Real-world application: the CMO as business leader

In March 2020 Airbnb's bookings collapsed and Brian Chesky's first large cost decision was to stop marketing outright. Roughly $800 million of planned spend came out that year, most of it performance. The decision that mattered came later: when demand returned, Airbnb did not switch the machine back on. Five years on the company still has no CMO, marketing reports to the founder, and the sales and marketing line sits at close to half the share of revenue it took in 2019. This lesson stays with that one company and follows the call through to traffic, margin, supply, and the two or three places where it nearly did not hold.

Core concept: the 2020 rebalance, in numbers

Airbnb's marketing already sat inside the commercial remit the foundations lesson defines, which is why the cut was argued as a capital allocation question and not a budget trim. The baseline was 2019: revenue of about $4.8 billion, sales and marketing of roughly $1.6 billion (near a third of revenue), and a net loss around $675 million. A large share of that spend was paid search, metasearch and bidding on terms for a brand travellers already knew by name.

What changed after March 2020:

• Performance spend fell to a test budget, concentrated in countries where Airbnb was not yet the default word for the category
• Brand work was judged on unpaid traffic share and direct app usage rather than last-click return
• Product marketing was folded into product management, so the roadmap and the message ship together
• The marketing number is now argued in front of the CEO against free cash flowfree cash flowFree Cash Flow is the cash a company generates from operations after funding the capital expenditures needed to maintain and grow its asset base.View full definition →, which raises the bar and also removes the departmental advocate who used to defend it

Airbnb has reported that roughly 90% of its traffic arrives direct or through unpaid channels. That single figure is the entire argument. A company with a 30% unpaid share running the same play is not saving money, it is opening a demand hole it will have to buy back at a worse price.

Sub-concept 1: the number airbnb decided to own

Marketing at Airbnb is held to two figures a CFO recognises: share of unpaid traffic, and free cash flow margin. The second is the harder one. Airbnb closed 2022 with about $3.4 billion of free cash flow on $8.4 billion of revenue, and 2023 with roughly $3.8 billion on $9.9 billion, a margin near 40% in both years. For a marketplace that owns no inventory, that margin only survives if demand is not re-bought every quarter.

Once marketing accepts that number, the internal arguments change shape. A campaign competes with buybacks and with host incentives, and it has to be sized in cash rather than 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 →. The trap is picking a number you cannot actually move: unpaid traffic share is partly a function of category dominance, so a challenger that adopts it as a marketing KPI will spend two years reporting a metric that its brand position, not its work, determines.

Sub-concept 2: where the model does not travel

Brand-led acquisition works where the brand is already a verb. In the countries Airbnb calls expansion markets, among them Brazil, Japan, Korea, Germany and India, it was a challenger against local incumbents and hotel metasearch, and there it kept buying demand. Airbnb has consistently reported those markets growing faster than its core, and it has funded them accordingly.

So the company runs two budgets on two logics inside one function: a brand-led core where paid spend is nearly optional, and a set of markets where paid is the only way in. The failure mode is letting the global story flatten the country plans. A regional lead who hears the founder say Airbnb does not do performance marketing, and quietly cuts a Sao Paulo or Seoul programme that was still the cheapest source of first-time bookers, is executing the headline instead of the strategy.

Sub-concept 3: brand spend pointed at supply, not demand

The least copied part of the rebalance is where the money went. Airbnb aimed a large share of brand investment at hosts rather than guests: "Made possible by Hosts" in 2021, then the "Airbnb it" campaign pushing ordinary homeowners to list. Active listings passed 8 million by 2024, and supply, not guest demand, has been the binding constraint on nights in most quarters.

The economics differ from a booking ad in one important respect. A guest campaign buys a transaction. A host campaign creates an asset that produces nights for years and needs no re-acquisition, which is why supply marketing is the line most defensible in a downturn and the one most CMOs never propose, because its payback lands two budget cycles after the person who approved it has moved on.

Sub-concept 4: the cadence that makes a brand-led model pay

Airbnb runs the review rhythm and decision rights the frameworks lesson sets out, with the founder in the chair. The mechanism worth copying is the release calendar: two product launches a year, each carrying the marketing plan. Categories and AirCover in May 2022, Rooms in 2023, Icons in 2024, Services and a rebuilt Experiences in May 2025.

The second-order effect cuts both ways. Two releases a year give the company two earned-media moments that a performance-led competitor has to pay for continuously. But if a release slips, the quarter has nothing to say, and press coverage cannot be bought back at short notice the way a search auction can be re-entered on a Monday morning.

