The Conversion That Died When a Button Changed Its Name

The Setup

A residential fence installer running one Google Search campaign. The call that started the engagement was the classic one: the phones are slowing down, what is wrong with the ads?

The surface diagnosis was real. Cost per click had climbed to roughly two and a half times its level of three months earlier. Between the costlier clicks and the 27 to 48% of impression share the account was losing to ad rank every week, click volume fell 75% on an unchanged budget. But underneath it sat something worse: the account’s website lead tracking was completely dead. The only conversions recording were taps on the ad’s call button. Every form fill on the site, the account’s second lead path alongside phone calls, counted for nothing.

How I Found It

The tell was in the conversion actions list, not in the performance reports. I open that screen before I open anything else on a new account, and I read the Status column rather than the numbers. One action read Inactive, which in Google Ads means no conversions recorded in the last 30 days. Nothing on the overview screen surfaces that. The account showed a cost per conversion, a conversion count, and a trend line, all of them built from call-button taps alone, with no marker anywhere saying half the lead paths had stopped reporting.

That is why this class of failure survives for months. A dead conversion does not produce an error. It produces a smaller number, and a smaller number reads as a worse quarter.

What Was Broken

The website conversion existed. It was wired to a Tag Manager trigger that fired on a click whose text equaled “Send.”

At some point, the form’s submit button was renamed to “Submit.” Nobody touched the ads, nobody touched the tags, and the conversion died anyway. It had been sitting there Inactive, invisible unless someone went looking. Meanwhile the account also had automated recommendations applying themselves, including a bid-strategy change, with no human review.

A click-text trigger has two failure modes and this was both: it breaks on any copy edit, and it counts button presses instead of successful submissions, so even when it works it counts failures.

The compounding matters here. Smart Bidding was optimizing against the only signal it could still see, which was the call button. Every lead the site’s form produced was invisible to the bid model, so the campaign kept steering budget toward the audience most likely to tap a phone number and away from the audience most likely to fill in a form. Nobody made that decision. The broken trigger made it.

What I Did

Replaced the trigger with the event the form software emits. The form plugin fires a mail-sent event only on a successful submission. A small listener pushes that into the data layer, a custom event trigger picks it up, and the conversion tag fires on that. Verified live with a network capture: conversion ping, 200 OK. The new wiring survives button renames and counts only real submissions.

The distinction is the whole point. A click trigger observes the visitor. An event trigger observes the software. A visitor can click a submit button on a form that fails validation, times out, or gets rejected by the spam filter, and a click trigger counts all three as leads. The form plugin’s own success event fires after the mail is sent, so it cannot report a lead the business never received. Anything the page builder does to the button afterward is irrelevant to it.

Cleaned up the account it was starving. Turned off the self-applying recommendations, which had already changed a bid strategy without anyone approving it. Added competitor, DIY, and out-of-area negatives on top of the existing set. Removed duplicate keywords that were bidding against themselves. On a single-campaign account these are small moves individually, and together they are what collapsed the rank-lost impression share.

The Result

The form conversion fired its first real lead shortly after the fix, confirmed in the following month’s audit. Rank-lost impression share, which had been running 27 to 48% weekly, collapsed to under 1% after the cleanup. Ninety-day account view at the follow-up audit: 582 clicks, 42 conversions, with the two winning keyword themes carrying 86% of conversions and both protected during the cleanup.

What I Did Not Claim

Cheaper clicks. The click-cost diagnosis traced to a weak-relevance keyword theme and query waste, part of which is still being worked, and no tag fix changes what an auction charges. What changed is that the account measures itself again, the waste stopped, and the follow-up audit showed the real growth constraint was budget, which is a decision, not a defect.

The Takeaway

Tracking does not fail loudly. A button rename in a page builder killed this account’s only website conversion, and the dashboards showed a business that looked like it was getting worse at marketing.

If your conversions are wired to what a button says, they are wired to the next copy edit. Wire them to what the software does. That principle is the base of the tracking stack I rebuild every account on.

Two checks you can run on your own account this afternoon. Open Goals, then Conversions, then Summary, and read the Status column on every action rather than the totals. Anything reading Inactive or No recent conversions stopped working and nobody was told. Then open your Tag Manager triggers and search for Click Text. Every match is a conversion tied to a string of copy somebody in marketing can change without knowing what it costs.

The same failure has a second, quieter version worth naming. A trigger can keep firing and still count the wrong thing, which is what a click trigger does every time a submission fails validation. A conversion action that is recording is not the same as a conversion action that is correct.

When did someone last verify your conversions fire? My free written audit answers that within 24 hours of read-only access.