Keeping a customer costs 5 to 25 times less than winning one.

Frederick Reichheld of Bain & Company published "Zero Defections: Quality Comes to Services" in Harvard Business Review in 1990. He studied credit card companies and insurers and found that a 5% increase in retention raised profit by more than 25%. In The Loyalty Effect (1996) he widened that to 25% to 95%, depending on the industry.

The "five times cheaper" line grew out of that work. Harvard Business Review's 2014 article "The Value of Keeping the Right Customers" puts the honest range at 5 to 25 times, depending on the study and the industry. So the number is a range, not a constant, and every figure behind it predates the smartphone.

Why the gap is wider in 2026

Acquisition keeps getting pricier. Profitwell's benchmark reports customer acquisition costs up 222% over eight years, with an 18.4% rise in 2025 alone. Triple Whale, across 35,000 ad accounts, measured Meta CPMs up 20% year over year in 2025 in every industry. WordStream saw Google Shopping cost per click jump 33.7%. SimplicityDX estimates that ecommerce brands lose an average of $29 on every new customer they acquire, so the first order rarely pays for itself.

Retention compounds the other way. Gorgias found that repeat customers are 21% of a store's customer base but 44% of its revenue and 46% of its orders. Bain's data shows customers spending 67% more per order by months 31 to 36 of the relationship than in their first six months.

Most store owners know these numbers. Econsultancy found that 82% of business leaders believe retention is more cost-effective than acquisition. The spending says otherwise, because the tools decide.

The ad platform finds fast converters

Meta and Google optimize for the cheapest conversion inside a short attribution window. They find bargain hunters, because bargain hunters convert fastest. They cannot know whether the person will buy again, so they cannot optimize for it.

The email tool reaches 1 in 10

An email tool only acts on shoppers who gave an address and consent. On a typical store that is under 1 in 10 visitors (see why one agent). Chat reaches under 1 in 30 visits. Everyone else browses, leaves, and hears nothing.

Nothing shares what it learns

Search, recommendations, chat, email and ads each learn on their own and bill on their own. What search learns about a shopper never reaches the email or the ad audience. Retention needs the opposite: one memory of each visitor and one goal for every tool.

1. Own every visitor from the first visit

Start the profile at the first interaction, not the first login: the ad that brought them, what they searched, what they viewed, what they asked in chat. Then recommendations, messages and audiences all work from the same memory.

2. Give every tool the same goal

Choose what you optimize for: quick sales, bigger baskets or lifetime value. Make the site, the messages and the ad audiences work toward it, per person. That is what the agent does; the how it decides page shows the mechanics.

3. Tell your ads who not to find

Send predicted customer value to Meta and Google, and keep your existing customers out of acquisition campaigns. In one retailer's data, shoppers who were already emailable were 4.9% of ad clickers and 29% of revenue, so the ads were paying for customers the store already had (use case).

4. Measure the lift, not the opens

A holdout group without the agent shows what retention added, in the store's own orders. In one fashion retailer's data, Releva-influenced traffic converted 130% better at half the ad spend (case studies).