CMO playbook & advanced tactics: pricing strategy that drives revenue

Three pricing decisions reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → a CMO in a form no test settles. Who is allowed to discount, and by how much. Whether you reprice people who already bought. Whether you commit in public to a ceiling on your own margin. At 200 customers all three are reversible. At two million they are not: discount authority pushed too far down the org becomes a permanent leak no campaign outruns, a repricing of the installed base can turn your best-referenced buyers into your loudest critics inside a weekend, and a published margin cap buys a brand asset by selling an option you cannot buy back. McKinsey's much-cited work on the S&P 1500 puts a 1% improvement in realised price at roughly 8% more operating profit, which is why these three calls deserve more leadership time than most campaign budgets get.

Core concept: the price you set is not the price you collect

Between the list price and the money that lands there are invoice discounts, promotional credits, extended payment terms, rebates, returns and free months. Finance sees the aggregate. Sales sees individual deals. Marketing sees the pricing page. Nobody owns the gap, and the gap is where the margin went.

The arithmetic is unforgiving. On a product with 70% cost of goods, letting the average discount drift from 12% to 18% removes a third of gross profit at identical volume. Six points is invisible inside any single negotiation and fatal in aggregate. The CMO contribution is not the discount policy itself, which finance will write. It is supplying the reason a price holds: the argument a rep uses at the moment they would otherwise reach for a number.

Key sub-concept 1: discount governance is an org design question

A workable regime has four parts: discount bands with a named approver at each level, a rule that every concession is traded for something (longer term, prepayment, reduced scope, a reference), an expiry date written into every promotional price, and a monthly report of realised price by segment rather than by deal.

Tesla shows how the leak can move upstairs even when reps have no discretion at all. Fixed, no-haggle pricing removed the individual negotiation entirely, then quarter-end incentives put it back at company level: free Supercharging transfers, paint and insurance credits, discounts on inventory cars. Buyers learned to wait for the final fortnight of a quarter. The metric worth watching is the share of volume closing in the last two weeks of a period. If it climbs year on year, your customers have priced your reporting calendar into their behaviour.

Key sub-concept 2: repricing an installed base

New buyers read a lower price as an offer. Existing buyers read it as a refund they did not get. That asymmetry is the whole problem.

In January 2023 Tesla cut US prices by up to roughly a fifth on some Model Y configurations, days after cutting in China. Owners who had taken delivery weeks earlier gathered at Chinese showrooms demanding compensation. None came. The second-order effect ran through residual values: lease quotes and fleet economics key off what the car is worth in three years, and Hertz began offloading around 20,000 EVs from its fleet in early 2024, citing depreciation and repair costs. The same pattern hit software buyers when Full Self-Driving went from $15,000 in 2022 to $8,000 in April 2024, with the subscription halved to $99 a month. Early purchasers held a non-transferable asset worth half what they paid for it.

The leader's choices are notice period, bounded grandfathering with an end date in writing (12 to 24 months is the usual range), or an explicit make-good. Open-ended grandfathering looks generous and costs more than it appears: you support a legacy price book forever, and every future packaging change has to be built twice.

Key sub-concept 3: capping margin in public

In 2018 Xiaomi's board resolved that net margin on hardware would never exceed 5%, with anything above returned to users. It worked as positioningpositioningThe mental space you want your brand to occupy in your target customer's mind relative to alternatives.View full definition → because it was not binding. Xiaomi's hardware net margin has run well under that ceiling, so the promise cost nothing in cash while buying years of trust. A company already earning 12% net on hardware cannot copy the move.

What it costs: you surrender the option to raise price when component costs spike, and you push the profit into monetising the device afterwards, which is why ads inside the phone software became a recurring complaint. Note also that the cap binds in one direction only. Xiaomi's SU7 launched in March 2024 at 215,900 yuan, below build cost by Lei Jun's own account. A ceiling on profit says nothing about how far below cost you may go to buy a market.

Before publishing any such promise, ask three things: can an outsider audit it, does it bind you in a downturn, and what is the plan the day you have to break it.

Key sub-concept 4: raise the ceiling instead of cutting the floor

Duolingo's answer to price pressure has been to add points above, not below. The free tier does the acquisition work the foundations lesson describes, a high single-digit share of monthly actives converts to paying, the family plan captures households, and Max sits above Super for learners who want the generative features. The floor never moved.

Two failure modes come with the ladder. Every tier consumes pricing-page space, support scripts and enablement, so the fifth tier usually earns less than it costs to explain. And a premium tier built on generative AI has a cost line moving underneath it, which means 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 → per subscriber is a variable, not a constant. Price the top tier as if inference costs could double.

Real-world cases with results

Case 1: tesla buys volume and reprices the brand

The 2023 cuts did what they were meant to do on units: deliveries reached about 1.81 million, up roughly 38%, while revenue rose around 19%. Operating margin went the other way, from 16.8% in 2022 to 9.2% in 2023. The deeper cost was semantic. A cut repeated four or five times stops reading as a promotion and starts reading as your position. Tesla moved from scarce premium to mass market without ever announcing it, and every used-car listing did the announcing instead.

Case 2: xiaomi's 5% pledge as a marketing asset

The cap made hardware a distribution channel rather than a profit centre. Internet services (advertising, app distribution, financial products) contribute roughly a tenth of Xiaomi's revenue but a disproportionate share of its gross profit, at margins several times those of phones. The pledge is credible because the business does not need hardware margin, and it holds a competitor with a costlier hardware model at arm's length: matching the promise would mean rebuilding the P&L, not rewriting a press release.

Case 3: duolingo grows without discounting

Revenue passed $531 million in 2023, up more than 40%, with daily actives growing over 60% in the same year. No headline price cut carried that. The growth came from more people entering the free tier, better conversion inside it, and a higher-priced option for the people already paying. The CMO implication: when acquisition is working, a discount is the most expensive way to buy growth you were getting anyway.

CMO action items

• Pull the last two quarters of closed deals and plot realised price against list by segment and by approver. If one region or one manager sits five points below the rest, that is a governance problem, not a market problem.
• Write the installed-base repricing protocol before you need it: notice period, who is grandfathered, for how long, and what the make-good is. Get finance and legal to sign it while nothing is on fire.
• Stress-test any public price or margin promise against a bad quarter. Anything you would break under pressure should not be said out loud.
• Before approving a price cut, cost the alternative: a tier above the current top, priced for the 5% of customers who already ask for more.

Common mistakes that kill results

• DELEGATING DISCOUNT AUTHORITY WITHOUT A LEDGER: Approval thresholds without reporting produce compliance on paper and drift in practice. Every concession should carry a reason code and appear in a monthly realised-price view. What is not counted becomes standard within two quarters.
• LETTING NEW LOGOS BUY BETTER THAN RENEWALS: When the acquisition promotion is cheaper than the renewal price, your own customers find out, usually in a community forum. Either match the base or give it something the new cohort does not get.
• CUTTING PRICE WITHOUT MOVING POSITION: A reduction with no repositioning story reads as weakness. If you are lowering price, say what changed (scale, cost curve, strategy) and change the marketing to match, or the market writes its own explanation.
• TREATING A PUBLIC COMMITMENT AS A CAMPAIGN: A margin cap, a price-lock guarantee or a no-discount policy is an operating constraint that outlives the CMO who announced it. Say it only if the finance model still works when demand falls 20%.