CMO playbook & advanced tactics for loyalty & retention

Every star, point and voucher your programme has ever issued sits somewhere on the balance sheet as money you owe. Starbucks carries well over a billion dollars in stored value card and loyalty deferred revenuedeferred revenueCash a company has collected for goods or services it has not yet delivered. It sits on the balance sheet as a liability until earned.View full definition → at any given point in its financial year: cash already collected, revenue not yet earned. Under IFRSIFRSThe global accounting rulebook that governs how companies report financial results, used across the EU and 140+ jurisdictions.View full definition → 15 and ASC 606, points handed out with a sale are a separate performance obligation, so a slice of today's revenue is pushed into a future period and the CFO signs off on the estimate. That is the moment loyalty stops being a marketing initiative and becomes a P&L negotiation: how much margin the scheme gives away, how much of the resulting behaviour is incremental, and who gets to decide when marketing and finance want opposite things.

CORE CONCEPT: THE THREE NUMBERS A CMO HAS TO BE ABLE TO SAY OUT LOUD

Assuming the mechanics the foundations lesson sets out, a CMO running a scheme at national scale should be able to state three figures without looking them up. First, the gross cost of rewards issued as a percentage of sales through the programme. Second, the incremental gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.View full definition → the scheme generates, measured against a holdout, not against non-members. Third, the outstanding liability and the breakage assumption sitting underneath it.

Most marketing leaders can quote enrolment and engagement and none of the three. That is why the programme loses the argument in the budget round. The finance team arrives with a liability number and a cost line; marketing arrives with member counts. One of those is auditable.

KEY SUB-CONCEPT 1: THE LIABILITY GROWS FASTER THAN YOU THINK, AND DEVALUING IT IS EXPENSIVE

Breakage, the share of points that will never be redeemed, is an accounting estimate you revisit. Set it too high and you book income you later have to give back. Set it too low and you starve the P&L of earned revenue while the balance sheet swells.

The trap is on the growth curve. A scheme adding members quickly issues points faster than the base redeems them, so the liability compounds while margin looks fine. When enrolment flattens, redemption rates catch up and the drag lands in a quarter nobody forecast. Airlines have lived this for decades; retailers meet it the first time growth slows.

The standard escape is devaluation: raise the points price of rewards, shorten expiry, narrow the catalogue. It works on paper. In February 2023 Starbucks raised star thresholds on several redemptions, doubling the cost of a brewed coffee, and took a visible round of member anger for it. The second-order effect is the one to plan for: members who bank points for a large reward stop banking once they learn the currency can be cut, and the behaviour you were paying for (deferred gratification, higher visit frequency between redemptions) is exactly what goes first. If you devalue, do it once, announce it early, and give the top spending decile something they keep.

Key sub-concept 2: cannibalisation is the default outcome, not the exception

Run the arithmetic before the pitch. A grocer with a 4 percent operating margin that pushes 70 percent of sales through the card and gives back 2 percent in value has committed roughly 1.4 percent of total sales, more than a third of operating profit, to the scheme. For that to pay, the incremental margin has to clear 1.4 points of sales. It usually does not clear it on its own; it clears it in combination with the data, the media value of owned channelsowned channelsMedia channels a company owns and controls directly, such as its website, blog, newsletter, social accounts and mobile app. No per-use payment to a publisher is required.View full definition →, and the supplier funding the data unlocks.

The measurement failure is almost always the same. Teams compare members to non-members, or redeemers to non-redeemers, and report a spend gap of two or three times. That gap is selection: heavy buyers join and redeem. The only honest read is a randomised holdout, a few percent of members who get no offers for a defined period, with the cohort read done the way the frameworks lesson lays out. Yes, you are deliberately under-marketing to real customers. The cost of that test is one or two basis points of sales. The cost of not running it is defending a seven or eight figure programme with correlations.

Watch for the reverse problem too: a scheme that is genuinely incremental at the margin can still be destroying value if it is funding purchases in low margin categories. Points on fuel and tobacco earned at the same rate as points on own-label groceries is a decision, whether or not anyone made it deliberately.

KEY SUB-CONCEPT 3: MEMBER PRICING SHIFTS THE COST, AND THE RISK, SOMEWHERE ELSE

Tesco has run Clubcard since 1995, and the shift to Clubcard Prices at the shelf changed the economics of the scheme rather than extending it. Instead of deferring value into points, the discount lands at the till and hits margin now. The upside is immediate: identification rates jump because the price on the shelf is only available to scanned members, and the liability stops compounding. Tesco has said the large majority of its sales now go through a Clubcard.

