Two segments said yes to the same feature for opposite reasons

The situation

A prepaid mobile carrier had a new plan ready to test: unused data rolls over to the next billing cycle instead of expiring. Product had built it to solve a specific complaint that kept showing up in support tickets and call center logs, data expiring unused at the end of the month. The plan was priced slightly above the standard tier and scheduled for a limited-market pilot in three weeks.

Marketing's working assumption was that heavy data users, people who burned through their allowance every month, would be the natural audience. Insights had two studies on the shelf that suggested this wasn't obvious: a usage and billing survey from the previous year, and a more recent plan satisfaction tracker. Neither had ever asked about rollover directly, because the feature didn't exist when they ran. But both had plenty to say about how different customer groups thought about data limits, overage charges, and plan switching, something that mattered a great deal in a US prepaid market where customers switch carriers freely and overage fees are a common point of frustration.

The team wanted to know, before building a campaign, whether "heavy users" was actually the right target or just the most obvious one.

What they put in

Two studies:

• Usage and billing survey: consumption patterns, overage incidents, plan-switching history
• Plan satisfaction tracker: pain points, price sensitivity, competitor comparison

They locked the brief to prepaid customers on mid-tier and top-tier data plans, described the rollover mechanic directly, and excluded postpaid customers since the plan wasn't available to them.

What PersonaAI created

PersonaAI matched respondents in that corpus to the locked brief and created three digital twins, each with contributing studies, respondent counts, a match score, and facet coverage.

The team selected two:

• Budget Maximizers: consistently use their full allowance, occasionally hit overage charges, strong coverage on billing pain points and price sensitivity
• Irregular Users: usage swings month to month depending on travel or work, sometimes have leftover data, sometimes run short, strong coverage on plan-switching history, thinner coverage on overage experience

A third twin, built from low-usage customers who rarely came close to their limit, was created and set aside. The team judged this group unlikely to value a rollover feature regardless of what the twin conversation showed.

What they asked

They didn't ask "would you buy this plan." They described the rollover mechanic exactly as it would appear on the carrier's app or recharge screen and asked what problem each twin thought it solved, when they'd actually notice the benefit, and what would make them switch from their current plan to get it.

What the twins showed, and what they did not

Budget Maximizers wanted it for the overage protection, not the rollover itself.

The twin kept returning to unexpected charges, not unused data. Rollover mattered to this group mainly as a buffer against running out mid-cycle and getting hit with an overage fee. Whether the leftover data carried forward was secondary to whether it prevented a bad month.

Irregular Users wanted it for exactly the opposite reason.

This twin cared about not losing data they'd already paid for during a light month, ahead of a heavier month they knew was coming, like a cross-country trip or a work travel stretch. The overage angle barely came up. The feature's value here was about not wasting money already spent, not about avoiding a future charge.

Both segments said yes. The reason they said yes pointed to two different messages.

Coverage on billing pain points was strong enough in Budget Maximizers to support an overage-avoidance angle with confidence. Coverage on plan-switching history in Irregular Users supported a value-retention angle just as clearly. A single campaign built around "never lose your data" would have undersold the plan to Budget Maximizers, who weren't primarily worried about losing data. They were worried about paying more.

Whether either segment would actually switch plans for this, versus just liking the idea, was a gap.

Both studies measured attitudes and past switching behavior, not intent toward a feature that didn't exist yet. PersonaAI flagged stated switching intent as uncovered rather than inferring a take-up rate from general plan-switching history. The team treated that as the number the pilot itself needed to produce.

What changed

Marketing built two versions of the campaign instead of one: an overage-avoidance message for channels that reached Budget Maximizers, and a "don't lose what you paid for" message aimed at Irregular Users. The single generic "never lose your data" concept was dropped before creative production started.

The pilot's success metric changed too.

• Before: Measure plan uptake among heavy data users.
• After: Measure uptake split by message variant, and check whether Budget Maximizer sign-ups show a drop in overage incidents in the following billing cycle, since that was the actual behavior the feature was meant to change.