The feature everyone wanted built turned out to be the one nobody would pay for

The situation

A mid-sized B2B software company was planning its next major release. Product had a list of six candidate features, all of them requested by existing customers at some point over the past year: advanced reporting, a mobile app, single sign-on, API access, a dedicated account manager tier, and custom branding options.

Every feature had a champion inside the company. Every feature also had a real engineering cost, and the roadmap only had room for three of the six in the next two quarters. The usual approach, a simple customer survey asking people to rate how much they wanted each feature, had been tried before and hadn't held up. Customers rated almost everything highly when asked in isolation, which told the team nothing about what they'd actually give up something else to get, or what they'd pay more for.

The team needed to know which features would genuinely influence a renewal or an upgrade decision, not just which ones sounded nice when asked about individually.

What they put in

The insights team didn't reach for a standard satisfaction survey. The question wasn't "do you want this feature." It was "if you had to choose, which combination of features and price would actually win your business," which meant a trade-off question, not a rating question.

They ran the study with existing customers and a sample of prospects who had evaluated the product but not purchased, structuring it to test the six candidate features against each other along with price, rather than asking about each feature on its own.

What the analytics showed

Using a conjoint analysis approach, the study presented respondents with different combinations of features and pricing and asked them to choose between realistic packages, the way they'd actually choose between competing plans in the real world. A MaxDiff exercise ran alongside it, asking respondents to pick the most and least important feature from rotating sets, which forced trade-offs a simple rating scale never would have.

Single sign-on mattered far more than its rating had suggested.

On a straightforward satisfaction survey, SSO had rated in the middle of the pack. In the trade-off exercise, it turned out to be the single strongest driver of purchase decisions among the prospect sample specifically, the group the company most needed to convert. IT buyers evaluating the product alongside competitors treated its absence as close to disqualifying, something a simple rating question had never surfaced because nobody explicitly down-rated a feature they assumed was table stakes.

Custom branding and a dedicated account manager tier barely moved the needle.

Both had scored well when customers rated them individually, largely because they sounded appealing with no cost attached. Once respondents had to trade them against other features or against a lower price, willingness to pay for either collapsed. Neither feature, on its own, changed a single simulated purchase decision in the choice model.

Mobile app and advanced reporting split by segment.

Existing customers barely valued a mobile app. Prospects evaluating the product for field teams valued it heavily. The key driver analysis run alongside the choice data showed advanced reporting correlated more strongly with renewal likelihood among existing customers than with new purchase decisions, meaning it mattered more for keeping customers than winning new ones.

API access showed a clear price-sensitivity threshold.

The pricing and choice analytics component of the study identified a specific price point above which demand for an API add-on dropped sharply, information a simple feature survey could never have produced, since it never asked anyone what they'd actually pay.

What changed

Product cut custom branding and the dedicated account manager tier from the roadmap entirely, reallocating that engineering time to single sign-on, which moved from an unprioritized request to the first feature in the next release. Advanced reporting stayed on the roadmap but got repositioned internally as a retention investment rather than an acquisition driver, which changed how the launch would be marketed.

• Before: Six features, each with internal support, competing for three roadmap slots based on individual popularity ratings.
• After: A ranked list based on actual purchase-decision impact, with SSO prioritized for its outsized effect on prospect conversion and two previously favored features removed before any engineering work began.