Describe your process and Makeform creates a structured A/B test log form for the hypothesis, variants, audience, metric, dates, results, and decision.
Send experiment records to Slack, Google Sheets, and Zapier.
Sample prompts for the builder
Choose a prompt, tailor it to your testing process, or send it into the Makeform builder. The field structure is an example, not a live AI result.
Audience
Growth marketers testing page messages
Format
Experiment brief and result record
Prompt size
278 chars
Example form structure
Experiment brief and result record
Experiment ID, owner, and page URL
Hypothesis: if, then, because
Control and treatment details
Primary metric, baseline, and target
Outcome and next action
Suggested routing tags
Planned
Running
Decision recorded
Name one primary success metric before the test begins. Separating it from guardrail and diagnostic metrics makes the final decision easier to interpret.
Step 1
Plan
state the hypothesis and success metric
Step 2
Configure
document variants, audience, and allocation
Step 3
Measure
enter exposures, outcomes, and guardrails
Step 4
Decide
record interpretation and next action
Why use a test log
A structured log preserves why the test existed, what each audience saw, and how the team planned to decide.
Required fields keep hypotheses, variants, metrics, and decisions comparable.
Dedicated variant fields prevent copy, screenshots, allocation, and outcome numbers from becoming mixed together in a long project note.
Store counts, anomalies, guardrails, interpretation, and follow-up beside the decision.
Experiment programs
Start with a shared experiment spine, then add channel-specific metrics, artifacts, and decision choices.
Log copy, layout, CTA, or offer variants alongside traffic source and conversion outcomes.
Compare subject lines, send times, or body content with delivery, click, and unsubscribe context.
Document eligibility, feature flags, activation windows, retention guardrails, and rollout decisions.
Track step-level changes, device scope, purchase outcomes, revenue signals, and error guardrails.
Build your experiment record
Generate the structure, align it with your measurement plan, and publish it.
Name the channels you test, the required planning fields, the metrics you review, and the decisions your team uses.
Add statuses, metric definitions, audience rules, stop criteria, and evidence links.
Use sections for pre-launch assumptions and post-run results so the original hypothesis is not rewritten after outcomes appear.
Share the form, notify the experiment owner, and send structured submissions to the workspace where your backlog is maintained.
Log vs document vs dashboard
The log connects planning inputs to the final decision.
Field guide
A useful log preserves the plan before launch and the evidence after the run. These six sections give every experiment a complete, reviewable story.
Ownership & status
A stable ID and owner prevent duplicates. Status choices show whether an entry is proposed, running, completed, or archived.
Question & hypothesis
Capture the problem and mechanism before metrics arrive. An if-then-because structure makes the reasoning explicit.
Variants & audience
Describe control and treatment with links or screenshots. Add eligibility, exclusions, allocation, platform, and geography.
Metrics & stopping
Define one primary metric and measurement window. Keep guardrails separate and document the planned stopping rule.
Results & caveats
Record exposures, outcomes, metrics, and uncertainty per variant. Note tracking changes, overlapping campaigns, and segment differences.
Decision & follow-up
Choose ship, iterate, stop, or inconclusive, explain why, assign an owner, and link the follow-up task.
Related tools
Build supporting forms for research inputs, product signals, analytics requests, bugs, and feature ideas.
Collect audience evidence before choosing what message or offer to test.
Open toolAdd qualitative reactions to the behavior measured in an experiment.
Open toolStandardize requests for event definitions, cuts, and experiment readouts.
Open toolGather structured evaluations of a product experience or proposed change.
Open toolCapture follow-up ideas that emerge from experiment results.
Open toolRoute defects or tracking problems discovered during a test run.
Open toolFAQ
Practical answers for marketing, growth, analytics, and product teams building an experiment history.
It is a structured experiment record that captures the hypothesis, variants, audience, allocation, metric, guardrails, dates, results, caveats, decision, and follow-up.
Include ID, owner, status, hypothesis, variants, audience, allocation, primary metric, guardrails, dates, and stop criteria. Add counts, results, anomalies, interpretation, decision, and next action after the run.
That is often useful. Keep a pre-launch section for the frozen plan and a post-test section for the readout. You can use one record with status-based updates or two linked forms if your team needs stronger separation between planning and analysis.
Keep them in the log and label the decision as inconclusive rather than forcing a winner. Record the observed effect, uncertainty, run duration, sample counts, data issues, and whether the next action is to stop, extend only under a documented rule, or design a new test.
Yes. Add fields for each treatment and describe assignment. For a holdout, record its eligibility, exposure rule, measurement period, and comparison metrics.
Yes. Add file-upload fields for variant screenshots and short-answer fields for dashboards, queries, tickets, feature flags, or design files. Keep a written variant description too, so the record remains understandable even when an external link changes.
Yes. Makeform is unlimited free for generating, editing, publishing, and using your form. The paid tier removes the Makeform badge; it does not unlock additional submission capacity.
Require core planning and readout fields, use controlled status and decision choices, define metrics, and assign a stable ID. Create a new schema version when field meanings change.
Turn scattered test notes into a reusable experiment history.
