What Comes After a Persona Document Like Make My Persona
On this page
A team with a persona document may already know whom it wants to reach. The next question is which message, price, or feature framing those buyers would choose. A profile can help define the audience; a choice study needs alternatives, a measured endpoint, and a comparison design.
What does a persona document do?
As of October 2026, HubSpot's free AI Persona Generator accepts a plain-language customer description and fills out a persona for review and PDF export. HubSpot also offers a Legacy App with a guided form. Both produce a shareable profile; neither profile by itself supplies evidence that a particular price or message changes buyer choices.
Use the document for sprint planning, onboarding, or agreeing whom the team is building for. Treat its goals and challenges as inputs to research. If two messages address the same stated challenge, the profile alone cannot establish which message performs better.
Three answers to "who is our customer," and what each one leaves open
| Approach | What it produces | What it cannot tell you |
|---|---|---|
| Static persona document | A reference profile of goals, demographics, and challenges | A tested preference between specified offers |
| Conversational persona interface | Generated answers that can suggest hypotheses | Human choice evidence from conversation alone; inspect any separate testing protocol |
| Randomized discrete choice study | A comparison of choices among specified alternatives | Actual purchases or psychological mechanisms from stated or generated choices alone |
A simulated persona transcript can suggest an alternative worth testing. Before using its answer to set a launch price, ask whether the team compared alternatives under a defined protocol and whether results were checked against relevant human evidence.
What does a controlled experiment add?
Subconscious uses controlled choice studies to compare defined options on generated responses. Specify the audience, varied attributes, and outcome before interpreting the comparison. McFadden's choice framework, mixed logit, and ICLV are modeling approaches; the study design determines what the estimate can answer. Request the configured estimator, effect units, and uncertainty calculation for the proposed study.
The public validation site and July 2026 working paper describe replication evidence. The paper reports mean Spearman correlation of 0.73 for estimated choice-parameter ranks across 43 design-filtered studies. That result concerns agreement in parameter ordering; it does not establish which offer a new audience will buy.
For a material pricing decision, plan a human study with aligned alternatives, attribute levels, and outcome definitions. Record changes needed for human fielding and compare the estimates, including disagreements, before deciding whether the simulated result transfers.
Where this does not fit
A defined choice set is a prerequisite. If the team is still discovering customer needs and has no message or price alternatives to compare, interviews and a shareable persona document can help define the next study.
For example, a persona describing a budget-conscious operations lead does not choose between a lower subscription price and an included onboarding service. A choice task can compare those offers. Generated preference between offer descriptions still needs relevant human validation before it becomes evidence about purchases.
Deciding which one to use
- Need a shareable reference document for internal alignment: use a persona generator.
- Need to explore a customer segment before knowing what to test: stay in open-ended research.
- Have specified message, price, or feature alternatives: design a choice comparison, define the minimum useful effect, and allow an inconclusive result.
Teams in the third case can see how the process runs or book a walkthrough.
