Design a statistically honest A/B test: hypothesis, primary metric, sample size reality-check, and the decision rule agreed before launch.
Design an A/B test for: {{change}} (what we're changing, current baseline numbers, traffic volume).
Spec: the hypothesis ('changing X will move Y because Z'), primary metric (ONE — everything else is secondary or guardrail), guardrail metrics (what must not degrade), sample-size reality check — with my traffic and baseline, how long to detect a {{effect_size}} lift? (If the answer is months, say so and suggest a bigger-effect test instead), randomisation unit and contamination risks, the decision rule written BEFORE launch (ship if X, kill if Y, extend if Z), and the 3 ways this test could mislead (novelty effect, seasonality, peeking) with countermeasures.
Fields like {{ this }} are placeholders — replace them with your own details, or let Aria ask you for them.
This prompt runs with Aria connected to your email, files, CRM and 40+ other services — and you can schedule it to run automatically.
Start Free WeekPull structured data out of messy text — emails, PDFs pastes, reports — into a clean table, with con...
Get the right Excel or Google Sheets formula for your layout — explained part by part, with a worked...
Get a structured first read of pasted data: observations with the numbers cited, trends vs noise sep...
Turn open-ended survey answers into themes and quotes.
Pick the best chart type for what you want to show.
Write or optimise a SQL query from a plain-English description of the data and the question — with t...