User Studies Overview
A User Study brings product research into one place. UX Research and A/B Studies run Personas through your product, while Surveys can collect answers from real people, simulated Personas, or both.
User Studies live on the User Studies page, scoped to the Application you’re working in. Survey questionnaires and responses are part of User Studies rather than a separate product.
The three study types
Section titled “The three study types”Every study is one of three types.
| Type | You provide | You get back |
|---|---|---|
| UX Research | A research question | Where Personas felt clear versus confused, as findings backed by their own words |
| A/B Study | A description of the two pages you’re comparing | Which variant converts better, and why |
| Survey | An internal questionnaire or external survey link | Organic and simulated responses, with per-question reports |
Creating a study
Section titled “Creating a study”Choose the study type, name it, and configure what that type needs:
- UX Research — Write the research question, review the drafted tasks and questions, and choose the Personas that will participate.
- A/B Study — Describe the study in your own words: the two pages you’re comparing and the conversion you care about. Marketrix reads the two variant URLs and the conversion goal out of that description and fills them in — every derived field stays editable, and it never invents a URL you didn’t write. Review the task every Persona runs on each version — your description itself is never shown to them — and the goal-tailored questions, then choose the Personas. An A/B Study compares two URLs today; testing the same page with your changes applied in the browser is marked Coming Soon in the wizard.
- Survey — Start with a flat internal questionnaire or an external survey link. Publish the internal questions for real respondents, choose Personas for a simulated run, or do both. See Surveys.
When an external Survey’s questions are unknown, the first external run discovers and saves them. Later runs use those internal questions by default; answering the external page again is an explicit option on each run.
Two settings shape how simulated runs are grounded, and you can change either before launching:
- Add context — attach Knowledge files from this Application so Marketrix works from what it has read about your product.
- Respondent transcripts — when on, each persona answers from its own transcripts, the real words of the people it stands for, rather than its trait summary alone. Personas with no transcripts are unaffected either way.
A third option, Screenshots, controls how Marketrix sees each page during a browser run: on, it studies a screenshot of every page for the best results; off, runs are faster with slightly less detail. It applies to A/B Studies and simulated Survey runs — UX Research always studies screenshots, so turning it off there changes nothing.
How a run works
Section titled “How a run works”Each time you run a study with Personas it produces a run, and a study keeps its run history so you can compare results over time. A run moves through a shared lifecycle — running → finalizing → completed (or failed / stopped) — and while it runs you can watch the simulations replay live. A run that reaches your Billing Cycle’s credit allowance goes Paused instead of failing, keeping everything it has already collected; Resume on the run starts only the work that never began. See Billing and Credits.
Every simulation produces a Persona Reaction: what that Persona made of each step and how it felt about it. Once every Persona has finished, Marketrix folds those reactions — plus each Persona’s own answers — into per-Persona findings and a run-level summary. Survey responses from real people arrive independently and are labeled Organic.
From a study you can:
- Switch runs — Compare the current run against earlier runs of the same study.
- Read per-persona results — See each persona’s reactions and the type-specific output.
- Replay — Follow the linked simulation replays and jump to the moment tied to a persona’s reaction.
- Export — Generate a PDF report of a completed run, or export the responses as CSV.
What each type returns
Section titled “What each type returns”UX Research
Section titled “UX Research”Personas perform neutral tasks in your app, then answer your research questions in character. The run returns per-persona key findings — each a claim backed by attributed quotes — folded into one plain-language answer to your research question, alongside how many participants said they would have finished each task and the step where the ones who gave up said they would have stopped.
A/B Study
Section titled “A/B Study”Each participant judges one of the two versions and answers about that one: where they would have stopped, or that they would have finished, and how sure they feel saying so. Participants are split across the two versions, so the two counts come from two separate groups and no answer is coloured by having already seen the other page. Each Persona does drive both versions itself and says which of the two it preferred — the order it sees them in alternates between Personas, so being shown second is not mistaken for being worse.
The headline is a count, not a score: how many participants out of how many said they would have finished on each variant. Because the two groups are separate, the two counts are read as two independent proportions — a difference is called only when the gap is bigger than groups this size could produce by chance, and when too few finished, or too few stopped, on either version, the comparison cannot separate them at all.
A variant is crowned only when that comparison actually separates the two. The report leads with No clear winner when the two versions were not equivalent to sit through, when the two groups that answered them were too different in size for the gap to mean anything, or when the counts are simply too close to tell apart — and it says which of the three it was. When only some Personas met versions that were not equivalent, the report discloses how many participants that covers. Every run also ends with ranked recommendations and the risks of shipping each variant.
When completions are too close to call but the Personas’ own preference does separate — each of them having driven both versions — the report says exactly that and names the preferred version. That preference is reported beside the completion counts and never folded into them: a stated preference is not evidence that anyone would have finished.
Simulated participants are not real visitors, so every figure is a direction to check rather than a measurement, and the report says so where it is measured. The PDF export makes it concrete: next to each variant’s count it prints the range that count would plausibly land in on another run of the same size.
Survey
Section titled “Survey”Start with an internal questionnaire or an external survey link. Publish the internal questions at /survey/{ref} to collect submitted organic responses, with optional password protection, email collection, a thank-you message and an optional post-submission redirect. Persona runs use the known internal questions by default and can explicitly use the external page instead. The Responses tab keeps individual submissions together, while Reports switches between Organic, Simulated, and Both. See Surveys.