There's a common assumption in lead generation that a lower cost per lead target means better performance. It's not true.
What a Cost Per Lead Target Actually Does
When you set a cost per lead target, you're not just telling Google what you're willing to pay.
You're telling it how selective to be about which searches and audiences it bids on. A tight target means the algorithm will only chase a click when it's highly confident that the click will turn into a lead, which sounds sensible, but in practice it shrinks the pool of homeowners you're eligible to show up for.
A campaign set to a £15 cost per lead target might generate 2 leads a month. A campaign set to £40 might generate 100. The second campaign looks worse on paper, but it's the one actually filling the diary.
Set the target too tight, and the platform becomes so cautious you're barely getting an impression, let alone a booked survey.
A Real Example From Home Improvement
We see this constantly with roofers, kitchen and bathroom remodelers, and window and door companies. A client wants the lowest possible cost per lead, so the target gets set aggressively low from day one. The campaign spends its budget slowly, delivers a handful of leads, and everyone concludes that lead gen "doesn't work" for their trade.
In reality, the campaign was never given room to find volume. It was built to be cautious, not productive.
What You Should Set Your Target At Instead
A more useful starting point isn't "what's the cheapest lead I can get." It's "what's the most I can pay per lead and still turn a profit, based on my close rate and average job value."
That's your ceiling. Set the target there, let the campaign run for long enough to gather data, and adjust based on what actually books.
The Goal Isn't the Lowest Cost Per Lead
The goal is a profitable volume of leads that convert into booked jobs, not the smallest number in a spreadsheet. A cheaper lead that never turns into a job is worth less than a slightly more expensive one that does.
If your campaigns are underperforming and you're not sure whether your targets are part of the problem, we're happy to take a look.
.png)
.png)
