Customer retention: 5% more, 95% more profit? The origin

The claim that increasing customer retention by 5% raises profits by 25% to 95% comes from Frederick Reichheld of Bain and W. Earl Sasser of Harvard Business School, but the original article, published in Harvard Business Review in 1990, said 25% to 85%. The examples were unnamed American service companies, and in the case the article details, “5%” meant five points less in the annual customer defection rate, which amounted to cutting defections in half. The 95% only appears ten years later, in an article about e-commerce. The other famous line, that acquiring a customer costs 5 to 25 times more than keeping one, has no study with a sample behind it: “five times” shows up in 1990 as expert opinion, and “25 times” arrives in 2014 in a piece that cites no source. Later academic research confirms that retention is the strongest lever of customer value and shows that long-standing customers are not always the most profitable, especially in B2B.

I hear both numbers in almost every conversation about Customer Success, and almost nobody in the room knows where they came from. I went back to the original texts to separate what they measured from what repetition added, with the board that decides how much to invest in the installed base and the leaders who answer for recurring revenue in mind.

Where does “retaining 5% more customers raises profit by up to 95%” come from?

Reichheld and Sasser published “Zero Defections: Quality Comes to Services” in the September and October 1990 issue of HBR (Reichheld and Sasser, 1990). The thesis was to bring manufacturing’s “zero defects” logic to services, treating the customer who does not come back as the service company’s scrap heap. The authors say they found the same pattern in more than 100 companies across two dozen industries analyzed by Bain: profit per customer grows each year of the relationship, because the customer buys more, costs less to serve, accepts a higher price and refers other customers.

What the 1990 article measured

The profit figures come from three examples: reducing defections by 5% generated 85% more profit in a bank’s branch system, 50% in an insurance brokerage and 30% in an auto service chain. The chart that goes with the text is titled with the 25% to 85% range, and the opening rounds the thesis up to “almost 100%”. The article describes no sample, method or data collection period, and the profit it calculates is customer value, the present value of the profits a customer generates over the relationship, and not the company’s accounting profit the following year.

The detailed example is a credit card company that spent an average of 51 dollars to acquire a customer. With the annual defection rate falling from 20% to 10%, the customer’s average life doubled from five to ten years and the customer’s value rose from 134 to 300 dollars. With defections falling from 10% to 5%, the average life doubled again and the value went to 525 dollars, a 75% increase. The curve uses the parameters of a single company and shows how the calculation works, without measuring several cases.

The “5%” is percentage points

In that example, the extra 5% of retention takes defections from 10% to 5% a year, or retention from 90% to 95%. For a company that loses 10% of its base each year, that means cutting defections in half and doubling the customer’s average life. In board presentations, the line usually arrives as if moving retention from 80% to 84%, a 5% relative improvement, were enough to double profit.

Where the 95% was born

The 25% to 95% range first appears, as far as I could trace it, in “E-Loyalty: Your Secret Weapon on the Web”, which Reichheld published with Phil Schefter in the July and August 2000 issue of HBR (Reichheld and Schefter, 2000). The piece, about online retailers, summarizes the 1990 research as 25% to 95% more profit, without explaining the new ceiling. In a short Bain piece, Reichheld uses a more modest version: in financial services, a 5% increase in retention produces more than a 25% increase in profit (Reichheld, Bain & Company, 2001). The version that spread came from a 2014 HBR piece by Amy Gallo, which attributes the 25% to 95% range to Reichheld with no link to the study (Gallo, HBR, 2014).

Version of the number	Where it appears	What can be verified
85%, 50% and 30% more profit	Reichheld and Sasser, HBR, 1990	Three unnamed service companies, no published method
25% to 85%	Chart in the same article	Range of the examples
25% to 95%	Reichheld and Schefter, HBR, 2000	Wider range with no explanation
More than 25% in financial services	Bain piece, 2001	Claim with no data
25% to 95% as Reichheld’s research	Gallo, HBR, 2014	Repetition with no link to the study

Does acquiring a customer cost 5 to 25 times more than retaining one?

I found no study with a sample and method that supports this ratio. The oldest version I located is in another 1990 HBR article, by Christopher Hart, James Heskett and the same Earl Sasser (Hart, Heskett and Sasser, 1990). The authors write that replacing a customer costs five times more than retaining one and attribute the figure to the agreement of most industry experts. The article’s footnote, a 1986 study on complaints prepared for the US government, supports a different figure, repurchase after a resolved complaint.

The “25 times” comes from Gallo’s 2014 piece, which says that, depending on the study and the industry, acquiring a customer costs five to 25 times more than retaining one, without naming any of those studies. The Reichheld and Sasser article, often cited alongside it, only gives the 51 dollars from the credit card example and calculates no ratio at all. The ratio also depends on what each company puts into the cost of retention, such as CS, support and renewal discounts. The useful question for the board is how much it costs to win a customer, in how many months that customer pays the cost back and how much margin it leaves while it stays, calculations I detail in RevOps metrics.

What does academic research show about retention and customer value?

