The Form Still Works. The Leads Are Dying Somewhere Else.
Every few months, a business owner tells me their website contact form "stopped working." Almost every time, it hasn't. The form submitted just fine. Somebody's spam filter ate the notification, or it landed in an inbox nobody checks, or it arrived on a Friday and got answered on Wednesday.
That difference matters, because it changes what you fix. Rewriting a form that works is expensive theater. Finding out where a working form's output goes is a twenty-minute job.
Think of the form as a pipe, not a page
A contact form is three things in a trench coat: a front end that collects data, a delivery mechanism that moves it, and a human who responds. Most audits only look at the first one — the layout, the button color, the number of fields.
The other two break far more often.
So here's the walk-through. Five places leaks happen, roughly in order of how often I see them.
Leak one: nothing is listening
The single most common failure. The form sends an email to info@ — an address that was set up in 2019 for a web designer who left, or that forwards to a personal account nobody migrated when the domain changed hosts.
You can spot this in about ninety seconds. Submit the form yourself from a personal email address, then go check the destination inbox directly. Not the notification on someone's phone. The actual inbox.
If the mail server changed recently, also check the sending domain's authentication records — SPF and DKIM. A form notification that fails authentication is a notification that may get quietly dropped by the recipient's provider. This is a two-minute check with a free online tool and it solves a surprising share of "the form is broken" tickets.
Leak two: the spam filter is overachieving
Spam filtering is a blunt instrument, and business inquiries look a lot like spam. Short subject line. Generic greeting. A link to a document. Someone typing "Hi, I'd like a quote" from a phone at 11 p.m.
Spam tools also get progressively more aggressive, which is why a form can work for two years and then stop. Nobody changed anything. The filter just learned.
Two fixes worth doing:
- Whitelist the form's sending address so it bypasses quarantine entirely and lands in a folder a human reads.
- Add a honeypot field to the form itself — a hidden input that bots fill in and people never see. That catches the actual spam so the filter doesn't have to be so paranoid about legitimate messages.
Skip the CAPTCHA if you can. Half the value, twice the abandonment.
Leak three: the phone kills it
Pull out your phone and fill out your own form. Right now. All the way through.
Most small-business forms were tested on a laptop at a desk. On a phone, a two-column layout collapses badly, autofill puts a zip code in the company field, and the submit button ends up under the keyboard. People don't fight forms anymore. They leave.
Three specific things to check on mobile: whether autofill works on the email and phone fields, whether the submit button is reachable without scrolling sideways, and whether the form still submits if the visitor loses signal for ten seconds in a parking garage.
That last one sounds fussy. It isn't — people fill out forms in cars.
Leak four: the dead end
A visitor hits submit. The page reloads to the same form, blank. No confirmation, no thank-you, no email receipt. They assume it didn't go through. So they submit again. Then they leave, or call, or give up.
You now have duplicate entries in a spreadsheet nobody monitors, and one fewer customer.
A thank-you page should do two jobs: confirm the message arrived, and set an expectation. "Thanks — we reply within one business day. Here's our phone number if it's urgent." That sentence alone reduces callbacks and sets up the fifth leak, which is the expensive one.
Leak five: the reply that never happened
Studies keep landing in the same neighborhood on this: the odds of ever reaching a lead collapse after the first hour, and get ugly after the first day. For a small business, the practical version is simpler — if you promise a one-day reply, the clock starts when the form is submitted, not when someone happens to open the inbox.
A few things that actually work without buying new software:
- Route form notifications to a shared inbox or channel, not one person's address.
- Assign a named responder for each weekday, even if the assignment is just a rotation on a whiteboard.
- Write a two-sentence acknowledgment template so replying doesn't require composing from scratch.
- Set a daily five-minute check as a calendar event. Not a habit. A calendar event.
The goal isn't a sales process. It's that a legitimate inquiry gets a human response before the prospect has moved on.
The twenty-minute monthly test
Once a month, put a reminder on the calendar and do this:
- Submit the form from a personal email on a phone, using a fake last name you'll recognize.
- Time how long until the notification appears in the shared inbox.
- Confirm the thank-you page renders and the phone number on it is current.
- Reply to the fake submission within an hour so you've tested the whole loop.
- Note the result somewhere — a line in a shared doc is enough.
Five lines a month. When something changes, you'll see it in the note instead of hearing about it from a customer who gave up in September.
What you're really auditing
Not a web page. A handoff. Somewhere between a stranger typing their email and a person on your team typing back, there are four or five places the message can vanish, and none of them are visible from the design tool you used to build the form.
The form was probably fine the whole time. The pipe behind it needed a look.