No invented statistics. Every number is sourced or labelled as an example. Our standards

Myth Busting

Myth: "A new customer costs five times more than keeping one" — the car-maker sum behind the slogan

The claim:
It costs five times more to acquire a new customer than to keep an existing one (some versions say 5 to 25 times).
Misapplied

You will have seen this one on a conference slide or in a CRM vendor's ebook: winning a new customer costs five times as much as keeping an existing one. Newer versions stretch it to "five to 25 times". It gets used to justify retention budgets and account management hires. If the number is wrong for your business, those decisions are resting on someone else's sums.

The good news is that this stat does come from somewhere real. The bad news is where: one company, in one industry, in one country, roughly forty years ago.

Where the claim comes from

Most people who repeat it credit Frederick Reichheld at Bain & Company, the consultant later known for inventing the Net Promoter Score. That is an understandable guess, as his loyalty work is usually quoted alongside it. But when the authors of the 2005 book Loyalty Myths went looking for the origin of the "five times" line, the earliest sources they found pointed somewhere else: to research by TARP (Technical Assistance Research Programs), a Washington, D.C. consultancy that studied customer complaints, in the late 1980s. They also noted that other firms were claiming the same finding as their own around that time, and that a 1990 Harvard Business Review article and Tom Peters' Thriving on Chaos helped spread it (Ipsos Loyalty excerpt, 2005).

What we could actually find

The clearest account of the original calculation comes from TARP itself. In a 1999 article, John Goodman, TARP's former president, set out to correct the record on the "facts" that had been attributed to his firm, some of which he said were garbled and some simply wrong (Goodman, Basic Facts on Customer Complaint Behavior, 1999; the copy we read is hosted on a third-party site).

According to Goodman, the five-times figure came from a study for a single US car manufacturer. TARP compared the advertising cost of winning a new car buyer with the goodwill spending (fixing complaints and problems) needed to keep an existing owner. The company spent $375 in advertising per car sold. With roughly half its buyers being repeat customers, TARP reckoned the real advertising cost per genuinely new customer was about $750. The goodwill cost of keeping a customer averaged $150. $750 divided by $150 is five. Goodman says this work helped justify the customer service freephone lines of the early 1980s, and points to a 1984 Business Week article as a record of it. That would put the calculation a few years earlier than the late-1980s sources the Loyalty Myths authors found.

Then comes the line that almost never gets quoted: Goodman wrote that TARP had since found the real ratio varies from 2 to 1 up to 20 to 1. By 1999 the firm behind the number was itself saying five was not a constant.

The "five to 25 times" version is harder to pin down. The most-cited source for it is a 2014 Harvard Business Review piece by Amy Gallo, which says acquiring a customer is five to 25 times more expensive "depending on which study you believe, and what industry you're in". It does not name the studies behind that range (Gallo, HBR, 2014). We couldn't trace the 25 upper bound to a primary source; the earliest widely cited version we found is that HBR article.

Gallo's article does cite Reichheld for a separate, related claim: that a 5% rise in retention lifts profits by 25% to 95%. The Bain document it links to is a 2001 one-pager by Reichheld, which says that in financial services a 5% increase in retention produces more than a 25% increase in profit. It does not mention a 95% figure or any acquisition-to-retention cost ratio (Reichheld, Bain, 2001). Reichheld and Sasser's 1990 HBR article "Zero Defections", which Bain summarises as showing profits can rise by almost 100% from keeping 5% more customers, is the root of that retention-profit idea (Bain summary). We could read only Bain's summary, not the paywalled article; it contains no five-times cost ratio.

That claim has critics too. Byron Sharp of the Ehrenberg-Bass Institute argues the 1990 figures came from a thought experiment that assumed defections could be halved at no cost (Sharp, Ehrenberg-Bass).

What the evidence does support

This is not an invented number. The TARP figure was a real calculation, and its direction is reasonable for many businesses: winning a stranger usually costs more than looking after someone who already trusts you. The problem is what has happened to it since:

The Loyalty Myths authors add that advertising reaches existing customers too, so booking all of it as "acquisition" inflates the ratio, and that your most valuable customers are often the most expensive to keep (Ipsos Loyalty).

So the verdict is misapplied. A real finding from a narrow context has been stripped of its limits and used everywhere.

What to do instead

The only ratio that matters is yours, and you can usually work it out in an afternoon.

  1. Work out your acquisition cost. Add up sales and marketing spend for a period, including salaries, and divide by new customers won. Our customer acquisition cost calculator walks you through it.
  2. Estimate what you spend keeping customers. Account management time, support beyond the contract, renewal discounts and any retention offers. Be honest about the discounts; they are often the biggest line.
  3. Put a value on a kept customer. A cheap-to-keep customer who spends little may be worth less than an expensive new one. The customer lifetime value calculator helps you compare like with like.
  4. Check your churn first. If you lose few customers already, extra retention spend may buy very little.

As an illustration, not a benchmark: a UK firm spending £60,000 a year on sales and marketing to win 40 customers pays £1,500 per customer. If keeping one costs £400 a year, the ratio is under four to one.

Retention is almost always worth taking seriously. Just make the case with your figures, not a car maker's from the early 1980s.

Sources we checked

Spotted a mistake or a number we can't back up? Tell us and we'll fix it.