Mentoring Program Evaluation: Framework, KPIs & Timing
Mentoring program evaluation is the structured check of whether a mentoring program is running as planned and moving the goals it was set up for. You track participation, the quality of the relationships and a few outcome indicators, at fixed points from baseline to follow-up, then use the findings to improve the next cycle.
Most programs collect something. Few collect the right things at the right time. This guide gives you a framework, a KPI set you can copy, a measurement calendar and the traps that make evaluation look better or worse than the program really is.
Why evaluate a mentoring program at all?
Three reasons, in this order:
- To fix the program while it runs. A midpoint check that shows a third of pairs have stopped meeting is worth more than a polished final report.
- To decide what to keep. Which format, matching rules and session structure worked well enough to repeat next cohort?
- To answer the sponsor. Leadership, a board or a funder will ask what the program achieved. Evaluation gives you an honest answer instead of anecdotes.
If the third reason is the only one driving your evaluation, it tends to become a reporting exercise that nobody uses. Start with the first two.
What framework should you use?
The most widely known model is the four-level framework named after Donald Kirkpatrick, originally designed for training. It maps well onto mentoring because it separates what people felt from what actually changed.
| Level | Question for a mentoring program | Typical evidence |
|---|---|---|
| 1. Reaction | Did mentors and mentees find the relationship worthwhile? | Post-meeting feedback, end-of-program survey |
| 2. Learning | What did mentees learn: skills, knowledge, confidence, network? | Self-assessment against goals at baseline and end, mentor observations |
| 3. Behavior | Do mentees act differently in their work or study? | Goal progress, examples, manager or tutor input, follow-up survey |
| 4. Results | Have the indicators the program was built for moved? | Retention, internal mobility, readiness, academic progression |
Two practical rules:
- Do not stop at level 1. High satisfaction is pleasant and easy to collect, but it does not tell you whether anything changed.
- Do not jump to level 4 alone. Results data without levels 2 and 3 cannot explain why an indicator moved, or did not.
Which KPIs should a mentoring program track?
Pick a small set from each group below. The participation group tells you whether the program is happening, the quality group whether it is worth people’s time, and the outcome group whether it is pointing toward its purpose.
| Group | KPI | How to calculate or collect it | What it tells you |
|---|---|---|---|
| Participation | Applications | Number of mentor and mentee applications per cohort, against target | Demand and the balance of supply |
| Participation | Match rate | Matched participants ÷ eligible applicants | Whether supply, rules or process left people out |
| Participation | Meetings held per pair | Meetings recorded ÷ active pairs, against the planned number | Whether relationships are actually running |
| Participation | Drop-off | Pairs that stopped meeting or left before the end ÷ all pairs | Where the program loses people, and when |
| Quality | Satisfaction | Rating plus one open question, after meetings and at the end | Perceived value, early warning signs |
| Quality | Goal progress | Mentee self-rating on each goal at baseline, midpoint and end | Whether mentoring is working on what was agreed |
| Outcome indicator | Retention | Share of mentees still in the organization after 6 or 12 months, alongside a comparison group | A signal worth tracking, not proof of effect |
| Outcome indicator | Promotion or internal move | Moves within the follow-up period, alongside a comparison group | Career movement connected to the program’s purpose |
| Outcome indicator | Readiness | Manager or self-rated readiness for a next role | Pipeline for leadership or succession programs |
| Outcome indicator | Progression (students) | Continuation to the next year, completion, or move into work or further study | The student-program equivalent of retention |
Treat every row in the outcome group as an indicator to track, not as a result the program has caused. Many things affect whether someone stays or gets promoted; mentoring is one of them.
A useful habit: write the target next to each KPI before the cohort starts (“at least four meetings per pair by the midpoint”). A number without a target cannot be read.
When should you measure?
Evaluation is a calendar, not an event at the end. Use four points:
- Baseline (before the first meeting). Mentees rate themselves on each goal, plus a few items on confidence, network and intention to stay or continue. Without this, end-of-program numbers have nothing to be compared with.
- Midpoint. A short pulse: is the pair meeting, is it useful, is anything blocking it? This is the moment to rematch or intervene, not to write a report.
- End of cycle. The full survey for mentors and mentees, goal re-rating, meeting totals and completion.
- Follow-up (three to twelve months later). Behavior and results take time. A short follow-up survey and a look at outcome indicators belong here.
Ready-made wording for each point is in the mentoring evaluation survey questions template.
Where does the data come from?
Use three sources and say which one each number comes from:
- Surveys. Short, consistent and repeated at the same points, so answers can be compared over time. Ask the same goal questions at baseline and end. Keep each survey short enough to finish in a few minutes.
- Session and meeting records. Dates, attendance, sessions completed and post-meeting feedback. These are behavioral data, not opinions, which makes them the most reliable participation evidence you have.
- HR, academic or membership records. Retention, internal moves, progression. Handle these with care: agree access with HR or the data owner, tell participants at enrollment what will be used and why, report only aggregates, and suppress groups so small that individuals could be identified. Under GDPR in Europe, KVKK in Turkey and similar laws elsewhere, data minimization and a clear purpose are not optional.
How do you compare results fairly?
Every evaluation eventually meets the question: would this have happened anyway?
Simple before-and-after comparisons and “mentored vs not mentored” tables are tempting, but they are weak evidence. People who sign up for mentoring are often already more motivated, better connected or more likely to stay. Comparing them with everyone else can make the program look better than it is.
Ways to get closer to a fair comparison:
- Matched comparison group. Compare mentees with non-participants who are similar in role, level, tenure or year of study.
- Waitlist design. When demand exceeds mentor supply, people on the waitlist form a natural comparison group for one cycle.
- Phased rollout. Launch in some units or campuses first and compare with the ones that start later.
Even with these, report findings as associations (“mentees showed higher 12-month retention than a matched comparison group”) rather than causes (“mentoring increased retention”), unless the design genuinely supports a stronger claim. A sponsor will trust a careful sentence more than a bold one that falls apart under a single question. When the conversation turns to money, the mentoring program ROI guide shows how to put costs and outcomes side by side without overclaiming.
What are the common evaluation mistakes?
- No baseline. You cannot show change if you never measured the starting point.
- Only satisfaction. A program can be well liked and change nothing. Pair satisfaction with goal progress and behavior.
- Too many KPIs. Twenty metrics loosely tracked are worse than six tracked well. Every KPI should have an owner and a target.
- Measuring only at the end. By then, the pairs that stopped meeting in week three are long gone. The midpoint check is where evaluation pays for itself.
- Claiming causation. Retention and promotion are indicators to watch, not effects you have proven.
- Ignoring drop-outs. Surveying only people who finished inflates every number. Ask those who left why, briefly.
- Collecting data nobody reads. Decide in advance who will receive the findings and what decision they will feed. The how to report mentoring results guide covers turning the numbers into a report leadership will read.
Evaluation is one part of running the program well; the rest of the cycle is in how to manage a mentoring program.
How can software help with evaluation?
If you run the program in a spreadsheet, most of the participation data has to be chased by email. In Mentornity, every meeting booked through the platform is recorded with its date, participants and session, and admins see which pairs have met, which are behind on their sessions and which have gone quiet. Sessions can carry questions before and after each meeting, and feedback forms are part of the program, with reminders going to anyone who has an open form or missing feedback. Reports export as charts, Excel or PDF, alongside a Program Health score updated every week. See how this works for employee mentoring programs, or try it free with up to 10 users.
For more on designing and running programs, browse all mentoring guides.