- Start with the form, not the audience
- The follow-up speed problem nobody wants to admit
- Choose the right lead objective and let Facebook's algorithm work
- Budget and testing structure that gets you data
- Creative that stops the scroll for lead gen specifically
- Retargeting the people who didn't submit
- Where the algorithm can't save you
- Fixing what happens after the click
- When to bring in outside help
- Frequently asked questions
- Official documentation
The short version: most Facebook lead ads fail because of the form, the follow-up speed, or the offer, not the targeting, and fixing those three things usually drops cost per lead by 30 to 50 percent faster than any audience tweak. I've run this exact diagnostic on client accounts more times than I can count, and the pattern never changes.
I want to start with something uncomfortable. Most people optimizing Facebook ads for leads are optimizing the wrong layer entirely. They're in Ads Manager fiddling with age ranges and interest stacks while their lead form takes eleven fields to fill in and nobody calls the lead back for four days. You can have the best targeting on the platform and still get rubbish results if the machinery behind the ad is broken.
Start with the form, not the audience
Facebook's own data has shown for years that every extra field on a lead form drops completion rate. I tested this directly with a UK-based training company in 2023. Their original form had eight fields: name, email, phone, company, job title, company size, budget range, and "what's your biggest challenge." Cost per lead was sitting at £34.
We cut it to three fields: name, email, phone. Nothing else. Cost per lead dropped to £19 within ten days, same budget, same audience, same creative. That's not a small tweak, that's nearly a 45 percent reduction from removing five questions.
The instinct to qualify harder on the form feels smart. It isn't. Qualification belongs in your sales process, not your ad form. Every field you add is a tax on volume, and most sales teams would rather have twenty warm leads to sort through than eight perfectly pre-qualified ones that never arrive.
What belongs on the form
- Name and email, always, because you need these regardless
- Phone number only if your sales process calls people within the hour
- One qualifying question maximum, and only if it changes what happens next (for example, routing enterprise leads to a different rep)
The follow-up speed problem nobody wants to admit
Here's the bit that makes marketers uncomfortable: your ad optimization doesn't matter if your sales team responds to leads in two days instead of two minutes. I've seen accounts with a beautifully low cost per lead and an atrocious conversion rate to sale, and the reason was never the ad. It was that leads sat in a CRM queue over a weekend while a competitor called within five minutes.
There's research from a widely cited Harvard Business Review study on lead response times showing that companies contacting leads within an hour were nearly seven times more likely to qualify them than those waiting even an hour longer. That's not a Facebook stat, that's a sales process stat, but it directly determines whether your Facebook budget is worth spending at all.
If you're running lead ads and your average first-contact time is over 24 hours, fix that before you touch a single ad set. No amount of algorithmic optimization on Facebook's side compensates for a slow human on your side.
Choose the right lead objective and let Facebook's algorithm work
Facebook's Conversion Leads objective (built on top of the standard Lead Generation objective) lets you optimize toward qualified leads rather than just form completions, provided you're feeding conversion data back through the CRM integration or Conversions API. This matters more than people realize. If you're only optimizing for "lead submitted," Facebook will happily find you the cheapest, laziest form-fillers on the platform. Feed it data on who became a customer, and the algorithm starts finding you more people like your real customers.
Practically, this means connecting your CRM (Salesforce, HubSpot, or even a Google Sheet via Zapier) so that when a lead becomes qualified or closes, that signal goes back to Facebook. Give the algorithm at least 50 conversion events in a 7-day window before judging performance. Cutting a campaign at day 3 because cost per lead looks high is the single most common mistake I see, because Facebook's learning phase needs volume before it stabilizes.
Budget and testing structure that gets you data
Underfunding tests is the second most common mistake. If your average cost per lead is £15, and you're testing three creatives with a £10 daily budget across three ad sets, you're not testing anything, you're guessing. Facebook needs roughly 15 to 20 leads per ad set within a week to make a reasonable read on performance.
My working rule: multiply your target cost per lead by 20, and that's your minimum weekly test budget per variation. If CPL target is £20, you need £400 a week minimum to get a real signal from one creative variant. Most small businesses split this too thin across too many tests and then wonder why nothing shows a clear winner.