The exposures move with it. Non-members pay a visibly higher price for the same item, which is a fairness question the CMO owns publicly. After consumer group pressure, the CMA reviewed supermarket loyalty pricing and concluded that the savings on offer are generally genuine, but the review itself is the point: member pricing at scale is a pricing policy, and pricing policy gets regulated. Two-tier shelf pricing also puts the promotional calendar, supplier funding and the loyalty scheme into one negotiation. If marketing sets the member price and commercial sets the base price, you will find categories where the two teams have quietly financed the same discount twice.

Key sub-concept 4: who owns it when marketing and finance disagree

Four functions have a legitimate claim: marketing (engagement and the customer promise), finance (liability, breakage, revenue recognition), commercial or merchandising (price and margin), and data or legal (consent, retention periods, what the programme is allowed to do with what it knows). Programmes fail organisationally when all four have a veto and none has the P&L.

The workable settlement names one P&L owner and writes down the arbitration rules in advance: a ceiling on reward cost as a percentage of programme sales, a minimum incrementalityincrementalityThe share of results (sales, conversions, revenue) that only happened because of a marketing action, not what would have occurred anyway.View full definition → hurdle tested at a fixed cadence, a notice period before any change to earn or expiry rules, and a rule that breakage released to income cannot be counted as marketing performance. That last one matters more than it sounds. A CFO who shortens expiry gets a one-off income release this year and a colder base next year, and the engagement drop shows up in the marketing scorecard two quarters after the decision was taken elsewhere.

Paid membership is the structural fix, because the member pays and the argument gets simpler.

Real-world cases with results

STARBUCKS REWARDS: around 34 million active US members, generating roughly 57 percent of US company-operated revenue. The scheme is also a treasury operation: preloaded cards and app balances mean customers fund Starbucks in advance, and unused balances eventually flow through as breakage income. Bonus Star challenges tied to specific behaviours (a new cold brew, a Tuesday visit) raise frequency without cutting the ticket price, which is the difference between paying for behaviour and paying for volume you already had. The 2023 threshold increase is the cost side of the same ledger: the liability was managed, and the goodwill bill arrived immediately.

AMAZON PRIME: launched in 2005 at $79 as a shipping offer, now $139 a year in the US with over 200 million members worldwide. Amazon reports subscription services net sales of roughly $40 billion a year, and Prime is the bulk of it. The organisational lesson is the accounting one: the fee is revenue, the benefits (shipping, video, music) are costs carried by the units that deliver them, so nobody can claim the programme is free. Each added benefit raises the perceived cost of cancelling, and the annual price rises (to $119 in 2018, $139 in 2022) test how much of that perception is real.

TESCO CLUBCARD: three decades of the same dataset, and the reason Tesco could price, range and site stores against actual basket behaviour rather than survey panels. dunnhumby's analysis of Clubcard data became a business Tesco owned outright, which is the clearest illustration of the point: the durable asset was the data and the supplier negotiation it supported, not the points.

CMO action items

• Get the liability balance and the breakage assumption from finance this quarter, and ask what happens to reported margin if redemption rises five points. If nobody has modelled it, you have found your next board paper.
• Stand up a permanent randomised holdout of members, sized to detect a one point difference in spend. Report incremental margin per member alongside enrolment, and stop quoting member versus non-member spend gaps.
• Price the reward by category margin, not by a flat percentage of spend, and check whether the tiering the application lesson works through is funded by the categories it actually drives.
• Write the arbitration rules with your CFO before the next cost-cutting cycle: who can change earn rates, with what notice, and how released breakage is reported.

Common mistakes that kill results

MISTAKE 1: NOT KNOWING THE LIABILITY. A marketing leader who cannot state the outstanding balance and the breakage rate will not be in the room when the terms are changed. The scheme then gets cut by people optimising a single line of the accounts.

MISTAKE 2: PROVING VALUE WITH SELECTION BIAS. Members spend more because heavy buyers join. Redeemers spend more because spending is what earns the reward. Neither comparison survives contact with a finance team that has seen the technique before, and using it once costs you the argument for years.

MISTAKE 3: DEVALUING QUIETLY DURING A BAD QUARTER. Shortening expiry or raising thresholds without notice releases income now and teaches your best customers that the currency is not safe. Banking behaviour stops, redemption spikes, and the saving reverses inside a year with the brand damage left over.