TEST TWO MESSAGES ON THE SAME PLACE

What you get

• Same twins, two versions

  An A/B reads both versions against one population, so the gap between them is not a gap between two different samples.

• The gap, not the level

  What you read is which version is ahead and by how much the interval allows — the part a simulation is suited to.

• Cut by group for free

  See whether the version ahead overall is also ahead with renters, older households or one neighbourhood.

• The interval travels with it

  Every reading carries an interval and says when the two versions are too close to call.

Studies you could run

• Residents

  A new bus route: announced as a time saver, or as a cost saver?

  A reading shows: Which framing is ahead overall, and whether it flips among households without a car.

• Residents

  A school bond: open with class sizes, or with building repairs?

  A reading shows: Where the two framings separate by age and by neighbourhood, and where they don’t.

• Businesses

  A delivery service pitched on speed, or on price?

  A reading shows: Which pitch is ahead in this reading among restaurants compared with retail.

Questions

Who answers?

Simulated twins of the place’s population. Each twin is modeled on the real make-up of the place — age, household, work, housing — and weighted back to it.

Can I test more than two messages?

An A/B compares two. For more, ask one question with up to 12 options and read which options are ahead.

Does the order of the versions matter?

Each version is read separately by the same twins, so neither one colours the other. Wording still matters — keep the two versions the same length and tone.

Is this a forecast of how the message will do?

No. It is a simulation. Read which version is ahead and with whom, not the level of support.