A zero in Ads Manager can send you into a tracking rebuild while a customer is already waiting for an inspection. Before spending money to fix the zero, establish which question it failed to answer: did the business receive the lead, did Meta receive the event, or did the campaign receive credit?
Those questions need separate answers. Once an event is visible at its destination, repeating the browser-to-server setup checklist may investigate a boundary that has already passed. The unresolved question is what happened between receipt and the campaign report.
Keep the customer in the picture
A roofing advertiser reported a real lead stored in GHL, a phone conversation and a booked inspection. Events Manager showed browser and server Lead events, yet Ads Manager still showed no website lead after more than 90 hours. The advertiser also reported a Lead match-quality score of 6.1. Those are bounded account observations, not a reproduced platform defect.
They are enough to challenge two convenient answers: the business got nothing, or a visible event proves the whole measurement job is finished. Neither follows from that record.
Keep booked enquiries and collected orders visible while investigating campaign credit. A booked inspection is not yet a paid roofing job, and a shop's Facebook label is not proof that advertising created an otherwise impossible sale. Business activity, recorded acquisition clues and incremental advertising value are different claims. Do not make one dashboard carry all three.
Give waiting a question to answer
There is a reasonable case for not rebuilding a fresh setup over an early reporting gap. Changing several senders, events and campaign settings at once makes the next observation harder to interpret. Some observation time can be the cheaper choice.
But patience needs a review point and a spending limit. Write down what you expect to check next and which result would justify further investigation. A generic instruction to wait a week does not explain a specific lead that remains uncredited. Nor does the passage of a familiar number of hours establish the explanation.
The explanation for the reported gaps remains unresolved; neither an iPhone label nor a match-quality score supplies a diagnosis. Use them to choose a check, not to close the case.
Follow one missing outcome
Pick an order or enquiry whose business record is clear. Record its time, event, stored acquisition information and the campaign report being compared. Check the selected attribution settings and how that report handles dates before deciding the outcome is absent. A same-day store total and a campaign total are not automatically a like-for-like comparison.
Then inspect what actually accompanied the received event. A UTM or ad ID in the shop gives the investigation something to follow. It does not establish that the same information reached the advertising system or satisfied its attribution requirements. Equally, the presence of an ad ID should stop a missing-ID theory being accepted without examining the specific record.
Keep the question small: which part is observed, and which part is still being inferred? Do not turn a plausible identifier problem into a universal claim that every uncredited conversion needs the same field or suffered the same mobile-browser failure.
Another sender needs its own proof
A server-side change may be worth testing when the current event lacks useful matching information. First establish what the existing integration sends, what the proposed integration would add and which business event each represents. A planned Booked event is not evidence that a missing Lead has been repaired.
Do not grade the work by a target match-quality number alone. The result that matters to this investigation is whether the intended outcome can be followed through the reporting boundary. The score can inform that work without becoming a guarantee of campaign credit.
Check the existing browser and server paths before adding another. Inspect deduplication for the event under investigation: a PageView warning does not by itself explain an uncredited Lead. Make a controlled test and examine the resulting records rather than assuming a connector name settles double counting.
Make the budget decision separately
Missing campaign credit raises a legitimate concern about optimization feedback. It does not reveal exactly what the system used internally, establish an account-wide failure or erase a customer's order. Keep the uncertainty attached to the measurement claim.
Use the business outcomes and an affordable spending boundary to decide what can continue while the gap is investigated. Keep the evidence needed to repair attribution separate from the evidence needed to justify the next advertising dollar. Fixing the report should make that decision clearer, not make it for you.