The Cheapest Lead Is Rarely the Cheapest Customer

Cost per lead is the most popular number in performance marketing for a simple reason: it's available immediately, it's easy to compare, and it moves when you change something. Lower it and the report looks better by the next morning. That makes it a very good number for managing a campaign, and a very risky one for managing a business.

The problem is what cost per lead doesn't include. It tells you what it cost to get someone to raise their hand. It says nothing about whether they were ever going to buy, whether anyone called them back, or what they were worth if they did. A business that optimises on cost per lead alone is, in effect, optimising for the cheapest hand-raise, and cheap hand-raises have a habit of being cheap for a reason.

The pattern shows up in very different businesses. In one engagement written up on Selected Engagements, a high-spend rehabilitation facility had campaigns whose cost per lead looked expensive enough to cut. The facility's own monthly verification and admission reports, which nobody had translated into anything usable, told a different story: some of the "expensive" campaigns consistently produced verified admissions, while some of the cheaper leads stalled at the very first verification step. Cutting on cost per lead would have cut the best source of patients. The full diagnosis is here.

In another, a regional builder's campaign was producing leads, until it emerged that around 43% of them were outside the service area entirely. Whatever those leads cost, they were worth nothing to the business, and they were quietly inflating the real cost of the ones that mattered. That story had a second twist: many of the leads that were in the area were never being called, so the "lead quality problem" was partly a follow-up problem wearing a disguise.

Fabien saw the same thing from the other side running his own lead generation agency across home-improvement trades. Raw cost per lead could swing wildly by vertical, sometimes tenfold on the same day, and installers who complained about "bad leads" were often losing the sale somewhere in their own follow-up. By the end, leads were being qualified on cost per acquisition, conversion rate and lifetime value before volume was even discussed. Across that agency's campaigns, documented cost per lead ranged from under £2 to nearly £20, numbers that only mean something once you know which stage of the customer journey they came from (Connecting the Click to What Came After).

A lead is an input into the commercial system. The customer is the output. Price the output.

None of this means cost per lead should be ignored. It means it should be read as the first number in a chain, not the last. The useful questions sit further down:

• Cost per qualified lead. What does it cost to generate a lead that meets a definition sales has actually agreed to? If there's no agreed definition, that's the first thing to fix, see What Counts as a Qualified Lead Isn't Universal.
• Cost per booked conversation. An appointment, a site visit, a demo, an intake call: the first point where the lead has done something a business can act on.
• Cost per customer. What it actually cost, in media and in sales time, to win one paying customer from each source.
• Value per customer. First-order value, margin and, where it's knowable, lifetime value. A source that costs twice as much per customer can still be the best one if its customers are worth three times as much.

Most businesses already hold more of this data than they think. It's sitting in a CRM, a booking system, an intake spreadsheet or a monthly report from another team. The gap is usually translation, not access, the same point made in "We Don't Have the Data" Is a Claim Worth Checking.

What people ask about lead cost and lead value.

If your budget decisions are still being made on cost per lead alone, a Commercial Diagnostic is a 90-minute session that traces where your leads actually go after they arrive, and which sources are worth paying more for.