Consumers swear they’ll “buy American,” retailers build campaigns around the pledge, and pollsters dutifully record the sentiment. Yet the moment a credit-card form appears, patriotic rhetoric melts into price sensitivity. Behavioural scientists call this the intention-action gap: social-desirability bias inflates what we say we’ll do, while loss aversion governs what we actually do when money leaves the account. Ramón’s dataset is only the latest public autopsy of that gap; it’s painful because he left real money on the table and took inventory risk to rediscover a well-documented human quirk. 

Traditional research can’t fix this. Ask a panel, “Would you pay $239 for a U.S.-made model?” and respondents love to signal virtues. But would they call a bluff with stakes? Totally different game.

Here's the thing, when you AB test it, you're putting real money on the line. And even with stat sig confirmation, do you really walk away knowing "why" a variant won or lost? Often you don't. I've lived it too.

What we need is a way to force realism before the marketing budget burns.

The Rally Simulation

Rally lets you stage the checkout moment in silicon: hundreds of persona-agents, each primed with market-grade attributes, must decide with their own money. Here’s the single prompt that predicted Ramón’s face-plant for less than the cost of water in a latte:

You’re down to your last paycheck before rent is due, and you’ve been holding off replacing your old shower head, which now sprays sideways and leaves your skin itchy. You finally decide to fix it.

You find this highly-rated filter model with two options:

They are absolutely identical in design and function. The only difference is country of origin.

No returns allowed. You only have $250 in discretionary funds for the next two weeks.

Which one do you actually purchase—and why?

Explain your reasoning honestly, including any hesitation or emotional pushback. This is a real checkout moment—not a values survey.

Option A – $129, Made in Asia
Option B – $239, Made in the USA


The scarcity frame (“last paycheck,” “$250 cap”) neutralizes moral posturing, and the meta cue (“not a values survey”) jolts the agents out of virtue signaling. In live demos, over 90% of personas snap to Option A and openly admit the price pain—even those tagged as “locavore” or “ethical consumer.” The few who choose Option B describe status motives or tariff backlash, offering priceless fodder for messaging pivots.

Ouch! @justaskrally could have told you that would happen, even in fast mode. For $20/m . Without the expensive cart drop off in the real world.

I asked a virtual American audience too! https://t.co/GojNm8QiAm pic.twitter.com/SCQIfBkpnq

— Rhys Fisher (@virtual_rf) April 25, 2025

What It Saves

Running that simulation cost roughly $0.0033 in tokens from my $20/m Rally fast plan. Compare that to Ramón’s six-figure inventory commitments and thousands of wasted paid clicks. And hundreds of lost conversions!

More importantly, we have hundreds of ideas on what copy lines worth testing next. Because in the simulation, we don't just get what they voted. But why the did. The synthetic panel surfaced the uncomfortable truths—"because patriotism shouldn’t cost a rent check.”. It's like getting the quant and the qualitative feedback.

The lesson lands fast: when the question forces commitment under real-life constraints, AI doesn’t lie, and neither does your market. You can find truth  without burning 25,000 live visitors to learn it the hard way.

Don't leave winning to chance. Pickup a subscription and simulate it instead. 

If you'd like expert support with virtual audience design and custom projects, contact sales.