Survey template

Feature prioritization survey

Ask people to rate features 1-5 and everything comes back a 5. This template forces a choice — what they'd pay for, what they'd drop, the one thing they'd keep — so the priorities are real, not polite.

Best for: PMs and founders deciding what to build next from a crowded backlog.

The questions

1. Of these features, which ONE would you keep if you could only have one?

   Single select

   List your top candidates. The single-choice constraint is what makes this a real signal.

2. Which feature here is the least valuable to you?

   Single select

   MaxDiff move — pairing 'best' with 'worst' on the same set is far sharper than rating each in isolation.

3. What's the last thing that made you wish the product did something it doesn't?

   Long text

   Past-behavior anchor — a real moment beats a hypothetical wishlist.

4. Which of these would you actually pay more to get?

   Multi select

   Willingness to pay filters out 'sounds nice' from 'genuinely needed'.

5. If we shipped your top pick next month, what would it let you stop doing?

   Long text

   Reveals the workaround a feature would replace — the true measure of value.

6. How are you handling that need today?

   Long text

   If the answer is 'nothing', the pain may be smaller than it sounds.

7. What's your role?

   Short text

   Lets you slice priorities by segment — power users and new users rank differently.

How to run it well

• Pair a 'most valuable' with a 'least valuable' question on the same feature set — that contrast is the MaxDiff signal that breaks the all-5s tie.
• Cap the list at 5-7 features per question; more than that and people stop reading and start clicking.
• Tie willingness-to-pay to specific features, not the product overall — that's where the prioritization actually lives.
• Segment results by role and tenure before you act; a feature your champions rank last may be the thing that converts new users.
• Re-run after each major release so the backlog ranking reflects the product people use now, not the one they signed up for.

FAQ

Why not just ask users to rate each feature's importance?

Importance ratings compress to the top of the scale — almost everyone marks almost everything 'very important', so you can't rank anything. Forcing a single choice or a best/worst pairing makes people reveal real tradeoffs.

How many responses do I need to trust a feature ranking?

For a clear winner, patterns usually stabilize around 30-50 responses. If two features sit neck-and-neck, gather more or segment — the tie often resolves once you split by user type.

Is this a real MaxDiff study?

It's a lightweight, MaxDiff-inspired version — true MaxDiff rotates many feature subsets and models utilities statistically. This template captures the same best/worst logic without the survey-science overhead, which is enough for most product calls.

Related templates