Peer-reviewed research confirms Reichheld and Sasser’s intuition and corrects the exaggeration: retention weighs more on customer value than margin and acquisition cost, and the size of the effect depends on the company.

Retention is the strongest lever of customer value

Sunil Gupta, Donald Lehmann and Jennifer Ames Stuart calculated the value of the customer base of five companies, Capital One, Amazon, Ameritrade, eBay and E*Trade, with public data from March 2002, and the result closely tracked the market value of three of them (Gupta, Lehmann and Stuart, 2004). Improving retention by 1% raised customer value by 3% to 7%, against about 1% for the same improvement in margin and 0.02% to 0.3% for a lower acquisition cost. The effect was larger in the companies that already retained more. The companies did not disclose their own retention, and the authors estimated it from analyst reports, an expert’s opinion and, for eBay, a market average.

Satisfaction, retention and market value

Gupta and Valarie Zeithaml’s review in Marketing Science brought together two decades of studies on customer metrics and financial results (Gupta and Zeithaml, 2006). In one of the reviewed studies, with 200 Fortune 500 companies across 40 industries, each 1% change in the American Customer Satisfaction Index was associated with a 1.016% change in Tobin’s q, a measure of market value, and only 14.5% of the difference in that relationship came from the industry.

The review also records criticism: one study argues that the 1990 results only hold with very low margins or already high retention, and another shows, in a retail bank, that the 5% of customers who left each year took 3% of deposits, while the 35% who reduced their balance without leaving took 24%. In B2B, that is the customer who cuts scope and stays on the active list, a loss I discuss in B2B churn.

A loyal customer is not always a profitable customer

Werner Reinartz and V. Kumar followed more than 16,000 customers of four companies for four years, among them a US high-tech service provider selling to businesses (Reinartz and Kumar, 2002). The correlation between relationship length and profit was weak to moderate in all four and stood at 0.30 at the B2B provider. There, long-standing customers paid 5% to 7% less than new ones and cost more to serve, because high-volume corporate customers know their own weight and use it to get discounts and dedicated service. The finding contradicts two of the 1990 explanations, the higher price and the lower cost.

How much to invest in retention

Reinartz, Jacquelyn Thomas and Kumar used 1998 to 2001 data from a multinational hardware and software manufacturer that sells to businesses, with 23,476 prospects contacted and 5,725 customers acquired, to calculate the budget split that maximizes profit per customer (Reinartz, Thomas and Kumar, 2005). In the working paper version of the study, the optimal split was 78.9% for retention and 21.1% for acquisition, close to what the company already did, and the problem was the total: it spent too much on its customers. The authors concluded that spending below the optimum hurts more than spending above it and that getting the retention budget wrong weighs more than getting the acquisition budget wrong.

Poorly targeted retention actions increase churn

In a field experiment at a wireless carrier, the company recommended a plan better suited to usage to part of its customers, and 10% of those who received the recommendation churned in the following three months, against 6% in the control group (Ascarza, Iyengar and Schleicher, 2016). For the authors, the recommendation broke inertia and led customers to reassess their contract. In another paper, with two field experiments, Eva Ascarza showed that the customers at highest risk of churning do not always respond best to retention actions, and that choosing targets by expected response works better (Ascarza, 2018).

What does this mean for customer retention in B2B?

The first consequence is in the definition. The 1990 article has a B2B example that applies to your company: a railroad almost never loses a customer entirely, but a customer who shifts 80% of its shipments to trucks cannot count as retained. In a B2B contract, the retention that matters is retention of revenue and margin, and counting customers hides scope reductions, discounts given to renew and active accounts that do not use what they bought.

The second is in the size of the effect, which depends on each company’s margin and acquisition cost. In the illustrative example below, with made-up numbers and no financial discounting, a customer cost R$ 40 thousand to acquire. With 15% annual defections it stays an average of 6.7 years, and with 10% it stays 10 years.

Illustrative example	Annual margin of R$ 30 thousand	Annual margin of R$ 10 thousand
Profit per customer at 15% annual defections, net of CAC	R$ 160 thousand	R$ 26.7 thousand
Profit per customer at 10% annual defections, net of CAC	R$ 260 thousand	R$ 60 thousand
Increase in profit per customer	62.5%	125%

The same five-point improvement yields 62.5% in one company and 125% in the other. A company with thin margins and high CAC gains more, in percentage terms, from each point of retention, which matches the criticism recorded by Gupta and Zeithaml and explains why the 1990 examples vary so much.

The third is in what retention measures. It results from work that starts with selling to the right customer and continues through onboarding, account management and renewals handled ahead of time, a routine I describe in B2B post-sale. The studies by Reinartz and by Ascarza point to what the CRM needs to store: margin and cost to serve per account, and each retention action with the customer’s response. Account portfolio management depends on these fields, which almost never exist in the databases I diagnose.

