A product page has been edited several times, but the merchant cannot tell which edit helped and which hurt orders. In this case, the immediate task is to design targeted improvements to product-page conversion. The workflow should produce a conversion diagnosis and experiment backlog tied to the page's own views, add-to-cart events and orders.
First replace every [bracketed] placeholder with real store scope and constraints, then add any actual records or data you have, and paste it into your preferred AI assistant or Spark agent. Spark will follow this page's workflow to produce structured output and verification steps.
Act as a Shopify operations specialist.
Help me design targeted improvements to product-page conversion.
My store context is [describe the business, products and target audience].
My scope is [specify the affected products, pages, customers or reporting period]. Use these inputs: [the page's views deduplicated per session, add-to-cart events, orders whose line items include that product, the dated list of page edits with what each changed, the price, variant and shipping conditions shown, and the source mix for the window]. Attribute a movement in the page's add-to-cart and order counts to a dated edit only where the edit precedes the movement and no source-mix shift or price change in the same window explains it better, marking the edit as unproven where the page's counts fall below the merchant's stated minimum for that window.
Return a conversion diagnosis and experiment backlog with success and guardrail metrics set for each test for the specified task.
Show the relevant evidence or before-and-after examples and explain how I can verify the result.
If essential data is missing, ask for it and mark unsupported conclusions as unknown.
Do not invent metrics, product claims or customer facts.
For any publication, message, account or store change, use only connected tools within my explicitly authorized scope and report success only after verification; otherwise provide the draft or proposed steps.Split the work into input, judgment, and deliverable phases. Verify each step's output before moving to the next.
Collect for the page, over one merchant-defined window, its views deduplicated to one per session, its add-to-cart events, the orders whose line items include that page's product, the dated list of edits made to the page with what each edit changed, the price, variant and shipping conditions shown, and the source mix of the sessions arriving on it.
Attribute a movement in the page's add-to-cart and order counts to a dated edit only where the edit precedes the movement and no source-mix shift or price change in the same window explains it better, marking the edit as unproven where the page's counts fall below the merchant's stated minimum for that window.
Deliver a conversion diagnosis and an experiment backlog in which each proposed change names the element it targets, the record linking it to lost carts or exits, the success metric and the guardrail metric measurable from the same records, and the rollback value of the element, then reconcile the orders used with the merchant's order export.
Cover decision rules, frequent misunderstandings, and required inputs so the team aligns on prerequisites before execution.
An edit is linked only when it predates the movement in the page's counts and no other dated change in that window, such as a price change or a source-mix shift, explains it first. Timing alone is not proof, so the pairing is recorded with the alternative considered; edits that cannot be dated are listed as unattributable.
The guardrail is a second metric that must not worsen while the success metric improves, such as the order count measured with the same denominator. Without it a change can lift add-to-carts while orders fall and still be called a win; both metrics are fixed before the test starts, so the result cannot be judged afterwards on whatever moved.
A page whose traffic is below the merchant's stated minimum for a window is reported with raw counts and given neither a rate nor a test. With few visits a single order moves the figure, so those pages either wait for more traffic or are grouped with similar pages so the combined volume supports a conclusion.
Start with scenarios under the same topic, then expand into cross-topic matches suggested by similarity.
Ciwi Spark builds Shopify-native AI tools and agents around the workflows merchants actually run: research, ads, translation, and customer support. Stop rewriting the same prompts by hand.