I sat through a growth seminar last week, run by one of the big global marketing and data agencies, expecting the usual deck of platform updates and case studies. What I got instead sent me back to first principles. Nothing in it was new; most of it was older than digital marketing. That was exactly the point.
The session was built on a framework they called customer economics. Strip away the branding and it’s direct-marketing maths that has been sitting in textbooks since the 1970s, dressed for a modern audience. I liked it enough that I went home and built a calculator to model it properly, which you can play with here: customer economics calculator. This post is the thinking behind it.
The three levers, and the number you can’t change
Here’s the uncomfortable simplicity of it. Every pound of revenue you’ll make next year comes down to how many customers buy from you and how much each one spends. That’s it. Break the customer count in two - the people who come back, and the new or reactivated people who arrive - and you have three levers:
- Retention - the share of last year’s customers who buy again
- Acquisition - new and reactivated customers you bring in
- Spend per customer - what each one is worth on average
Sitting underneath all three is a fourth number that isn’t a lever at all: the base of customers you start the year with. That number is fixed. It’s the sum of everything you did last year, and you can’t touch it now. Worth sitting with for a second, because most planning conversations skip straight past it.
If any of this feels familiar, it should. It’s Jay Abraham’s “three ways to grow a business”, which he was teaching decades ago: get more customers, get them to spend more, get them to come back more often. The maths hasn’t changed. What’s changed is that we now have enough data to be honest about it, and mostly choose not to be.
The baseline nobody forecasts
Most forecasts I’ve seen are built bottom-up. List every campaign you plan to run, estimate the bookings or sales from each, add them together, call it next year’s number. It feels rigorous. It rarely survives contact with scale, and it takes credit for revenue that was coming anyway.
The more honest starting point is what the agency called the “as-is” scenario, and what forecasters have always called business-as-usual: what happens to the business next year if you change absolutely nothing. Same retention, same acquisition, same prices. The intuitive answer is “we stay flat”. The real answer is usually that we grow a little, because the retained base rolls forward and compounds - you start next year with more customers than you started this one.
That has a sharp edge to it. If your target for the year sits just above the as-is line, you’ve set the team up to celebrate standing still. You’ll hit the number, pat yourselves on the back, and never notice that the business would have got most of the way there on its own. First principles: know your baseline before you claim a win.
What you can actually afford to pay for a customer
This is the bit digital marketing half-forgot, and it’s the reason I built the calculator the way I did.
Long before anyone talked about ROAS, catalogue and direct-mail marketers worked to a number called the allowable acquisition cost. The logic was plain: work out what a customer is worth to you over their lifetime, and that tells you the most you can afford to pay to acquire one. Spend less than the allowable and you make money; spend more and you’re buying revenue at a loss, however good the campaign looks. Bob Stone and Jim Kobs formalised the sums; over here, Victor Ross at Reader’s Digest drilled it into a generation with the line “profit is a cost”, and Drayton Bird has been teaching the maths of the allowable for as long as I can remember. Dave Chaffey carried it into the digital frameworks most of us actually use.
It also changes what I ask for. I’ve more or less stopped asking a channel for its ROAS. The question I care about is whether we’re paying less to acquire a customer than that customer is worth over time - a different question, and a much less flattering one. The gap between the two is where a lot of budget bleeds out without anyone noticing.
The ROAS trap
Which brings me to the metric everyone reaches for and few interrogate. ROAS measures efficiency, not effectiveness, and the two get confused constantly.
A high ROAS often isn’t a sign you’re winning; it’s a sign you’re underspending. Push more budget into a channel and, past a point, your ROAS falls, because you’ve exhausted the easy demand and started paying to reach people who need more convincing. The really awkward part is the selection effect: a lot of what paid channels report is people who were always going to buy - brand-name searchers, retargeted browsers who’d already made up their minds. The channel takes the credit; the customer had decided anyway. Lionel Yeo’s write-up of the ROAS trap puts it well, and Tom Roach has been making the same argument about ROI for years.
The corrective is the one Les Binet and Peter Field landed on after chewing through a thousand IPA case studies: roughly 60% of budget on brand building, 40% on activation. Not an iron rule - it flexes by category and situation - but a useful counterweight to the pull of the efficiency metric. I treat ROAS as a hygiene check, not a number to grow the business by.
Retention is the cheap lever
The last lever is the one that gets the least airtime and does a lot of the real work. It is far cheaper to bring a lapsed customer back than to win a stranger. Fred Reichheld’s research at Bain found that a 5% lift in retention can raise profits by 25% and up in some sectors, and the tools have caught up - win-back flows in Klaviyo, reactivation scoring, a whole category of software now pointed at the base you already paid to acquire.
The catch: retention is a stubborn metric. Best-in-class experience and CRM might add a handful of percentage points, no more; you’re not going to swing it by twenty. But those few points compound through the whole base, year after year, which is exactly why they’re worth chasing even when they look unglamorous next to a shiny acquisition number.
The bottom line
None of this is clever. That’s what I liked about it. Three levers, one fixed base, a baseline you should know before you set a target, and a clear-eyed view of what a customer is worth before you decide what to pay for one. The frameworks that survive tend to be the boring ones.
If you want to see how the numbers move on your own business, the calculator lets you plug in your retention rate, your acquisition, and your average spend, and watch what happens to the top line. I built it mostly to stop myself fudging the baseline, which is the part I’m most tempted to fudge.
