What a repeat customer is really worth

Two firms in Varna sell the same service to the same people. One of them can pay 180 leva to win a customer and still run a healthy business. The other starts losing money somewhere around 60 and has to stop bidding. Every time they meet in the same auction, the second one loses the click and assumes the first one is being reckless.

Nobody in that story is cleverer in the ad account. The whole difference was decided after the first sale, in delivery, scheduling and follow-up, months before either of them opened Google Ads. Retention is the quiet variable that sets how much you are allowed to spend, and almost no small business measures it.

The first sale is the wrong unit of analysis

If you judge a channel on the profit from the first order, you will shut down channels that work and protect channels that do not. A customer who buys once at a fat margin looks better on that measure than a customer who buys four times a year at a thin one. The second customer is worth several times more and your spreadsheet says the opposite.

This bites hardest in anything auction-based. Whoever can afford the highest cost per customer, and still profit, takes the impression. Retention raises that ceiling. It is a marketing advantage manufactured almost entirely by operations.

Four numbers, from your own records

You do not need a model. You need four figures and an afternoon.

• Average gross profit per order. Revenue minus the direct cost of delivering it. Not revenue. Revenue will flatter you by a factor of three or four and the mistake is invisible until it is expensive.
• Average orders per customer per year. From transaction history, not memory. Memory rounds up.
• How many years a customer stays. If your data is thin, use the share of customers still buying twelve months later as a stand-in.
• Multiply the first three. That is your rough lifetime gross profit, and the share of it you are prepared to hand to acquisition is now a decision you can actually argue about.

Be mean with the horizon. One or two years, not ten. A model that needs a customer to stay a decade before the spend makes sense is not a plan, it is a wish with arithmetic attached. And money arriving in 2029 is worth less than money arriving this month, which matters a great deal if you are funding acquisition out of cash flow rather than capital.

What the number buys you

Three arguments stop being arguments.

The first is your bid ceiling. If lifetime gross profit is several times the first order, you can spend more than the first order returns and still be building something profitable, so long as you can fund the gap in between. That gap is where businesses die, so size it before you widen it.

The second is which customers you want. Segments differ enormously on repeat purchase, and the segment with the cheapest acquisition cost is very often the worst on coming back. Once you can see both columns, targeting stops being a conversation about volume.

The third is whether a discount is defensible. A first-order discount that buys a customer who stays for years is an investment. The same discount aimed at a segment that never returns is a loss you have paid for twice. Same tactic, opposite verdict, and the only thing that separates them is a number you probably have not calculated. The same logic runs in reverse on price, which we have written about in raising prices.

The second order is the hinge

RJMetrics looked at 176 online retailers and 18 million customers for its 2015 buyer behaviour report and found that just 32 per cent of customers bought again within a year of their first purchase. The same study put the top 1 per cent of customers at eighteen times the value of the average one.

Those are e-commerce figures and your business is probably not e-commerce, so take the shape rather than the digits. The shape is that the gap between the first and second purchase is where most customers are lost, and that the odds improve with every order after it. Which means retention work should pile up almost entirely on one transition rather than spreading politely across the customer lifetime.

Things that move it, in rough order of how often they work:

• Deliver the first order noticeably well. Overdelivering once, at the start, does more than a year of newsletters.
• Make contact before they need you again. Work out your natural repurchase interval and reach out just inside it, with something useful rather than an offer.
• Strip out the friction of ordering again. Saved details, a named contact, a standing arrangement.
• Ring ten customers who did not come back and ask why. Ten phone calls will teach you more than any dashboard, and the answer is usually small, dull and fixable.
• Give them a structure to sit inside. A maintenance plan, a retainer, a scheduled review. Structure outlasts goodwill.

There is a fish restaurant on the Varna seafront that has bought from the same boat for eleven years. No contract, no app, no loyalty scheme. The skipper rings at six, the kitchen says yes or no, and that is the entire relationship. The boat has never advertised and has a waiting list.

Two numbers you will be handed, and should refuse

The first is that it costs five times more to acquire a customer than to keep one. Chase it back and it dissolves. The trail runs to work by the Technical Assistance Research Project in Washington in the late 1980s, picked up by Christopher Hart, James Heskett and Earl Sasser in their 1990 Harvard Business Review piece “The Profitable Art of Service Recovery”, and then repeated for thirty years by people quoting each other. Keiningham, Vavra, Aksoy and Wallard take it apart in Loyalty Myths (Wiley, 2005): the ratio assumes existing customers spend more, pay full price and cost less to serve, and none of those hold reliably. Your own ratio is whatever your own records say, and it is worth half an hour to find out.

The second is Reichheld and Sasser’s claim that retaining 5 per cent more customers nearly doubles profit, from “Zero Defections”, Harvard Business Review, September 1990. It is the most quoted sentence in customer loyalty and it came from a worked example about a credit card company, not from a measured result across businesses. The underlying point is sound: churn is expensive and compounding. The specific figure is arithmetic someone did once about somebody else’s industry.

Most of this is not a marketing job

This is the part owners resist, so we will put it bluntly. Churn is usually caused by delivery, response time, staff turnover, or a product that quietly stopped being competitive. Email campaigns fix none of that. They remind unhappy customers that you exist.

So before you spend on retention marketing, go and find the operational cause. How long does an existing customer wait for a reply. How many complaints turn into fixes. How often does the person your customer actually knows get replaced. In service businesses, continuity of one human relationship is frequently the single biggest retention factor, and it has never once appeared in a marketing report.

Pull twelve months of transactions and answer one question: what share of your customers bought more than once. Then compare gross profit from repeat customers against gross profit from first-timers. If the repeat column is large and nothing in your business is deliberately built to produce it, you are getting that result by accident. Accidents are not a strategy, but they are evidence that the demand is there. If you want a second pair of eyes on the arithmetic before it turns into a bidding decision, that is what our consultations are for, and the true cost of acquiring a customer is the other half of the same sum.

Fix the first-to-second purchase before you spend another lev on acquisition. Everything downstream of it is cheaper than it looks, and everything upstream of it is more expensive.