Summary table of the research cited

Source	Year	Sample and method	Main finding
Reichheld and Sasser, HBR	1990	More than 100 service companies, no published method	Reducing defections by 5% generated 30% to 85% more profit in the examples
Hart, Heskett and Sasser, HBR	1990	Management article	“Five times more” to replace a customer, as expert opinion
Reichheld and Schefter, HBR	2000	Bain analysis of e-commerce sectors	First located appearance of 25% to 95%
Gallo, HBR	2014	Popular article	Repeats 25% to 95% and cites “5 to 25 times” with no source
Gupta, Lehmann and Stuart, JMR	2004	Customer value model for five companies	1% more retention raises customer value by 3% to 7%
Reinartz and Kumar, HBR	2002	16,000 customers of four companies	Weak to moderate link between customer tenure and profit
Reinartz, Thomas and Kumar, JM	2005	5,725 customers of a technology manufacturer	There is an optimal retention budget; spending below it hurts more
Gupta and Zeithaml, Marketing Science	2006	Review of two decades of studies	Satisfaction linked to market value, with variation across companies
Ascarza, Iyengar and Schleicher, JMR	2016	Field experiment in wireless	Proactive action raised churn from 6% to 10%
Ascarza, JMR	2018	Two field experiments	Highest-risk customers are not always the best target

What these data do not say

The 1990 number does not say how much a Brazilian B2B company’s profit rises with retention. The examples were American service companies analyzed before 1990, unnamed, with no sample or margin of error, and the profit calculated was an estimate of customer value, with assumptions about margin, financial discounting and average life, far from the next year’s accounting result. The 95% does not even have that backing.

The academic studies have limits too. Gupta, Lehmann and Stuart work with five consumer companies and retention estimated by third parties. The studies by Reinartz and co-authors use databases from a few companies, all outside Brazil, and I cite the working paper version of the budget study, whose numbers may differ slightly from the published one. Ascarza’s experiments were run with consumers, and the effect of a proactive action on a B2B contract negotiated by a buying committee may be different. Much of the evidence shows association, which does not prove causation, and no study I found measures how many times more it costs to acquire than to retain.

What leadership does with these numbers

1. Write the definition of retention. Record in the team’s guide what counts as a retained customer, how scope reduction shows up and which time window applies, with customer and revenue retention side by side.
2. Measure actual defections over the last two years. Pull customer and revenue losses from the CRM and finance by segment and by cohort, with no estimates made up in meetings.
3. Record margin and cost to serve per account. Without these fields, the company does not know which long-standing customer is profitable.
4. Calculate what each point of retention is worth. Run the illustrative example with your own numbers before discussing the CS budget.
5. Record each retention action and the response. Discounts, value reviews and executive visits go into the CRM with date, account and result.
6. Review the installed-base budget once a year. Compare spending on CS, support and renewal discounts with the margin the base delivered.

Customer retention and L&D: what HR needs to assess

The request usually reaches HR as “train CS to retain customers”. The team needs to calculate revenue retention for its own portfolio, choose the accounts that receive an action by expected response and record the action and the result in the CRM. CS, onboarding and sales leadership take part, because the customer outside the ideal profile shows up months later in retention. To measure, take revenue retention, the share of accounts with recorded margin and the share of actions with a result in the CRM before the program, and compare after one renewal cycle. The data problems that get in the way of this reading are in CRM data quality.

If your board wants to decide the investment in the installed base with the company’s own numbers, that is the work of the in-company CRM adoption training. I run the program with CS, sales and leadership on the CRM and the portfolio you already have, and together we build the definition of retention and the margin and action fields. The company leaves with the CRM best practices guide and the portfolio and customer base management manual, written for your operation.

Frequently asked questions

Is it true that increasing retention by 5% raises profit by 95%?

Reichheld and Sasser’s original 1990 article said 25% to 85% and gave three examples of American service companies, with no published method. The 25% to 95% range appears in 2000, in an article about e-commerce, with no explanation for the new ceiling. In the example the 1990 article details, “5%” means five points less in the defection rate.

Does acquiring a customer really cost 5 times more than retaining one?

I found no study with a sample that supports this ratio. “Five times” appears in a 1990 HBR article as expert opinion, and “5 to 25 times” in a 2014 piece that cites no source. The board decides better with CAC, payback and margin per customer calculated on its own base.

What does academic research say about retention and profit?

Studies with a published method indicate that retention is the strongest lever of customer value, with an effect of 3% to 7% on value for each 1% improvement in the Gupta, Lehmann and Stuart study. They also show that relationship length and profit have a weak to moderate correlation and that overspending on the base also reduces profit.

Is a long-standing customer always more profitable in B2B?

The research says not always. At the B2B service provider studied by Reinartz and Kumar, long-standing customers paid 5% to 7% less than new ones and cost more to serve. A company only finds out which old accounts are profitable when it records margin and cost to serve per account in the CRM.

How do I calculate what a point of retention is worth at my company?

Use the annual defection rate measured in the CRM to estimate the customer’s average life, multiply it by annual gross margin per account and subtract CAC. Redo the calculation with five points less defection and compare. The difference shows how much it makes sense to invest in post-sale to gain those points.

How do I prepare the CS team to improve retention?

Train the team to calculate revenue retention for its own portfolio, choose accounts by expected response to an action and record each action and result in the CRM. Train sales leadership alongside them, because part of the loss starts in the sale, and measure the effect after one renewal cycle.