Facebook lead ads produce junk because the default form is built to be submitted by accident.
Meta pre-fills the name, email and phone number from the person's profile, picks the "more volume" form type for you, and serves the ad into Reels and Stories where people are scrolling with a thumb. One tap opens the form. A second tap sends it. Nobody has read anything.
The fix is in what the form asks. This article covers why the leads come out the way they do, the form settings Meta gives you, the four questions that filter a UK enquiry, and how to get the lead to a phone within minutes.
Why instant forms produce low-intent leads
An instant form removes every step that used to prove someone was interested. A website enquiry means the person clicked an ad, waited for a page, read some of it, found the form and typed their details. An instant form skips all of that. Ben Heath of the UK agency Heath Media wrote back in 2018 that lead quality from lead ads was "consistently worse" than from a website for this reason: people are not invested in the process.
Three mechanics do the damage. The first is pre-filling. The form opens with every field already populated from the Facebook profile, so the person sees a completed form and taps the button. A September 2026 write-up from upGrowth describes the whole sequence finishing in under three seconds, which is why so many of these people later say they never filled anything in. They did, but it was a reflex.
The second is the default form type. Meta offers three and "more volume" is the one you get unless you change it. Jon Loomer, who has written about Meta ads since 2011, describes it as built for maximum sign-ups. Volume is what you get.
The third is placement. Advantage+ placements put the ad into Reels, Stories and the Audience Network, where accidental taps on a call-to-action are common. A lead from a fast-scroll placement at 11pm with a pre-filled phone number has done nothing to earn a call back.
The format itself is fine. Meta's own lead generation guide reports a 20% lower cost per qualified lead for instant forms against website forms, from 15 advertiser A/B tests run between November 2022 and February 2023. The leads are cheap and some are good. The job is to stop paying for the ones that are not.
The form settings Meta gives you
Meta gives you three form types, a review screen, up to 15 questions and conditional logic, and most accounts never touch any of them. The PPC Hero walkthrough of the form editor lists the types as more volume (the default), higher intent (the same form plus "an additional review step that gives people a chance to confirm their information before submitting it") and rich creative, which adds images and product sections but only shows in the Facebook mobile app.
The review screen is the cheapest fix available. It appears only on the higher intent type, there is nothing to set up, and it turns a one-tap submission into two taps with a chance to see what is about to be sent. Loomer calls it a good starting point and says you can do better, which is fair. It catches accidents. It does not catch the person who meant to tap but has no budget and no timeframe.
Custom questions are where the filtering happens. Loomer's guide to instant forms lists the types as multiple choice, short answer, conditional, appointment request and slider, and notes that any type except conditional can be made optional. Meta also suggests questions based on what you asked before. Treat those as a starting point.
Conditional logic lets a multiple choice answer end the form. Once it is on, a logic column appears against each answer with three options: go to a question, submit form, or close form. Close form sends the person to an end screen and their contact details are never submitted. Anna Sonnenberg's Social Media Examiner piece notes that Meta itself warns conditional logic may raise cost per lead. That is the point.
One more setting sits outside the form. The performance goal at ad set level defaults to maximising the number of leads. The alternative, conversion leads, needs your CRM to send lead status back through the Conversions API, and Meta's guide reports a 28% higher rate of leads becoming quality leads and a 16% lower cost per quality lead across 68 advertisers tested in August 2023.
Do it later, once the form and the follow-up are fixed. It will not rescue a bad form.
The four questions that filter a UK enquiry
Four questions filter an enquiry well enough for a service business: postcode, timeframe, budget band and a free-text description of the job. Each does a separate job and the order matters.
- Postcode. Ask for the outward code (NN1, MK9, OX16) as a typed short answer, not a dropdown. Meta can pre-fill a city, and "Northampton" tells a roofer very little. A typed outward code tells you in two seconds whether the job is in your area, and the act of typing it is the first bit of friction.
- Timeframe. Multiple choice with conditional logic on: "within a month", "one to three months", "later this year", "just looking". Route "just looking" to close form. If you want those people for a newsletter, run a separate campaign for them with its own form.
- Budget band. Multiple choice again, with bands that match how you quote. A kitchen fitter might use "under £5,000", "£5,000 to £12,000", "over £12,000". A solicitor might ask about the matter type instead. Route the band you never take on to close form.
- Describe the job. One short answer field: "Tell us what you need in a sentence or two." This is the question Meta tells you not to ask, and it is the one that sorts the enquiry list better than any other.
