What this page is

It is one screen that follows a single person from the ad they clicked all the way to the money they paid. That sounds obvious. No tool in the standard coaching stack does it.

Meta knows about the click and the opt-in. The CRM knows about the call. Stripe knows about the payment. None of them share a key, so nobody can answer the only question that matters: which ad produced the person who actually paid?

So the page does one job. It stamps every registration with the ad that produced it, then joins that to what happened on the call. Once those two facts live on the same row, everything else on this screen is just a query.

It took an afternoon. It replaces the entry tier of a tool that costs $459 a month.

How we built it

Four steps. None of them are clever, and that is rather the point.

utm_source=facebook&utm_medium=paid_social
&utm_campaign={{campaign.name}}
&utm_content={{ad.name}}
&utm_term={{adset.name}}
&utm_placement={{placement}}

// registrant row
{ email, name,
  utm_content:  // the ad name
  utm_term:     // the ad set
  variant:      // which page they saw
  qualified:    // from the step-2 answer
  created_at }

registrations  LEFT JOIN  calls    ON email
               LEFT JOIN  enrolments ON email

SELECT utm_content,
       count(*)                        -- registrations
       count(call_id), count(enrol_id) -- the invisible part
FROM joined GROUP BY utm_content

Where the ad platform stops looking

Meta can report the top two rows. It has no idea the bottom four happened, because they happen on calls and in a CRM.

All 17 ads, as the page sees them

Sorted by registrations. This is the part the platform can see, and it is still the part almost nobody reads properly. Note how badly click-through rate predicts anything.

The creative behind the numbers

Nine statics and three video ads ran in this account. Every row in the table above is one of these. The statics did the heavy lifting: about 40 of the 62 registrations came from them, not video.

These are the assets, not the answers. Nothing about looking at them tells you which one produced an enrolment — the two best-looking ads in this set include one that produced 18 registrations and one that produced none. That is the entire argument for the table above.

What it changed

Four decisions came out of this screen, and all four were the opposite of what the ad platform would have suggested.