Real-world cases with results

CASE 1: 2020 to 2021, the cut and what came back

Airbnb stopped marketing, cut staff, and went public in December 2020 at $68 a share, closing the first day near double that. Traffic returned without the paid layer: management described switching spend off market by market and watching volumes hold. By Q3 2021 revenue was above $2.2 billion, ahead of the same quarter in 2019, with net income of roughly $834 million and the first quarter of adjusted EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition → above $1 billion. Full-year 2021 revenue of about $6 billion beat 2019 on a marketing budget well below it.

CASE 2: 2022 to 2024, the model at steady state

2022 was Airbnb's first full year of GAAPGAAPThe standard set of accounting rules companies follow to prepare consistent, comparable financial statements, dominant in US reporting.View full definition → profitability, near $1.9 billion of net income on $8.4 billion of revenue, with sales and marketing around 18% of revenue against 34% in 2019. Read 2023 carefully: the reported $4.8 billion of net income included a one-off release of a tax valuation allowance worth roughly $2.9 billion, so the operating picture looked much like 2022. By 2024 revenue reached about $11.1 billion, growth had slowed to the low teens, marketing dollars were rising again in absolute terms, and the company guided to $200 to $250 million of investment in launching new businesses in 2025. The ratio fell; the spend did not disappear.

CASE 3: BOOKING.COM, the counter-example that pays for demand and wins anyway

Booking Holdings, whose largest brand is Booking.com, does the opposite and out-earns Airbnb doing it. In 2023 it spent roughly $6.8 billion on marketing against $21.4 billion of revenue, near a third, sold more than a billion room nights against Airbnb's roughly 448 million nights and experiences, and reported net income around $4.3 billion with an operating margin in the high twenties, above Airbnb's mid-teens. The expense lines are not perfectly comparable, but the gap is not an accounting artefact.

The reason is structural: hotel supply is largely undifferentiated, so intent is bought at the moment of search, and Booking's repeat rate through its app absorbs the acquisition costacquisition costCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → over multiple trips. Cutting paid there would hand the auction to a competitor selling the same room. Airbnb's homes are unique, its brand carries the category name, and its guests arrive already deciding. The correct lesson is not that performance marketing is a tax. It is that the mix follows brand position, supply differentiation and repeat behaviour, and a CMO who imports Airbnb's answer without checking those three conditions is copying a conclusion rather than a method.

CMO action items

• Calculate your unpaid traffic share before you propose any reallocation. Under about half, an Airbnb-style cut removes volume you cannot replace inside a quarter.
• Run a geogeoThe practice of making your brand and content visible and citable inside AI-generated answers from tools like ChatGPT, Gemini and Perplexity.View full definition → holdout: switch paid off in two or three comparable markets for six to eight weeks and measure total sessions and bookings, not attributed ones. This is the only evidence that survives a CFO conversation.
• Price one supply-side or ecosystem campaign against a demand campaign of the same cost, over a three-year horizon. If your business has a supply constraint, the arithmetic will surprise the room.
• Split your plan into brand-led markets and challenger markets, with different budgets and different metrics, and say so out loud so nobody applies the global narrative to a country where it is wrong.

Common mistakes that kill results

Mistake 1: Copying the cut without the brand behind it

Airbnb had ten years of word of mouth, a name used as a verb, and roughly 90% unpaid traffic before it stopped spending. Those conditions were the precondition, not the result. A challenger that mimics the move usually discovers within two quarters that its paid channel was not inefficient, it was load-bearing.

Mistake 2: Treating the reallocation as a permanent cost saving

Marketing as a share of Airbnb revenue halved, but the dollars climbed back above $2 billion by 2024 as the company funded expansion markets and new lines of business. Boards that book the ratio improvement as a structural saving will resist the next increase, and the CMO who let them make that assumption owns the problem.

Mistake 3: Going dark on measurement at the exact moment you need it

Killing paid removes the attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → reporting most teams live on, and 2021's recovery coincided with a travel rebound the whole category enjoyed. Without holdouts, brand search tracking and a market-by-market counterfactual, you cannot tell brand-led growth from a tide lifting everything, and you will not be able to defend the budget when growth slows.

Key takeaways

• The cut worked because roughly 90% of Airbnb's traffic was already unpaid. Establish that number before you touch the paid budget.
• Marketing at Airbnb is judged against free cash flow, which held near 40% of revenue in 2022 and 2023. A campaign that cannot be sized in cash does not get funded.
• Brand money aimed at supply (the host campaigns behind 8 million-plus listings) buys an asset that keeps producing; brand money aimed at demand buys a quarter.
• Booking Holdings spends about a third of revenue on marketing and still earns a higher operating margin than Airbnb. Mix follows brand position and supply differentiation, not fashion.
• Read the reported profit properly: Airbnb's 2023 net income of $4.8 billion carried a roughly $2.9 billion one-off tax item, and a CMO who quotes it as marketing's result loses credibility the day someone checks.