“No purchases” can describe two different problems. Customers may not be buying, or the advertising report may not be showing the purchases that happened. The first needs a commercial diagnosis. The second needs a measurement investigation. Mixing them leads teams to change campaigns while the real issue remains untouched.
Start by fixing the scope: the account, campaign, reporting dates, timezone, conversion event, and attribution view. Save that snapshot so later changes to the report do not erase the evidence available at the start of the investigation.
Check the order ledger first
Look for completed orders in the commerce system during the relevant period. Confirm whether they are paid, canceled, refunded, test transactions, or duplicates. An order ledger does not tell you which ad caused a sale, but it can establish whether the business received purchases at all.
If orders exist while advertising purchases are zero, inspect the measurement path before interpreting the campaign as commercially unsuccessful. Follow a permitted test transaction through the website event, analytics system, and ad-platform event diagnostics. Keep test orders separate from business results.
Google's GA4 ecommerce documentation describes purchase and refund events and the values they carry. Use the purchase-event QA guide to organize that check. Correct event collection is a prerequisite for reliable reporting, not proof of ad attribution.
Confirm what the campaign is trying to optimize
Inspect the selected objective and optimization event. A campaign designed to obtain traffic should not be judged as if its configuration explicitly prioritized purchases. This does not mean traffic can never produce sales; it means the delivery goal and business evaluation may be misaligned.
Also check the intended destination. Is the ad sending people to a product page, a collection, a lead form, or another surface? Does the destination support the action the team expects?
Meta's campaign evaluation course outline distinguishes reporting from tests used to measure campaign performance. The workflow below is an editorial diagnostic procedure; it does not substitute for a controlled experiment establishing lift.
Follow the funnel in order
Build a simple stage table for the same scope and period:
| Stage | Question | Useful evidence |
|---|---|---|
| Delivery | Did the ad serve to the intended market? | Status, impressions, spend, targeting settings |
| Arrival | Did people reach the destination? | Outbound clicks, landing-page views, website sessions |
| Product intent | Did visitors inspect the offer? | Product views and meaningful interactions |
| Checkout | Did they attempt to buy? | Cart and checkout events |
| Completion | Did payment and order creation succeed? | Paid order records and purchase events |
Definitions can differ between reporting systems, so do not expect every adjacent count to reconcile perfectly. Look for a material break that is supported by more than one observation.
For example, many clicks and very few observed page arrivals suggest a different investigation from many checkout starts and no completed orders. The former points toward destination or measurement issues; the latter makes payment, shipping, availability, and checkout behavior more relevant.
Use a worked example without inventing a benchmark
Suppose an illustrative campaign spends $300, records 150 outbound clicks, and shows 20 landing-page views with no purchases. The immediate question is why so few arrivals are observed. It is premature to conclude that the product needs a new price solely from that table.
Now suppose another campaign records 140 landing-page views, 18 checkout starts, and no paid orders. The team should inspect checkout availability, payment errors, unexpected shipping charges, and the event definitions. Replacing the opening video hook may leave the actual problem unchanged.
Neither example establishes a normal conversion rate or a spending threshold for your account. The figures show how the location of the apparent loss changes the next question.
Inspect the destination on a realistic device
Open the exact ad destination with its parameters on a mobile device or a representative browser environment. Check redirects, page loading, product availability, variant selection, offer terms, and the path to payment.
A fast office connection can hide a loading problem experienced by visitors elsewhere. Web Vitals provides a framework for examining loading, interactivity, and layout stability. Performance diagnostics should inform a concrete fix, rather than become a claim that a particular score guarantees sales.
Use the click-to-arrival investigation when the page-arrival gap is the strongest signal.
Review the promise made by the ad
If the technical path works, compare the ad's promise with the destination. Does the headline imply a price, outcome, product version, or delivery condition that the page does not support? Does a visually striking ad attract curiosity without explaining who should buy?
Read the first screen as a customer. Confirm that the product, price, key terms, and next step are understandable. Use actual customer questions and approved product evidence to improve the message. Do not fabricate scarcity or testimonials to compensate for weak conversion.
A high-CTR, low-sales diagnosis can help distinguish attention from purchase intent.
Decide what evidence would change the diagnosis
End the investigation with a specific hypothesis, one bounded correction, and a review condition. If the issue is a broken checkout, verify the repair through a permitted end-to-end test before increasing traffic. If the issue is an unclear offer, define the proposed message change and the outcome it should affect.
When the sample is small or conversions are delayed, state that the result is still uncertain. The right next step may be to continue observing within the approved budget, not to manufacture a confident explanation.
The goal is to locate the strongest supported failure point. A campaign rebuild is justified only when the evidence points to the campaign structure or strategy itself.