I Ran 4 Customer Advisory Boards. 3 Turned Into Wishlist Sessions. Here's What Fixed It.
I used to think Customer Advisory Boards were about asking customers what they wanted. That's exactly why my first three sessions at Sonic Linker and Finvestfx were disasters.
The pattern was always the same. I'd send out invites to power users, book a conference room (or Zoom link), prepare a deck with our roadmap, and show up ready to "gather feedback." Thirty minutes in, I'd have a Google Doc with 47 feature requests, zero clarity on what mattered, and a sinking feeling that I'd just wasted everyone's time.
The problem wasn't the customers. It was me. I was asking open-ended questions without giving them context to push back.
Stop Asking What They Want. Start Asking What They'll Trade Off.
The shift happened when I ran a CAB at Finvestfx with 8 treasury managers from mid-sized exporters. Instead of showing up with "what features do you need?" I brought three specific product bets we were considering:
- Real-time FX rate alerts via WhatsApp
- Automated hedge ratio recommendations
- Multi-currency cash flow forecasting
But here's the key, I told them we could only ship one in the next quarter. And I made them vote. Not just thumbs up/down, but rank them and explain why.
Suddenly the conversation shifted. Instead of "we want all of this," it became "if you're only doing one, here's why #2 matters more than #1 right now." One treasurer said WhatsApp alerts sounded cool but didn't solve the actual problem, which was that their teams couldn't model hedge scenarios fast enough during volatile sessions. That's the insight I needed.
We shipped the hedge ratio tool first. Retention for that cohort went up 18% over the next two quarters.
The trick is forced trade-offs. Customers will always say yes to free features. But when you make them choose, they reveal what actually hurts.
Bring Half-Baked Ideas, Not Polished Roadmaps
At Sonic Linker, I made the mistake of presenting a polished product deck in an early CAB. It looked professional. It also killed the conversation. When everything looks done, customers assume you've already decided. They won't push back.
So I started bringing wireframes that looked intentionally rough. Sketches. Figma prototypes with placeholder text and broken flows. I'd say upfront, "this is 40% baked, and I need you to tell me if we're solving the wrong problem."
In one session, I showed a workflow for auto-generating LinkedIn carousels from blog posts. I thought it was clever. Two customers immediately said, "this feels like a feature for PMs, not for the sales teams who'd actually use this." They were right. We pivoted to auto-generating prospect research summaries instead, which got 3x more usage in the beta.
The rougher the artifact, the more honest the feedback. Polish signals confidence. Rough signals you actually want their input.
Set a Pre-Mortem, Not Just Objectives
This one I learned from a founder I worked with at Stampede Capital. Before every CAB, we'd run a 10-minute pre-mortem. Not "what do we want to learn," but "imagine this session bombs, what went wrong?"
Usually the answer was: we didn't frame the problem clearly enough, or we invited the wrong people, or we let one loud voice dominate.
So I started being way more deliberate. For a CAB on improving Finvestfx's onboarding, I didn't just invite power users. I invited two customers who'd churned and one who was still stuck in trial after 60 days. I told them upfront, "you're here because onboarding didn't work for you, and I need to know why."
That session was uncomfortable. But it was useful. Turns out our onboarding assumed users knew what a hedge ratio was. Most didn't. We added a 2-minute explainer video and saw trial-to-paid conversion jump 14%.
Pre-mortems force you to admit what you're afraid of. And that's usually where the real insights are.
The Real Output Isn't Features. It's Conviction.
Here's what I tell PMs now: a good CAB doesn't end with a list of features. It ends with you being able to say, "I'm confident we should build X instead of Y, and here's the customer evidence."
After my fourth CAB at Sonic Linker, I killed a feature we'd been planning for weeks. Three customers said it solved a problem they had once a quarter, not once a week. That's not a roadmap item. That's a nice-to-have.
The goal isn't to make customers happy in the room. It's to get the signal you need to make a decision you can defend. Sometimes that means saying no to things they want. Sometimes it means building something they didn't ask for but you now know they need.
The CAB is just the forcing function. The real work is what you do after, when you turn those messy conversations into a roadmap that doesn't try to please everyone.