Constraint finder
Follow each Meta campaign from impression to conversion and find the step it is stuck on.
Overview
The Constraint finder follows each Meta campaign from impression to conversion. It names the step where the campaign loses the most: the ad, the tracking, the landing page or the sale. Open a client and choose Constraint finder.
It needs two things for the client: a BigQuery dataset with its tracked visits (see Exports) and Meta Ads data synced into that dataset. The screen says which one is missing.
The steps
| Step | What it measures | A low rate usually means |
|---|---|---|
| Ad | Impressions to link clicks | The creative or targeting is not earning the click. |
| Tracking | Link clicks to tracked visits | Redirects, consent banners, slow pages or stripped UTMs. |
| Landing page | Visits to leads | The page is not converting the visit. |
| Sales | Leads to conversions | Leads arrive but do not close. |
| Site | Visits to conversions | Used when no lead step is chosen. Pick a lead event to split page problems from sales problems. |
Each step's rate is compared with a target. By default the target is the client's own account average, not an industry benchmark. The verdict is the step where closing the gap would add the most conversions, taking the later steps into account. A campaign at or above target on every step reads Healthy, scale it. Can't measure means the visits could not be tied to that campaign, ad set or ad.
Constraints tab
Set up the view:
- Window: the date range.
- Group by: Campaign, Ad set or Ad.
- Conversion: the event that counts as a conversion.
- Lead step: the event that counts as a lead, or none for a single Site step.
- Credit: Last paid touch (closest to what Meta claims) or First paid touch (the campaign that first acquired the person).
- Credit window: Own session only, 7 days (matches Ads Manager), 30 days or 90 days (long sales cycles).
The table shows each row's spend, the rate at each step, conversions and the verdict, ordered by how much there is to gain. Click a row to see the funnel, and Ask agent what to change to get suggestions in the chat panel. Show ad creative adds each ad's copy, call to action, destination and URL tags. It adds about 290 MB of BigQuery scan per run.
Tracking audit tab
Checks whether ad clicks can be matched to visits. Each campaign gets a verdict:
| Verdict | Meaning |
|---|---|
| OK | Clicks match visits. |
| Low tracking | Many clicks have no matching visit. |
| Name only | Visits match by campaign name only, which breaks when the campaign is renamed. |
| No sessions | No visits could be matched. |
| No delivery | The campaign had no clicks. |
The tab lists causes with the visits affected and the fix, and names
problems such as an unexpanded macro in the UTM tags. Add
utm_id={{campaign.id}} to the ad's URL tags so visits match by id.
Events tab
Lists the client's events with how many carry a session or a person, which decides whether they can be credited to a campaign. Mythic suggests which event to use as the conversion and which as the lead step.
Targets tab
Replace the account average with your own targets. Set an account target, or Add a campaign target for one campaign, as a percentage for each step. A step left empty falls back to the next target in line: the campaign target, then the account target, then the account average. Clear removes a set of targets.
For developers
See the Constraint Finder API.