ADS & LANDING PAGES

Post-click drop off diagnosis

Use when ads are performing but the traffic they send does not convert, and you need an ordered list of suspects.

post-click-drop-off-diagnosis.md

Download .md

You are a paid media diagnostician. Traffic is arriving and not converting. Work out where it is going wrong and in what order to check things.

Ad copy and keywords or audience: {{AD_AND_TARGETING}}
Landing page text as a visitor reads it: {{PAGE_TEXT}}
Funnel numbers I have (impressions, clicks, sessions, scroll depth, form starts, form completes, conversions): {{FUNNEL_NUMBERS}}
Device and traffic split if known: {{DEVICE_SPLIT}}
Anything that changed recently: {{RECENT_CHANGES}}

Produce:
1. A drop off table using only my numbers: Stage, Volume, Drop from previous stage, Percent lost. Mark any stage I did not supply as "not supplied" rather than filling it in.
2. The largest single drop, named plainly, with the three most likely causes for that specific stage.
3. A ranked list of eight hypotheses across targeting, message match, page clarity, form friction, trust, speed, device, and tracking. Rank by expected impact divided by effort to check, and show that reasoning for each.
4. For each of the top four, the exact check to run and what result would confirm or clear it.
5. A tracking sanity check: what would make these numbers wrong rather than the funnel being broken.

Rules:
- Never estimate a number I did not give you.
- Do not recommend a redesign before the tracking check has been cleared.
- No em dashes. British spelling.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{AD_AND_TARGETING}}
{{PAGE_TEXT}}
{{FUNNEL_NUMBERS}}
{{DEVICE_SPLIT}}
{{RECENT_CHANGES}}

Getting a better result

Clear the tracking hypothesis first - broken conversion tags mimic every other problem.
Supply the device split; mobile-only drops point straight at layout or form issues.
Note recent changes honestly, including ones you think were unrelated.

Questions about this prompt

When do I use this rather than commissioning a redesign?

When ads perform, traffic arrives, nothing converts, and the meeting is drifting towards a redesign. A redesign changes everything at once and is usually proposed before anyone has checked whether the conversions are being recorded at all. This gives you an ordered list of suspects with the cheap checks first and the expensive ones last.

What do I need in front of me?

Whatever funnel numbers you actually hold, stage by stage: impressions, clicks, sessions, scroll depth, form starts, form completes, conversions. Gaps are fine, since unsupplied stages are marked rather than filled in. Include the device split and {{RECENT_CHANGES}}, listing changes you assume were unrelated, because those are the ones that turn out to matter.

What comes back?

A drop off table built only from your numbers, the largest single drop with three likely causes, eight hypotheses ranked by impact over effort to check, exact checks for the top four, and a tracking sanity check. Start with the tracking section wherever it ranks, because broken conversion tags imitate every other symptom on the list.

What is the mistake?

Approving a redesign before the tracking hypothesis is cleared, which is why the prompt refuses to recommend one until then. The second common error is supplying a blended device figure. A mobile-only collapse points straight at layout or form friction, and averaging it with desktop hides the finding you ran the diagnosis to get.