# From a list of ideas to a prioritized queue. > Bixlow hypotheses are step-by-step guided briefs, linked one-to-one to a workflow, with attached evidence the AI analyzes and a 1–100 priority score calculated from reach, impact, confidence and difficulty. - Source: https://bixlow.com/en/platform/hypotheses - Language: en - Publisher: Bixlow Every idea becomes a hypothesis with a problem, a change and a metric, backed by data and scored so you know what goes first. The agent interviews you, analyzes the evidence and drafts the brief with you. - **Brief:** Guided, step by step - **Evidence:** Files and CSVs - **Priority:** Score from 1 to 100 - **Link:** One brief, one workflow ## Why ideas stay ideas There is no shortage of ideas. What is missing is knowing which deserve traffic, why, and in what order. ### The list lives in a spreadsheet nobody opens Ideas get written down in a meeting, sit in a document and never reach the site. ### "I think" is not a hypothesis Without a specific problem, a change and a metric, you do not know what you are testing or what the result means. ### The data is there, but nobody looks at it The Analytics or Shopify export holds the clue, but working through it takes an afternoon nobody has. ### The most senior person's idea gets tested Without a shared criterion, priority is set by the highest-ranking opinion, not by expected impact. ## From finding to a hypothesis ready to build The brief walks through the steps of a good hypothesis and stays linked to the workflow that will put it live. 1. **Tell the assistant your idea.** The assistant interviews you: what you saw, on which page, who it affects. From that it drafts a first brief. 2. **Add the evidence.** Attach screenshots, documents or a CSV exported from Google Analytics, Shopify or a spreadsheet. The AI analyzes it and summarizes what it finds. 3. **Score and rank.** Rate reach, impact, confidence and difficulty. Bixlow calculates a 1–100 score so the queue sorts itself. 4. **Hand it off to build.** Every brief is linked to its workflow. When its turn comes, the agent builds the change or the test from what the brief says. ## Everything you need to decide what to test One place for the ideas, the evidence behind them and the order they will be built in. ### Guided brief: A complete hypothesis, not a loose sentence The brief is filled in step by step, so no hypothesis leaves without a problem, a change and a metric. - A step-by-step flow with what needs deciding at each stage. - An assistant that interviews, drafts and edits the brief with you. - Linked one-to-one to the workflow that puts it live. ### Evidence: Your data, analyzed by the AI Attach what backs the idea and let the AI draw the conclusions that would take you an afternoon. - Attachments stored alongside the brief. - CSVs exported from Google Analytics, Shopify or a spreadsheet. - A summary of what the data says, inside the brief itself. ### Priority: A queue ordered by expected impact One score for every idea, so the conversation is about data rather than opinions. - Reach × impact × confidence ÷ difficulty. - A 1–100 scale, comparable across hypotheses. - The queue sorts itself by score. ## The agent turns a finding into a hypothesis Tell it what you saw or hand it an export, and it comes back with a complete brief you can edit. ### You decide - Which problems matter to the business. - Which evidence is reliable. - The final rating of each hypothesis. - What gets built and when. ### The agent does - Interviews you and drafts the brief. - Analyzes CSVs and summarizes what it finds. - Proposes problem, change, audience and primary metric. - Builds the workflow once the hypothesis is ready. ### A hypothesis in Bixlow - **Evidence:** GA4 export · mobile sessions - **Problem:** On mobile, the buy button sits below the fold - **Change:** Sticky add-to-cart bar - **Audience:** Mobile visitors - **Primary metric:** Add to cart - **Priority:** 72 / 100 ## Bixlow Studio: Want a roadmap, not just a list? Bixlow Studio audits your site, turns the findings into prioritized hypotheses and builds them every week, with a strategist approving every step. - Conversion audit with hypotheses backed by your data. - A prioritized roadmap reviewed with you. - Weekly delivery of the changes and tests in the queue. ## FAQ ### What goes into a hypothesis brief in Bixlow? The problem you found, the change you propose, who it affects and the metric you will measure it with, together with the evidence behind it and its priority score. ### How is priority calculated? Reach, impact and confidence are multiplied and divided by difficulty. The result is mapped to a 1–100 scale so hypotheses can be compared with each other. ### What kind of data can the AI analyze? CSVs exported from Google Analytics, Shopify or any spreadsheet, plus the documents and images you attach to the brief. ### Do I have to use the assistant to create a hypothesis? No. You can fill in the brief yourself, step by step. The assistant is there to save time: it interviews, drafts and edits, and you review. ### What happens to the hypothesis once the change is built? Each brief is linked to its workflow, so the context travels with the change: the agent builds from the brief and the team always knows why that workflow exists. ## Get your ideas in order today Tell the assistant what you saw, attach your data and walk away with your first prioritized hypothesis.