Customer value, acquisition and retention

At a newspaper and an online grocer, deeper acquisition discounts produced customers who repurchased less and were worth less. From the abstract: "a 35% acquisition discount results in customers with about one-half the long-term value of non-promotionally acquired customers."

The deeper the discount you acquire customers with, the less they end up being worth.

At a web-hosting firm, "marketing-induced customers add more short-term value, but word-of-mouth customers add nearly twice as much long-term value."

The channel a cohort came in through predicts what it'll be worth.

At a digital-TV provider, behavioral differences "cause the average customer lifetime value of free-trial customers to be 59% lower than that of regular customers." Free-trial customers do respond more to marketing and to usage.

Judge your free trial on whether it pays back in lifetime value. Sign-ups alone won't tell you.

"Deeper price discounts in the current period increased future purchases by first-time customers … but reduced future purchases by established customers."

Discounts don't always bring in worse customers. It depends on who gets the discount.

Merchants' Yelp ratings fall after they run Groupon deals. Across "over 7 million Yelp reviews," the authors test explanations and propose that deal customers are "real, unbiased customers" compared with inflated baseline reviews.

Customers who come in through deals behave differently, and so do the signals they leave, like reviews.

The Pareto/NBD model estimates "the probability that any particular customer is still active" from the number and timing of past purchases.

How do you count live customers when there's no contract? This model does it.

The BG/NBD model is "vastly easier to implement" (it fits in Excel) and gives "very similar results."

This gives you a practical forecast of repeat purchases and lifetime value.

The shifted-beta-geometric model fits in a spreadsheet. In the paper's words, "the observed phenomenon of increasing retention rates is simply a sorting effect in a heterogeneous population."

A cohort's retention rises with age because the leavers leave first. Nobody got more loyal. So read churn by cohort instead of as one blended rate.

"at the cohort level, retention rates typically increase over time." Using a single retention rate "yields a downward-biased estimate" of the customer base's remaining value.

Value your subscription book with one blended churn rate and you'll get it wrong.

Life tables that ignore differences in frailty "overestimate current life expectancy … while underestimating rates of individual aging." The frail die first, which changes what the survivors' averages seem to say.

Fader and Hardie took this from demography and applied it to customers.

Across five public firms, "a 1% improvement in retention, margin, or acquisition cost improves firm value by 5%, 1%, and .1%, respectively."

Retention is the biggest lever you have on firm value. So the next question is whose retention.

"Companies can boost profits by almost 100% by retaining just 5% more of their customers." The case examples are 85% more profit at a bank's branch system, 50% at an insurance brokerage and 30% at an auto-service chain.

It's the most-quoted retention claim in business. The figures are case examples from the article, so don't treat them as a general law.

Across four companies and more than 16,000 customers over four years, the link between loyalty and profit was "much weaker—and subtler" than its proponents claimed. At one of the four, a high-tech corporate service provider, about half the customers who had bought regularly for at least two years "barely generated a profit."

Tenure doesn't tell you value, so put your retention effort where the value is.

"Customers who are selected on the basis of their lifetime value provide higher profits in future periods than do customers selected on the basis of several other customer-based metrics."

Rank your accounts on expected future value instead of past spend or count.

"customers identified as having the highest risk of churning are not necessarily the best targets." Targeting by sensitivity to the intervention was "significantly more effective."

Aim your retention effort at the customers it actually changes, weighted by value, instead of at the highest churn scores.

"whereas only 6% of customers in the control condition churned during the three months following the intervention, 10% did so in the treatment group."

Reach out to save customers and you can wake up "sleeping dogs" who would have stayed anyway.

Rank customers by the campaign's incremental effect on churn and on future cash flows, net of cost. "Two field experiments affirm that our approach leads to significantly more profitable campaigns than competing models."

Save your valuable customers instead of just chasing the ones most likely to leave.

Four groups:

• Persuadables: "made more likely to stay by treatment."
• Sure Things and Lost Causes: stay or leave "whether treated or not."
• Sleeping Dogs: "tend to be driven away by treatment."

Only spend where the effort changes the outcome. That goes for your ads as much as your retention, since brand search reaches sure things.

Targeting Boston Globe discounts on a surrogate for long-term retention beat targeting on short-term proxies. "Over three years, our approach had a net-positive revenue impact in the range of $4–$5 million."

Validate your early signals against long-term value, then use them to decide who to target.