Measurement
Why paid-search learning often stops at the form fill
Most fertility advertising accounts optimize toward the moment someone raises a hand, because that is the last event the platform can see. The stages the clinic actually values live in other systems, and until they are connected, the bidding system learns from the wrong signal.
A paid-search account learns from the conversions it receives. In most fertility accounts, the conversion is a form submission or a call, because those are the events the website can report without any further plumbing. Everything the clinic cares about after that point, whether the person scheduled, showed up, or started treatment, happens in a CRM, a booking system, or a phone log the advertising platform never sees.
The result is a quiet mismatch. The clinic evaluates marketing by consultations and treatment starts. The bidding system evaluates itself by inquiries. Both are doing their jobs. They are just reading different scoreboards.
How the gap forms
When a campaign is told that a form fill is success, automated bidding does what it is designed to do: it finds more people who fill out forms. Some of those people are well-matched patients. Some are researching for a relative, comparing prices with no intent to travel, outside the clinic's geography, or seeking a service the clinic does not offer. The platform cannot tell these apart, because from its side of the wall every submission looks identical.
Over months, budget drifts toward whatever produces submissions most cheaply. Query mixes shift. Ad copy that attracts thin inquiries beats ad copy that attracts fewer, better ones. The account can report improving cost per lead while the front desk reports that lead quality is getting worse. Both reports are accurate.
What the platform can actually receive
Google and Microsoft both support conversion imports: the advertiser sends back a record that says, in effect, the click with this identifier eventually reached this stage, with this value. Google documents this as offline conversion import and enhanced conversions for leads. The mechanics matter less than the principle: the platform will optimize toward any event you can return to it reliably.
- A click identifier arrives with the ad visit and is captured at the form or call.
- The identifier is stored on the inquiry record in the CRM.
- When the record reaches a stage the clinic values, a neutral event name and optional value are returned against the identifier.
- The bidding system begins to distinguish clicks that became consultations from clicks that only became submissions.
Why this rarely gets built
The connection crosses three owners: the agency runs the ad account, the clinic owns the CRM, and a web vendor owns the forms. Each can reasonably say the join is someone else's job. It also needs maintenance, because form changes, CRM stage renames, and consent-banner updates all silently break the path. A one-time setup that nobody monitors degrades into false confidence within a quarter or two.
The practical test for any fertility account is simple to state. Pick ten recent consultations and ask which campaign, ad group, and search term produced each one. If the answer requires guessing, the learning loop is open, and the bidding system has been optimizing toward form fills all along.