A filing-led analysis of Airbnb’s disclosed marketing mix and a framework for testing paid reliance without confusing brand strength, product distribution and cost ratios.
Documented fact: Airbnb’s 2025 annual filing says performance marketing remains part of its approach and attributes lower reliance on paid channels to brand strength and communications; it also reports increased brand and performance marketing expense.
CDM inference: CDM interprets the disclosures as a system-level lesson and proposes an evidence ledger; the filing does not prove what mix another company should adopt.
What Airbnb disclosed
Airbnb’s 2025 Form 10-K describes a marketing strategy combining brand marketing, communications, performance marketing and strategic partnerships. The company says its brand and communication strategy allowed it to maintain lower reliance on paid marketing channels while performance marketing remained a component.
The same filing reports brand and performance marketing expense of $1.595 billion in 2025, up from $1.455 billion in 2024, and total sales and marketing expense of $2.588 billion, up from $2.148 billion. The percentage of revenue reported for total sales and marketing rose from 19% to 21%.
These facts prevent a simplistic reading. “Lower reliance” does not mean no paid marketing or falling marketing expense. It describes the role paid channels play in a wider acquisition and demand system.
Fact/inference boundary: Airbnb states the mix and reports costs. CDM’s explanation of persistent demand mechanisms is an analytical framework. The filing does not disclose a complete causal decomposition of bookings.
Define paid reliance before trying to reduce it
Paid reliance is the extent to which incremental demand, customer access or commercial continuity depends on continued marginal paid exposure. It is not the share of spend labelled “performance” and not simply paid traffic as a percentage of sessions.
A marketplace can receive direct and unpaid demand because customers remember the brand, open the application, follow a saved link, respond to communications, value product features or encounter supply that itself creates distribution. Paid media may still support new markets, retargeting, category capture and specific launches.
Do not rename performance activity as brand to improve a ratio. Define channels by mechanism, buying model and decision use. Sponsorship can create memory, distribution or both. Search can harvest existing demand or introduce a new category. The label alone does not reveal incrementality.
Build a Paid-Reliance Evidence Ledger
| Mechanism | Evidence to record | Counterfactual question | Common error |
|---|---|---|---|
| Brand memory | Direct navigation, branded demand, research and repeat behaviour | Would suitable customers seek the brand without current paid exposure? | Treating all direct traffic as brand-caused |
| Communications | Earned reach, message penetration, referral and sentiment with context | Did the information change discovery or evaluation? | Equating coverage volume with demand |
| Product distribution | Saved use, notifications, supply, sharing and product-led loops | Does the product create a reason to return or invite? | Calling every product event organic acquisition |
| Performance media | Spend, marginal response, saturation, audience and incrementality | What demand disappears when exposure changes? | Using platform-attributed conversions as causal proof |
| Partnerships | Audience access, integration, content and commercial terms | What capability or distribution exists only through the partner? | Assigning all partner demand to brand |
CDM asset: Paid-Reliance Evidence Ledger
For each mechanism, record the date, segment, intervention, spend, observed demand, comparison design, customer quality, uncertainty and next decision. Keep management statements, accounting categories and analytical inference in separate fields.
Use the ledger to identify durable demand and hidden dependency. Branded search can rise because paid media introduced the brand. Direct visits can depend on offline campaigns. An application install acquired through media can later generate unpaid sessions. Attribution labels do not show the full sequence.
Test dependence at the margin
Use controlled reductions, geographic holdouts or staggered changes where feasible. Predefine the outcome window and monitor total demand, new customers, repeat use, contribution and spillovers across channels. Avoid switching off activity where the risk to customers or partners is unacceptable.
Measure recovery and substitution. When paid exposure falls, customers may arrive through another trackable route even though total demand is unchanged. That is different from demand disappearing. Conversely, reported efficiency can improve because the most difficult audiences are no longer reached, while growth slows.
Evaluate product and brand changes alongside media. A new service, improved supply or clearer proposition can alter direct demand. If several actions move together, report the combined system and resist single-channel causal claims.
Apply the lesson without copying a marketplace
- Define paid reliance for the business and decision.
- Map brand, product, communications, partnership and media mechanisms.
- Reconcile spend and outcome definitions with finance.
- Identify a reversible marginal test.
- Monitor total demand and customer quality, not platform credit alone.
- Record substitution, lag and external changes.
- Invest in the mechanisms that create useful persistent demand.
A young brand may reasonably depend on paid distribution while building memory and access. A category with low repeat may never resemble a travel marketplace. The objective is not the lowest paid share; it is a resilient, economically valid demand system.
Limits and risks
Public filings aggregate costs and outcomes and reflect management’s reporting categories. They do not expose every campaign, model or experiment. Airbnb’s two-sided platform, supply network and global brand make direct comparison difficult.
Reducing paid investment can damage discovery before other mechanisms are ready. Tests require adequate duration, power and operational guardrails. Refresh the case after material filings.
CDM verdict
Airbnb’s filing supports a nuanced lesson: paid marketing can remain important while the business builds stronger direct and unpaid mechanisms. Define dependence, measure the margin, and account for brand, product and communications together. The goal is not to escape paid media; it is to avoid confusing rented reach with the entire demand system.
Related guides
Frequently asked questions
Did Airbnb stop performance marketing?This article is editorial guidance. Apply the principles in proportion to your market, evidence, and responsibilities.
