In shortA cost per lead far below what this client normally pays proves nothing on its own — it is what a breakthrough and a junk audience both look like from inside Ads Manager — so hold the budget where it is and call twenty of the leads before deciding which it is.
Most of what goes wrong in a Meta account looks bad. This one looks wonderful. The cost per lead halves, the client is happy with the volume, and the obvious move is to put more budget behind whatever is working. Sometimes that is right. Often the campaign has simply found people who fill in forms easily, which is not the same as people who buy.
Why the cost can’t tell you which
Meta optimises for the event you asked for. On a lead campaign that event is a submitted form, so delivery drifts towards whoever submits forms most cheaply: people who tap through on autopilot, people outside the target area, people who will not answer the phone. Every one of them lowers the cost per lead. Nothing in spend, clicks or cost per lead records whether the person was worth calling.
That is why this finding never says the campaign is bad. It says the number has moved far enough from this client’s normal that it needs checking before anyone acts on it, in either direction.
A lead that costs a third of normal is either your best week or the wrong audience, and only a phone call can tell which.
Where the line is
If the client has said what a lead of this quality should cost, the line is the bottom of that range. If not, it is about a third of what this client’s leads usually cost, measured from its own past campaigns. Not an industry figure: a ₹150 lead is suspicious for a real estate developer and ordinary for a salon.
How to check it
- Hold the budget where it is. Don’t raise it, and don’t cut it either: cutting a breakthrough is as costly as scaling junk.
- Call twenty of the leads, or all of them if there are fewer. Mark each one good or junk the same day.
- Look at where the cheap ones came from: Breakdown by region, placement and age. Delivery outside the target area is the most common source.
- Work out the cost per real lead: cost per lead divided by the share that were good. That number, not the headline, decides what happens next.
What happens after the calls
If most of them were real, you have found something, and the campaign can be scaled on evidence instead of hope. If few were, the cost per real lead is probably higher than the old normal, and the fix is in targeting and the form: tighter locations, a qualifying question, a higher-intent form type. Either way, send the marks back to Meta so the next round of delivery learns which leads were real.