Storefront · Runs on a schedule

Find and fix weak product pages every week

Every Monday, Tolstoy ranks up to 10 weak product pages, rules out stock, price and traffic causes, proposes one reversible fix each and measures the fixes you approve.

Works with: Shopify, Google Analytics (optional), Tolstoy Library

Product page leaderboard ranking products in a grid

What you need

  • A connected Shopify store with enough product-page traffic to compare products
  • Google Analytics, if you use it, for page-level traffic

What you get

  • A Weekly PDP Brief: decisions on active changes and a ranked list of weak product pages with the evidence for each
  • Change proposals with the before and after state, a verification plan and a rollback path
  • An execution receipt and a measurement status for every change you approved

How it runs

Tolstoy runs this playbook in 5 steps.

  1. Close open loops first

    The routine measures each earlier change against its baseline and decides to keep, iterate, revert or keep monitoring it, or marks it inconclusive or blocked.

  2. Rank products, not URLs

    It joins every URL that leads to the same product, then ranks up to 10 low performers and treats missing data as unknown, never zero.

  3. Rule out ordinary causes

    Before it blames the page, it checks stock, sizes, price, promotions, traffic mix, seasonality and lifecycle against similar products.

  4. Propose one fix per product

    For each justified opportunity, it shows the current and proposed state, how it will verify and measure the change, and how to roll it back.

  5. Make only approved changes

    After you approve the exact scope, it makes that change, reads the live product again to confirm it and starts measuring.

Check the result

  • Each ranked rate names its metric, time window and traffic volume. A rate with no volume next to it is not ready to act on.
  • Each proposal changes one thing on one product, and you can read the exact before and after values.
  • After an approved change, the receipt shows the live product read again with the new value in place.

Know the limits

  • No live change without your approval of that exact scope. If the live product changed since the proposal, the routine asks again.
  • Before-and-after results are directional. Without a randomized test, the routine does not claim that a change caused a sales lift.
  • A week with no proposal is a valid result. The routine proposes a fix only when the evidence supports one.

When to use this playbook

Use it when some product pages convert worse than others and you cannot tell whether the page is the cause, or the stock, price or traffic behind it. The routine keeps a record of every change and its result, so each Monday starts from what the last change did. For common page problems to look for first, read five ecommerce product page mistakes to avoid.

Get better results

  • Approve one or two products at first. Small cohorts are easier to measure and to roll back.
  • Hold other changes to a product while its fix is measured. A price change or a promotion in the same window makes the result hard to read.
  • Connect Google Analytics if you use it, so the ranking has page-level traffic next to your Shopify data.

Schedule

Every Monday at 10:00 in your time zone; you can change the time. “Set up this routine” opens the routine in Tolstoy, where you check its connections and time. Nothing runs until you save it.

FAQ

How do I use UGC on product pages to lift conversions?

Put UGC where it answers the question that stops the purchase on that product, such as fit or how it looks in use, and measure the change. When the evidence points to a product page’s media, the routine can propose one media change, such as an image or video from your Tolstoy library, and measures it after you approve.

What kind of fix does it propose?

A small, isolated and reversible change to one product page, for example to its media or copy. It does not combine media, copy, price, inventory or template changes when a smaller change can answer the question.

What if the problem is not the page itself?

The routine says so. When stock, price, traffic or merchandising explains the gap better, it marks the product as a store-side action and proposes no page change.

Run it on your own catalog

Start free, connect your store, and Tolstoy opens this playbook for you. You review every result before anything goes live.