A simple testing sequence that works
- Week 1 to 2: test three distinct creative angles (problem-focused, proof-focused, offer-focused) with the same audience and form
- Week 3: keep the winner, test two form variations (three fields vs. one field) against it
- Week 4: keep the winning combination, test broad audience vs. one interest-based audience
This sequence isolates variables instead of changing five things at once and guessing which one moved the needle.
Creative that stops the scroll for lead gen specifically
Lead ad creative works differently to e-commerce creative. You're not selling a product on the spot, you're earning the right to a name and email. The best-performing lead ads I've built or reviewed almost always lead with a specific, believable outcome rather than a generic benefit.
"Get a free marketing audit" underperforms "See exactly why your Facebook ads aren't converting, in a 20-minute call" by a wide margin in every account I've tested this in, sometimes double the click-through rate. Specificity beats polish every time in lead gen creative. This is also exactly what makes the difference between an ad campaign that converts leads into actual customers and one that just generates cheap, unqualified names.
Retargeting the people who didn't submit
Most businesses running lead ads only build campaigns targeting cold audiences and never retarget the people who clicked but didn't submit the form. That's a mistake, because these are your warmest non-converters. Build a custom audience of everyone who engaged with the ad or landed on the form in the last 14 days but didn't convert, and run a second, shorter ad at them with a slightly different angle, often a testimonial or a deadline.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
In one account I managed for a B2B SaaS client, this retargeting layer alone accounted for 22 percent of total leads at a cost per lead nearly 40 percent lower than cold traffic, simply because these people already knew who we were.
Where the algorithm can't save you
There's a limit to optimization, and I say this as someone who has spent years being paid to squeeze more out of ad accounts. If your offer is weak, if nobody wants what you're selling at the price you're selling it, no amount of form-shortening or budget structuring fixes that. I've had honest conversations with clients where the answer wasn't "let's test another creative," it was "your offer doesn't match what this audience needs right now." That's an uncomfortable thing to say to someone paying you to run their ads, but it's true more often than the industry admits.
If you're not sure whether your problem is execution or offer, that's usually the moment to bring in outside eyes. Working with an AI consultant for small business or someone who's audited enough accounts to spot the pattern quickly can save months of guessing.
Fixing what happens after the click
All of the platform troubleshooting in the world doesn't help if your pixel is misfiring or your form has a bug nobody caught. Before assuming a targeting problem, check the basics: is the lead form delivering to your CRM, is the pixel firing correctly, is your Page in good standing. Most of these live in the standard diagnostics covered in a good Facebook troubleshooting guide, and I've seen a "targeting problem" turn out to be a broken Zapier connection that had been silently failing for three weeks.
When to bring in outside help
If you've done all of the above and the numbers still don't move, it might not be a Facebook problem at all, it might be a lead generation strategy problem. That's worth separating out clearly, because the questions to ask are different. If you're evaluating outside support, it's worth knowing how to find lead generation services that deliver, and equally worth knowing what to ask a lead generation agency before you sign anything, because plenty of agencies will happily take your budget and optimize the wrong layer for months.
If you want the full set, start at the Facebook Help: 80 Guides to Pages, Groups, Ads, Stories and Fixes.
Related: my ai consultant vs ai agency small business page.
Related: writing for us on b2b saas.
Frequently asked questions
How many fields should a Facebook lead form have?
Three or fewer whenever possible: name, email, and phone. Every additional field reduces completion rate, and in my testing, cutting from eight fields to three dropped cost per lead by 45 percent with no drop in lead quality once follow-up speed improved.
What is a good cost per lead on Facebook in 2026?
It varies enormously by industry, but B2B service businesses in the UK typically see £15 to £45 per lead, while consumer offers can run from £3 to £20. The number matters less than the ratio between cost per lead and your actual close rate to paying customer.
How fast do I need to follow up with a Facebook lead?
Within five minutes if possible, and within one hour at the absolute latest. Leads contacted within an hour are dramatically more likely to convert than those contacted later, and this single factor often matters more than any ad optimization.
Should I use the standard Lead Generation objective or Conversion Leads?
Use Conversion Leads and connect your CRM data if you have at least 50 qualified conversion events a month to feed the algorithm. Without enough data volume, stick with the standard lead objective and optimize your form and follow-up process instead.
