Someone asks you to reduce cost per lead by 30%, and the fastest way to do it is also the worst: broaden the targeting, shorten the form, and start collecting people who will never buy. The figure drops within days and the business gets worse.
There is an order that works, and it is close to the reverse of where most people start. Tracking, then the offer, then creative, then the page, and bidding a long way last.
This is for you if the number is under pressure and you would rather fix it than fake it.
To reduce cost per lead, check what you are measuring first
Before changing anything, make sure the figure is real.
A conversion tag that fires twice halves your reported cost per lead. One that broke in a site update doubles it. Meta browser tracking has been unreliable since Apple’s tracking changes, so a share of your leads may never be attributed at all.
Count your actual enquiries by hand for one week and compare against the dashboard. If the gap is large, you have found your problem, and no amount of campaign work would have fixed it.
This is the least interesting hour in paid media and it is where I start on every account, because a measurement problem wearing the costume of a performance problem wastes months.
Fix the offer before you touch the targeting
No amount of clever audience work rescues something nobody wants.
The offer is what you are asking someone to exchange their details for. A free quote is not an offer, it is a formality. A specific, concrete promise is an offer, and it changes response rates more than any setting in the account.
This is also the cheapest thing to test, because it needs no new creative and no new budget. Change the words describing what happens after someone enquires, and watch what moves.
Creative is where the gains are on Meta
Audience targeting stopped being the lever some time ago. Creative now does that job.
Broad targeting with strong creative routinely beats hand-built interest stacks, because the algorithm reads who responds to the creative and finds more of them. Your ad is now the targeting.
That means a genuinely new angle, not a new colour. Same offer photographed differently is a refresh. A different reason to care is a new test.
Refresh on a rhythm rather than when something breaks. Waiting for fatigue to show in the numbers means paying for a week of decline before you react.
The optimisation event most accounts get wrong
This one deserves its own check because it is so common and so invisible.
If your campaigns optimise for landing page views, link clicks or add-to-cart when what you want is enquiries, Meta is doing exactly what you asked and finding the wrong people. The figure may even look good, because page views are cheap.
Most Meta accounts I audit are optimising on the wrong event, and it is usually because someone set it once and nobody looked again.
Fixing it often makes the headline number worse and the business better. Fewer, more expensive, more serious enquiries. That is the trade to want, and it is why this metric should never be judged alone.
The page is where you reduce cost per lead for free
Conversion rate is the lever with no auction attached to it.
Every point you gain lowers cost per lead without buying anything. Nobody bids against you. The whole search market proved the point this year: WordStream’s 2026 study of 13,474 US campaigns found costs fell because conversion rate rose for 87% of industries, not because anyone won a bidding war.
Match the page to the ad that brought the person there. Cut the form to what you genuinely need, since every extra field costs you responses. Make the phone number tappable on a phone.
What this looks like when it works
The clearest example on my books is a US auto transport account on Meta that settled at $2.68 per lead across 5,599 leads.
It had previously sat above $6, and every attempt to scale had pushed it higher. The fix was not clever. Three things changed: the optimisation event was corrected, creative moved to a weekly refresh rhythm, and the form dropped to the fields the sales team actually used.
The harder part was that $2.68 held while budget went up, which is the bit most accounts fail. Holding a figure steady while spend grows is the real skill. Getting a low number at low spend is mostly luck.
Why you cannot reduce cost per lead forever
Efficiency work has a floor, and it is worth knowing where yours is before you promise anyone another 30%.
An account with broken tracking or the wrong optimisation event can improve dramatically, because something is genuinely wrong. Once that is fixed, the gains get smaller and harder. Creative and landing page work might buy you another tenth or fifth. After that you are grinding.
At that point the honest advice changes. If you are already below your ceiling and demand exists, the better move is to increase budget rather than squeeze further, because efficiency has a limit and growth does not.
The mistake is doing both at once. Scaling pushes costs up while you are trying to pull them down, and you end up unable to read either result. Pick one for the quarter.
What to leave alone
Three things I would not do under pressure.
Do not rebuild the account because performance dipped for four days. You throw away everything the bidding has learned and spend weeks getting back to where you were.
Do not run a split test of two creatives on a small daily budget and expect an answer. Without enough conversions you are reading noise, and acting on noise is worse than not testing.
Do not buy a guaranteed cost per lead. Whoever offers the guarantee controls the definition of a lead, and they will meet the number by loosening it.
What to change this week
Four steps, in this order.
Count a week of real enquiries against your dashboard. Check which event your campaigns optimise for, in the settings rather than from memory. Look at the age of your best-performing creative. Then cut one field from your form.
The wider piece on judging this metric covers where the ceiling sits, why your close rate sets the target covers how to build one, and what a lead is worth is the calculation underneath both. For the specific ways budget disappears, the three places Meta accounts leak money is the short version. If you want this done properly, Meta lead generation is most of what I do, and you can book a teardown.