics | Dan, with Ana's team |
| Invoices above threshold without approval | 14% of those cases | 0% | Conformance, then a control test each quarter | Dan and Nadia |
Illustrative course data: KPIs that came from Dan's findings, each with baseline, target, source and owner.
Value analysis in Process Insights and volumes from mining give the value axis; the solution architect and process experts give the effort axis. A finding that is valuable and cheap is done in the next release. A valuable but expensive one becomes a roadmap item with a business case. Recording the prioritization, with numbers, protects the owner when someone asks later why one improvement came before another.
The loop closes in three places that other courses cover. The testers use the same variants to scope regression (sap-tst-311). The run team monitors the operational KPIs in SAP Cloud ALM (sap-alm-304). And the model in the Collaboration Hub is updated whenever a decision changes how the process should work, so the next person who reads it sees the improved process, not the old one.
What would you do?
Process Insights shows that 9% of Detroit supplier invoices are paid late. The AP team blames late approvals; purchasing blames missing goods receipts. Dan has one meeting to decide. What should he bring?
- Validate, then classify findings: data, configuration, behavior, design or opportunity.
- Use variant volumes and values to decide deviations in fit-to-standard.
- Turn findings into KPIs with baseline, target, source and owner.
- Close the loop through tests, monitoring and updated models.
Check your understanding · not graded
Lesson check
1. What should a fit-to-standard workshop bring from mining?
2. Which finding class calls for a control test?
3. When is a finding closed?
4. What should happen to the model when a decision changes the process?