Start with the creative — it moves CPL the most
The single biggest lever on cost per lead is the ad itself. A stronger hook in the first three seconds lifts click-through rate, which lowers CPM-driven costs and feeds the algorithm better signal. Before touching budgets or bids, test genuinely different angles — problem-first, proof-first, offer-first — not just colour swaps.
When a winner emerges, don't just duplicate it. Keep the winning variable (hook, offer, or format) and vary the rest to fight fatigue.
Reduce form friction — carefully
Shorter forms and instant (in-platform) lead forms almost always lower CPL. But there's a catch that connects straight to lead quality: the easier the form, the more low-intent leads you collect. A single extra qualifying question can raise CPL slightly while dramatically improving closing rate.
That trade-off is the whole game. Optimise for cost per qualified lead, not cost per lead — sometimes a higher CPL is the cheaper path to a customer.
Then, and only then, tune audience and bidding
Broad audiences with strong creative usually beat narrow interest-stacked audiences on modern Meta. Give the algorithm room and a clean conversion signal. Avoid over-segmenting budgets so thin that no ad set ever exits the learning phase.
- •Consolidate ad sets so each can gather enough conversions to learn.
- •Optimise for the deepest reliable event (lead, or better, qualified lead/purchase).
- •Don't edit winning ad sets constantly — every significant change resets learning.
The quality caveat
You can always lower CPL by asking for less and targeting broader — but if those cheaper leads don't buy, you've made the wrong number better. Track what each change does to closing rate, not just to CPL. Cheap leads that never convert are the most expensive kind.