Lesson 5 of 6
Managing partner performance
9 min
Listen to this lesson
A good contract makes performance measurable; it does not make it happen. Partner performance is managed through a rhythm of reviews, independent checks of quality, honest joint retrospectives and a clear escalation path. The aim is a partner that improves every quarter, not one that is punished every month.
The review rhythm
| Forum | When | Who | Focus |
|---|---|---|---|
| Daily overlap | Each working day | Process twins, partner service lead | Failures needing judgment, blockers, handover gaps |
| Monthly service review | Monthly | Priya, partner service lead, Sam, Theo, Nadia | Service levels, credits, sampled evidence, improvement backlog |
| Release retrospective | After each release | TCoE and partner delivery team | What worked, what escaped, what changes next release |
| Quarterly business review | Quarterly | Priya, partner delivery executive, Contoso contract owner | Scorecard trend, improvement plan, staffing, proposals |
Contoso's partner governance rhythm. The quarterly business review with the partner feeds the quality council described in sap-tcoe-406.
Sampling evidence, properly
The course on test teams (sap-tst-310) showed why passed results need spot checks: Priya sampled 20 passed tests and found seven with no evidence beyond a status. For a permanent service, sampling becomes a rule rather than a reaction. Contoso samples each regression cycle: 5% of passed tests, at least 20, weighted toward critical processes, chosen by the TCoE rather than the partner. Each sample is checked against the evidence standard: system and client, date, data used, document numbers, expected and actual result. The service level is 95% complete. If a cycle falls below it, the next cycle's sample doubles, and it returns to normal after two good cycles, an idea borrowed from acceptance sampling, where inspection tightens when quality slips.
Retrospectives that are joint and blameless
After each release, the TCoE and the partner hold one retrospective together. It is not a performance review; it asks what the system of work did. When an escape is traced to a missed test, the useful question is why it was missed: was the scenario outside the agreed scope, was the expected result unclear, was the data wrong, or was the test skipped under time pressure? Each answer points to a different fix, most of them shared. Partners who are blamed for every escape learn to hide problems; partners who are part of the fix report them early.
The partner scorecard
A scorecard only helps if it is read the same way every quarter. Contoso's scorecard for the second quarter of the managed service shows what a typical "compliant but not yet good" partner looks like: service levels mostly met, evidence slipping, and little improvement initiated by the partner itself.
| Layer | Measure | Q2 result | Trend | Reading |
|---|---|---|---|---|
| Service levels | Regression cycle time (year-1 level 10 days) | 9 days | Improving | Met |
| Service levels | Automation repair turnaround, critical (3 days) | 4 days in one cycle | Flat | Missed once; credit applied, earn-back possible |
| Service levels | Evidence quality (95%) | 91% | Worsening | Missed; sampling tightened to 10% |
| Outcomes | Attributable severity 1 and 2 escapes | 1 severity 2 | n/a (component starts year 2) | Tracked, not yet paid on |
| Improvement | Partner-initiated proposals delivered | 0 of 0 | Flat | Ask for two proposals next quarter |
| Relationship | Key personnel continuity | No unplanned changes | Stable | Good |
Contoso's partner scorecard for one quarter. Illustrative course data.
Escalation without drama
- Can the issue be fixed by the people doing the work?Yes → resolve in the daily overlap · No → next step
- Is a service level or quality standard at risk this month?Yes → monthly service review with an action and date · No → improvement backlog
- Has the same miss repeated after an agreed action?Yes → partner delivery executive and a written improvement plan · No → track to closure
- Has the improvement plan failed by its dates?Yes → contract governance: remedies, step-in or termination clauses · No → close the plan
What would you do?
Two cycles in a row, sampled evidence for the partner's regression is 88% complete. The partner says the cause is time pressure in the last two days of each cycle. What should Priya do?
- Manage performance through a rhythm: daily overlap, monthly service review, release retrospective, quarterly business review.
- Sample evidence every cycle, chosen by the TCoE and weighted to critical processes; tighten sampling when quality slips.
- Run joint, blameless retrospectives; ask why a test missed, not who missed it.
- Use a written escalation ladder and never quietly redo partner work.
Check your understanding · not graded
Lesson check
1. Who should choose the sample of passed tests to check?
2. What happens to sampling after a cycle falls below the evidence standard?
3. What is the "shadow team" anti-pattern?
4. An escape is traced to a test that was skipped under time pressure. What question should the joint retrospective ask first?