A customer advisory board is a small group of your best customers who meet with you every quarter to react to plans before you commit to them. This SOP sets one up in two to three weeks and runs it for a year at 90 minutes per quarter. Done well, the board gives you a standing panel of people who live with the problem every day and will tell you when a plan misses. Done badly, it turns into a monthly sales pitch that customers stop attending by the second session.

What is a customer advisory board and when should you use it?

A customer advisory board (CAB) is a fixed group of 8 to 12 customers who agree to meet regularly, share honest views on your direction and talk with each other as peers. It differs from its neighbours in time and commitment:

A focus group is a one-off session with people who may not be customers.
Customer co-creation means building a feature together with customers.
A board meets repeatedly over a year, with the same members, about direction and priorities.

The peer element is what makes boards work. Members hear how other companies handle the same problem, and that conversation tends to surface needs that never come up in one-on-one calls.

Use this SOP:

When you sell to businesses and a handful of accounts shape your revenue.
Before big bets: a new product line, a pricing change, a move upmarket.
When feature requests conflict and you need customers to weigh trade-offs together.
When you want deeper relationships with accounts you can't afford to lose.

What do you need before you start?

At least 20 customers to choose from, so you can pick strangers who fit and leave friends out.
A list of your best-fit accounts, ideally those who answered "very disappointed" in a Sean Ellis test or score green in your customer success metrics.
Two or three questions you need answered this year.
A small budget for thank-you gifts, a shared dinner or travel if you meet in person.
A founder who will attend every session. Delegating the board to a junior hire tells members it doesn't matter.

How do you run it, step by step?

1. Write a one-page charter

State the purpose, the two or three questions for the year, the term (12 months), the commitment (four 90-minute meetings), what members get (early access, a voice in direction, a peer network) and the confidentiality rules. A clear charter is what makes a busy person say yes.

2. Choose members on purpose

Pick 8 to 12 people who make or influence buying decisions, from your best-fit segment, with a mix of company sizes and use cases. Leave out your single biggest account if it would dominate the room, and leave out prospects: the board is for customers.

3. Invite each member personally

Send a personal email from the founder with the charter attached, then follow up by phone. Expect about half of the people you invite to accept. Schedule all four dates at the start so they land in calendars.

4. Design each meeting around exercises

Use 10 minutes for a short update, then spend the rest on two structured blocks. Good formats: ranking five options on a Miro board, a peer round where each member shares how they solved a problem, and Luke Hohmann's Buy a Feature game, where members split a fixed budget across features and must negotiate trade-offs aloud.

5. Capture decisions separately from opinions

Assign a note-taker. Record three things: votes and rankings, points where the group agreed, and individual requests. Group agreement is a signal; individual requests are interview leads.

6. Close the loop within two weeks

Send members a "you said, we did" summary: what you heard, what you decided and what you won't do, with reasons. Members who see their input used keep coming.

7. Test conclusions beyond the board

A board of 10 is still 10 people. Check each conclusion with a short survey, a fake door test or a round of interviews with other customers before you commit build time.

Template you can copy

CUSTOMER ADVISORY BOARD CHARTER

Purpose: Help [Company] decide [question 1], [question 2], [question 3]
in the next 12 months.
Members: 8 to 12 [role] at [segment] who use [Product] weekly.
Term: [start month] to [end month]. Four 90-minute video sessions.
Members receive: early access to new features, a direct line to the
founder, a peer network, [gift or perk].
We ask: honest views, attendance at 3 of 4 sessions, confidentiality
about unreleased plans.
Not a sales meeting: we will never pitch upgrades in a session.

MEETING AGENDA (90 minutes)
0:00  Welcome, round of names and one recent win (10 min)
0:10  Founder update: what we did with your last input (10 min)
0:20  Block 1: [exercise, e.g. rank 5 options on Miro] (30 min)
0:50  Peer round: "How did you handle [problem] this quarter?" (25 min)
1:15  Block 2: quick vote on the next question (10 min)
1:25  Close: what we will report back and when (5 min)

NOTES TEMPLATE
Votes/rankings | Group agreement | Individual requests | Follow-up owner

Worked example: an HR tool for restaurant groups

This example is illustrative. A founder sells shift-scheduling software to restaurant groups with 3 to 30 locations and has 64 customers. She invites 16 operations managers and 9 accept. Her question for the year: build a payroll module or deepen scheduling?

In the first session, a Buy a Feature round gives members 100 points each. Payroll gets 180 of 900 points; shift-swap approvals across locations gets 410. The peer round explains why: every member already has a payroll provider they don't want to leave, but swaps between locations happen weekly and live in WhatsApp groups.

She tests the conclusion with a two-question survey to all 64 customers: 41 of 52 respondents rank cross-location swaps above payroll. Over the year, attendance averages 8 of 9, and 3 members agree to case studies.

What should you measure, and when do you stop?

Metric	Pass signal	Warning signal
Attendance rate	75%+ of members per session	Under 50% by the second session
Decisions per meeting	1+ clear decision per session	Sessions end with "interesting discussion" only
Board conclusions confirmed by wider customers	Most conclusions hold in a survey or test	Wider customers disagree with the board
Member renewal rate	Members renew at or above your average	Members churn during their term
Referrals and case studies from members	2+ per year	None

Run the board for its 12-month term, then rotate a third of the members. Stop or rebuild the board if attendance falls under half for two sessions in a row, or if members stop disagreeing with you, because a board that only agrees has stopped being useful.

Which mistakes should you avoid?

Pitching during sessions. The first sales slide is the moment members stop speaking freely.
Letting the loudest member set direction. Use votes and written rankings so quiet members count too.
Inviting only your biggest accounts. Their needs pull the product toward enterprise before you're ready.
Treating board votes as market proof. Ten people are a starting point. Test with the wider base.
Going silent after the meeting. No follow-up within two weeks costs you the next session.

How can AI speed this up?

Record sessions with consent and ask an AI assistant to produce a summary with three sections: votes, group agreements and individual requests, with quotes. Ask it to draft the "you said, we did" email and the follow-up survey for wider customers. Edit both yourself: AI summaries flatten disagreement, and the disagreements between members tend to hold the most useful information.

Where does this SOP fit in your validation plan?

A board fits late in the Product-Market Fit stage, after you know who your best-fit customers are. Before it, run the Sean Ellis test to find them. Related steps:

Turn board rankings into a build order with feature prioritization.
Bring in outside views from people who aren't customers with an expert interview.
Track whether board members reach their goals with customer success metrics.

See all Product-Market Fit SOPs. The F/MS audience and community module covers how to build and keep a group of engaged customers, and the Fe/male Switch startup game lets you rehearse founder conversations before the real board meets.

FAQ

How many members should a customer advisory board have?

8 to 12. Fewer than 8 makes the peer conversation thin; more than 12 leaves too little time for each voice in a 90-minute session.

How frequently should a customer advisory board meet?

Four times a year works for most startups: frequent enough to keep momentum, rare enough that busy members say yes.

Should I pay advisory board members?

Payment is unusual for customer boards. Members join for influence, early access and the peer network. A gift, a dinner or covered travel is a fitting thank-you.

What is the difference between a customer advisory board and a focus group?

A focus group is a one-time session, and participants may not be customers. A board meets repeatedly over a year with the same customers and discusses direction.

Customer Advisory Board SOP