Broken Forms: Why Months of Leads Vanish Unseen

How can a broken form cost months of leads if the campaign looks healthy?

A broken form drains leads silently because the two systems you watch — your ad platform and your analytics dashboard — can both report success while the actual handoff to the inbox or CRM is dead. Clicks rise, the form 'submits,' a thank-you page fires a conversion, and the count climbs. Nobody sees that the submission never left the building. That gap can run for weeks before a single human asks where the replies went.

This is exactly what happened to Danny Gavin, founder of the Houston agency Optidge. As recounted on Search Engine Land, a landing page built for an autism-therapy provider kept generating qualified enquiries for roughly one to two months — while not one of them reached the client. Inside Google Ads the campaign looked strong: clicks climbing, cost per lead healthy. The client, meanwhile, grew convinced the whole effort was a failure because their phone never rang and their inbox stayed empty.

The most dangerous failure in lead generation is the one your dashboard rates as a success. A green conversion count tells you a script fired, not that a human received a lead.

The lesson most write-ups of this story miss: the technical bug was only half the problem. Gavin himself points to the communication failure underneath it. Across all those weeks, nobody — agency or client — asked the one blunt question that would have surfaced the issue in a day: 'Are you actually receiving the leads?'

Why doesn't conversion tracking catch a dead form?

Because conversion tracking and lead delivery are two different pipelines, and most teams only instrument the first one. A typical lead form has at least four independent stages, each of which can break without touching the others:

Tie your 'success' metric to stage two — the conversion fire — and you are measuring whether a tag loaded, not whether a lead survived. That is the trap. A thank-you page can render beautifully for months while the handler behind it returns a swallowed 500 or posts into a deactivated integration.

The 2026 web stack makes this worse, not better. Consent-mode v2 and modeled conversions mean a chunk of your reported conversions in Google Ads and GA4 are now statistical estimates, not raw events — so a partial drop in real submissions can be masked by modeling that fills the gap. Add frequent CMS and form-plugin auto-updates, third-party spam filters, and serverless cold-start timeouts, and a form has more silent failure points than it did five years ago.

What checks would have caught the failure in a day?

The fix is to stop trusting proxies and start verifying the actual outcome. Three layers, from cheapest to most robust:

• End-to-end test leads on a schedule. Submit a real form entry — tagged with an obvious test marker — from an incognito session at least weekly, then confirm it arrives in the destination inbox and the CRM. Not the thank-you page. The destination. This single ritual would have surfaced Gavin's bug almost immediately.
• Submission-count monitoring with anomaly alerts. Track raw submissions stored in your database or form tool, not just tag fires, and alert when the number drops to zero against a baseline. A form that averaged 15 submissions a week dropping to zero should page someone — not wait for a quarterly review.
• Cross-source reconciliation. Once a week, line up three numbers: ad-platform conversions, analytics events, and leads actually delivered to the client. When the first two climb and the third flatlines, you have found your failure before the client does. This is where SERP and traffic analytics earn their keep — Sentinel SERP's reporting makes it easy to watch traffic and engagement trends beside your conversion data, so a divergence between rising sessions and falling outcomes stands out instead of hiding across four separate tools.

Build the human question into the process

Tooling catches the mechanical failure, but Gavin's real takeaway was organizational. Optidge's response was to make client communication a named job: dedicated account managers whose primary responsibility is keeping clients informed and asking the obvious questions early. 'Are the leads arriving and are they good?' became a standing agenda item, not an afterthought. No alerting system replaces a person whose job is to ask.

What's the right way to handle it once a form failure is found?

Honesty, immediately, with proof of recovery. When Optidge discovered the problem there was no attempt to soften or hide it. The team investigated at once, pulled every lead that had been captured in the database, and handed the client a full export of everything they could recover. Many of those enquiries — submitted weeks earlier — were salvageable precisely because the data had been stored even though delivery had failed.

That recovery step is the under-appreciated insurance policy. If your handler writes submissions to a database before attempting email or CRM delivery, a delivery failure becomes an inconvenience rather than a catastrophe: the leads still exist, and you can replay them. If your form only emails and stores nothing, a broken pipe means those leads are gone forever. Build storage-first form handling and you turn a relationship-ending event into a recoverable one.

The communication script when it happens is simple and survives scrutiny: state plainly what broke, for how long, and what was lost; show what you have recovered; and explain the specific monitoring you are adding so it cannot recur quietly. Clients forgive a bug caught and owned. What ends contracts is discovering, months later, that the agency never noticed — or worse, knew and stayed quiet.

How do you build a lasting QA system around your forms?

Treat every lead form as production infrastructure that degrades over time, because it does. A durable system has four habits working together:

1. Monitor outcomes, not proxies. Alert on stored submissions and delivered leads, never on tag fires alone.
2. Schedule synthetic tests. A weekly tagged test submission verified all the way to the CRM, owned by a named person.
3. Reconcile sources regularly. Compare ad conversions, analytics, and delivered leads side by side; investigate any divergence the same week.
4. Re-test after every change. CMS update, plugin update, theme change, DNS or SMTP change, CRM remapping — any of these can sever delivery, so each one triggers a fresh end-to-end test.

The deeper point Gavin's story makes is that great PPC or SEO performance is worthless if the leads never land. Optimization upstream — better keywords, sharper ad copy, faster pages — only multiplies value when the pipe at the end actually delivers. Most teams pour energy into the top of the funnel and assume the bottom works. The agencies that keep clients are the ones who verify, on a schedule, that a real human on the other end received a real lead. Watch the whole path, not just the part that's easy to chart.