A falling conversion rate leaves your team with several plausible explanations: a different campaign, a new theme release, a sold-out bestseller, a slow page or a broken cart. The useful next step is to turn that list into an investigation someone can act on.
Start with three questions: is the comparison consistent, where did the journey change, and what evidence would justify the next action? You can use this process with your existing analytics and your developer. Dozenfold helps connect the storefront evidence along the way.
First, make sure the two numbers mean the same thing
Write down the current and previous periods, the time zone and the exact conversion definition. Comparing a partial Monday with a full week, or store visits with checkout entrants, can send you toward the wrong explanation.
Use the same source and a comparable period. In a fictional example, a change from 3.2% to 2.4% is a decline of 0.8 percentage points, or 25% relative to the previous rate. Those describe the same change; mixing them makes it look different.
Then check what changed in the audience and measurement. More mobile or paid-social traffic can move the overall rate even when each group’s experience is unchanged. A change in consent behavior or collection coverage can also change which sessions appear in a monitoring tool. Record those facts before assigning a cause.
Locate the step that needs attention
Use the funnel to choose a starting point, then look for direct evidence. A drop-off identifies a place to investigate; it does not identify a bug by itself.
| Where the change appears | What to inspect next | Useful evidence to bring |
|---|---|---|
| Store visit → product view | Landing pages, campaign mix, product availability and page loading | Affected pages, traffic mix, page performance and observed navigation |
| Product view → cart | Variant selection, add-to-cart behavior, inventory and relevant app changes | Affected journey, specific failed interaction or request, browser and device |
| Cart → checkout | Cart updates, discounts, shipping expectations and navigation | Captured steps before the exit, cart request failures and recent changes |
| Checkout → purchase | Checkout progression, shipping/payment availability and applicable Shopify diagnostics | Funnel events and an actual reported or captured failure, with its source |
In Dozenfold, start with Overview for the broad change, Pages for where it concentrates, then the linked issues and sessions. The product tour shows the workflow with illustrative data. For a specific cart failure, use the cart investigation guide.
Put recent changes beside the affected group
Write a short timeline of theme releases, app changes, promotions, campaigns and stock changes. Compare it with when the symptom first appeared. If the change concentrates on one browser or page, have development examine that path first.
A release near a regression is a useful lead. It is not enough to choose a rollback. Look for a relevant code change, captured error or reproducible failure. Dozenfold’s release view separates early comparisons from a release’s ongoing activity; the release guide explains those windows.
Avoid beginning with a site-wide redesign when you already have a specific failing interaction to check. Equally, if that interaction works and the evidence points to traffic or merchandising, let the investigation follow it.
Give the next fix a business case and an owner
For the issue you select, bring four things to the review: the shopper action that failed, the observed affected sessions, the technical evidence and the next action’s owner.
Where the evidence supports a revenue estimate, use it to explain the priority. In Dozenfold that amount belongs to one issue and its selected period; the comparison, its 95% interval and the measured or directional label are available to inspect. An unavailable estimate does not prevent your team from investigating a broken interaction. The revenue guide explains how an amount qualifies.
After a fix, record the release and revisit the same journey under relevant observed traffic. Separate what your team changed from what subsequently improved. That gives the next review something more useful than “the chart went up.”
Bring a review your whole team can use
The storefront review template gives you a reusable structure: business change, priority issue, action owner and follow-up evidence. It is a manual working document, not an automated report, and works for a merchant and an agency reviewing a client together.
Want to work through your storefront’s buying flows with us? Book a demo. See the investigation workflow and discuss setup with the person responsible for your storefront changes.