How to Start a Mentorship Program, and How to Grow the Pilot You Already Have

Most guides on this topic assume you have nothing yet. Many readers here already have a pilot. A dozen pairs are meeting, people have asked to join, and the person running it is the same person who set it up. This post covers both starting points. The steps are the same. What changes is how much weight you put on the last one.

What do you decide before anything else?

Four things. Write each one down in a sentence.

Who owns it. One named person. Not a committee. If the owner changes jobs, the program needs a second name on the charter.

Who it is for. Pick one audience. New hires, first-time managers, an employee resource group, or one department. A program for "everyone" has no way to say who gets matched first.

What it is for. One purpose, stated in terms the program can observe. "Give new hires a named person to ask" is observable. "Improve engagement" is not something a mentoring program can show on its own.

Which format. One-to-one is the default. Group, reverse and peer circles are choices you can add later. The mentorship program examples post compares them.

The mentorship program charter template holds all four on one page. It is the shortest document that still gets sign-off.

How do you recruit the first mentors?

Make a specific ask. Say how many hours a month, for how many months, and what the mentor does and does not do. Invite people by name from manager nominations, past mentees and leadership-program alumni. A general call for volunteers brings in the people who have spare time, which is not always the people your mentees need.

The full plan is in how to recruit mentors. Prepare the mentors you accept before their first meeting. The mentor training checklist is a good start.

How do you match the first group?

By hand, with a rule you can write down. Collect each person's goals, skills, seniority and availability on a short form. The mentor matching questionnaire has the questions. Then read the answers and pair people. Start with goals: does the mentor's experience map to what the mentee wants to do?

Record why you made each pair. You will want that note when a pair asks to change, and when you have 40 pairs instead of 10.

Add a check-in at about 30 days. Ask both people if the conversations are relevant, if they can meet at the agreed rhythm, and if they want to continue. A rematch is a normal outcome, not a failure.

What meeting rhythm should you set?

One both people can keep. Monthly is a practical floor to test. Every two weeks works when calendars allow. Put the first meeting on the calendar at kickoff, not "when you can." The meeting cadence post covers what to track.

Give each pair a first-meeting agenda. A first meeting with a shape is easier to follow with a second. The first mentor meeting agenda is free.

What do you tell participants up front?

Tell them what the program is, how long it lasts, and what you will ask of them. Say how often they should meet and who to contact if the pair does not work. Say what you will track, so nobody finds out later. A short kickoff email does most of this, and the kickoff email templates save you the drafting. People who know the rules at the start ask fewer questions in week four.

How long should the first cohort run?

Give it a start date and an end date. Eight to twelve weeks is a reasonable design range to test, then adjust. A time box gives the pair a reason to book the next meeting and gives you a point to review. Time-boxed cohorts explains the tradeoff against open-ended pairs.

What should you count before you add people?

This is the step that matters if your pilot already works.

Take the work each pair creates and write it as a list. For most programs it looks like this: one matching decision, one kickoff message, a reminder before each meeting, a check for quiet pairs, a survey, and a line in the progress report. None of these is hard. Each one repeats with every pair.

Now multiply. If one pair costs you 30 minutes a month in follow-up and you plan to go from 12 pairs to 60, that is a different job. Do the sum with your own numbers.

Signs the sum has gone past one person:

• Matching takes longer than checking the matches.
• Reminders depend on your memory.
• Nobody can say which pairs went quiet last month.

If two of these are true, the pilot has reached its operating limit. That is a good problem. It means people want mentoring. When a mentoring pilot outgrows a spreadsheet walks through the five checks in more detail, and running mentoring on spreadsheets versus MentorStack compares the two honestly.

What can software take over?

The repeated work. In MentorStack that means scoring matches on skills, goals, seniority and availability and showing ranked suggestions for you to review, sending reminders, flagging pairs that have gone quiet, and letting you send a one-click nudge to both people. Reports and export start on the Starter plan.

It does not re-pair people for you, and it does not decide a program's purpose. Those stay with the owner. The scale a mentoring pilot page covers how to move work over in stages, and what employee mentoring software is defines the category.

If you are comparing vendors, mentor matching software: what to look for lists the questions to ask each one.

Where should you start today?

Pick the step you have skipped. If you have no charter, write the one-page version this week. If you have a pilot and no count, do the sum above. If you are not sure where your program stands, the mentorship maturity assessment takes about two minutes and tells you which area to work on first.

The Free Pilot runs up to 10 participants at no cost, with no time limit. It is enough to test the setup with a real cohort.