Here is how that compares with the form most accounts are running.
| Element | Default form (what most accounts run) | Filtering form |
|---|---|---|
| Form type | More volume | Higher intent (review screen on) |
| Contact fields | Full name, email, phone, all pre-filled | Full name, phone, email. Phone marked required |
| Location | None, or city pre-filled | Typed outward postcode |
| Timeframe | Not asked | Multiple choice, "just looking" closes the form |
| Budget | Not asked | Multiple choice bands matching your quoting |
| Job description | Not asked | Typed short answer, required |
| Conditional logic | Off | On, with two close-form routes |
| Questions in total | 2 or 3, all pre-filled | 6, three of them typed |
| Thank-you screen | "Thanks, we'll be in touch" | States who will call, from which number, and when |
Heath's 2019 list of lead quality fixes puts the phone field first: the default form collects name and email only, and email follow-up is easy to ignore. Make the phone field required even though it is pre-filled, because the follow-up that works is a call, the same reason Google Ads for tradesmen leans on call assets over forms.
Meta says avoid open-ended questions, and we disagree
Meta's advice is to "avoid asking too many questions, and avoid open-ended questions." The same guide tells advertisers to prioritise questions that auto-populate. That advice protects completion rate, and it is correct if completion rate is what you are measuring. For a UK trade or professional firm with one person doing the call-backs, completion rate is the wrong number.
"If you want to increase the quality of the leads generated from these forms, the answer is simple: Add more friction." Jon Loomer, How to Get High Quality Leads from Facebook Lead Ads
A typed sentence about the job does two things a dropdown cannot. It gives the caller something to open with ("you mentioned the back extension") instead of "did you fill in a form?". And it exposes the enquiry that was never a job: a one-word answer, a competitor's name, a string of letters.
Expect fewer leads and a higher cost per lead. One PPC, a UK agency, works a hypothetical example where the higher intent form costs £26.67 per lead against £20, and the cost per qualified lead falls from £133.33 to £66.67. Their numbers are illustrative, but the shape is what we see: lead cost up, cost per usable enquiry down.
If your reporting stops at cost per lead, the better form will look like a worse result. Report cost per lead you called back and wanted. It is the Meta version of rankings that rise while enquiries stay flat: the vanity number moves, the business one does not.
Get the lead to a phone in minutes, not days
The lead needs to reach a phone within minutes because a Facebook lead goes cold faster than any other kind. The person did not search for you. They tapped something between two videos. By the next morning they may not remember doing it, and Heath's experience is that a lead called two or three days later is far less likely to become a customer.
The research on this is old but has not been overturned. A 2011 Harvard Business Review audit by Oldroyd, McElheran and Elkington, The Short Life of Online Sales Leads, submitted test enquiries to 2,241 US firms. As summarised here, the average first response took 42 hours, 37% answered within an hour and 23% never replied. Those were website leads from people who had gone looking. Instant form leads are weaker to begin with.
Downloading a CSV from Ads Manager is not a follow-up process. Whoever downloads the file on Friday is calling Tuesday's leads on Monday. The fix is a webhook. Make lists a Facebook Lead Ads app with a "New Lead" trigger that fires on submission, and Zapier has the same. Either one can run a scenario that does three things at once:
- Posts the lead to a WhatsApp group or Slack channel with the postcode, timeframe, budget band and the typed job description, so whoever is on call can ring within five minutes.
- Writes the row to a CRM or a Google Sheet with a timestamp, so response time is measured, not guessed.
- Sends the enquirer a text from a named person saying when they will call, which heads off the "I never filled in a form" conversation.
Setting this up takes an afternoon. In Business Manager you assign Make or Zapier under Settings, Integrations, Leads Access, then map the form fields in the scenario. We build the same flow for clients as a workflow automation job, and it is usually the cheapest change in the account.
Once that is running, the CRM status becomes the signal you send back to Meta for the conversion leads goal. Fix the form, fix the follow-up, then let Meta learn from the leads you marked as good.
The rebuild checklist
- Change the form type from more volume to higher intent. The review screen needs no setup.
- Keep full name, email and phone. Mark phone as required.
- Add a required short answer question for the outward postcode.
- Add a multiple choice timeframe question. Turn on conditional logic and route "just looking" to close form.
- Add a multiple choice budget or matter-type question with bands that match your quoting. Route the band you do not serve to close form.
- Add a required short answer question: "Tell us what you need in a sentence or two."
- Write the thank-you screen with a named person, the number they will call from and a time window.
- Check the placements. If Audience Network or Reels is sending the junk, exclude it for a fortnight and compare.
- Connect the form to Make or Zapier and build the alert, the CRM row and the text message. Test it with a live submission from your own phone.
- Change the report. Track cost per lead you called and wanted, and response time in minutes, alongside the cost per lead Meta shows you.
- Once you have a run of leads with CRM statuses, switch the performance goal to conversion leads and connect the Conversions API.
Expect the lead count to drop when you publish the new form. That drop is the people who were never going to pick up the phone.
The decision rule: if the person running your call-backs would rather have ten enquiries they can act on than forty they have to sift, add the friction. The way to judge whoever runs the account is the same checklist you would use to pick a UK Google Ads agency. If you are building a list for a long nurture sequence and volume is the goal, leave the default alone and know what you are buying.
If a Meta account is producing leads nobody wants to call, our paid media team will go through the forms, placements and follow-up with you.