The Business Model Canvas, drawn for AirAsia and Airbnb

On 8 September 2001 Tony Fernandes and Kamarudin Meranun bought AirAsia for RM1; by the airline's own account it owed RM40 million at the time (AirAsia Newsroom, 2022). Six and a half years later, in March 2008, Airbed & Breakfast, the first name of Airbnb, launched during SXSW and took just two bookings (Airbnb, n.d.). Both sell travel, yet on the same Business Model Canvas the two pictures look almost nothing alike.

What does the canvas ask, and how much can you draw from public sources?

It asks how a company creates, delivers and captures value. Osterwalder and Pigneur (2010) define a business model with exactly those three verbs (Osterwalder and Pigneur, 2010, p. 14) and say the nine blocks cover four areas: customers, offer, infrastructure and financial viability (Osterwalder and Pigneur, 2010, p. 15).

Osterwalder, Pigneur and Tucci (2005) had reached nine blocks earlier under different names, and deliberately excluded competition and implementation, which they saw as related to the model but not part of it (Osterwalder et al., 2005).

I have three sources for these two companies: The Star's 2018 interview with Fernandes (Sidhu, 2018), AirAsia's 21st-anniversary release (AirAsia Newsroom, 2022) and Airbnb's own about page (Airbnb, n.d.). Two of the three are the companies telling their own story. Each cell below is marked as what the source says, what I infer, or what the source does not say.

AirAsia: nine blocks around one cost block

Cost structure is the block that holds this model together, because the value proposition is a low price. Osterwalder and Pigneur note that low-price propositions have important implications for the rest of a business model, and name no-frills airlines as an example (Osterwalder and Pigneur, 2010, p. 24). The library links AirAsia to Porter's generic strategies as an illustration of a low-cost strategy.

• Customer segments: people who rarely or never flew. Fernandes describes someone who used to fly home to Kota Kinabalu or Kuching once a year because of fares and now makes the trip every month (Sidhu, 2018); the head of cabin crew says many passengers were flying for the first time (AirAsia Newsroom, 2022).
• Value proposition: price. One-way fares as low as RM1 and the tagline "Now everyone can fly" (Sidhu, 2018).
• Channels: the sources do not say where tickets are sold.
• Customer relationships: in my reading, a brand and one person. A report quoted by The Star calls Fernandes the airline's most effective marketing and public relations tool (Sidhu, 2018).
• Key resources: a fleet that grew from two aircraft in 2001 to 216 A320s and 31 A330s in 2018, and 24,000 staff (Sidhu, 2018).
• Key activities: operating routes. The first international route, to Phuket, opened in 2003; by 2018 there were about 90 routes and in 2017 the group carried 65.7 million passengers at an 88 per cent load factor (Sidhu, 2018). Filling seats is the core activity in my reading, because it links the low fare to revenue.
• Key partners: the sources name only the seller, DRB-Hicom (Sidhu, 2018).
• Revenue streams: tickets are the obvious one. Revenue of RM9.7bn and net profit of RM1.59bn to the end of 2017, as The Star reports them (Sidhu, 2018); the sources give no split by stream.
• Cost structure: Fernandes says the strategy is "cost, cost, cost" and that cost discipline let the airline absorb shocks from oil prices to pandemics; the article says its costs are among the lowest globally (Sidhu, 2018). No figures.

The canvas also prompts a question. Fernandes says the airline went from 90 million passengers a year to grounding its entire fleet at one point (AirAsia Newsroom, 2022), and Osterwalder gives salaries as an example of fixed cost, which does not fall when volume does (Osterwalder and Pigneur, 2010, p. 41). The sources do not give AirAsia's fixed costs, so that is a question for whoever draws the canvas, not a finding.

Airbnb: nine blocks around two linked groups

The customer segments block holds this model together, because there are two groups and each needs the other. Osterwalder says a multi-sided platform serves two or more interdependent segments and works only if all are present (Osterwalder and Pigneur, 2010, p. 21). Göcke and Meier (2021) use Airbnb itself as their example: it has to appeal to guests (demand) and hosts (supply) alike (Göcke and Meier, 2021, p. 89). The figures below come from the Airbnb page, marked as updated in May 2026.

• Customer segments: over 5.5 million hosts and over 2.5 billion guest arrivals to date (Airbnb, n.d.).
• Value proposition: for guests, stays, experiences and services that, in the page's words, connect them with communities. For hosts, income: hosts have earned over $380bn and the typical US host earned $15,600 in 2025. Trust is layered on top: a $1m Host Guarantee (2012) and AirCover for hosts and guests (2021, 2022) (Airbnb, n.d.).
• Channels: website, app and Instant Book (November 2010); international expansion began with a German office (2011), and a separate Chinese brand followed (2017) (Airbnb, n.d.).
• Customer relationships: in my reading, programmes rather than one-to-one contact, such as 1 million Superhosts (2023) and the Co-Host Network (2024) (Airbnb, n.d.).
• Key resources: the platform itself. Its own payments system launched in August 2008 and processed about $70bn in 2019; over 9 million listings are active (Airbnb, n.d.). The page describes the stays as offered by hosts and does not say Airbnb owns any.
• Key activities: for a model built on a platform, Osterwalder lists platform management, service provisioning and platform promotion (Osterwalder and Pigneur, 2010, p. 37). The page's timeline fits: payments, yearly releases, Experiences (2016), Services (2025), the HotelTonight deal (March 2019) (Airbnb, n.d.).
• Key partners: city governments, from housing displaced residents after Hurricane Sandy with New York (2012) to City Portal (2020); the International Olympic Committee (2019) (Airbnb, n.d.).
• Revenue streams: the page does not say. Osterwalder lists brokerage fees as the stream of an intermediary linking two parties (Osterwalder and Pigneur, 2010, p. 32), so that is a candidate, but the page does not say which side pays or how much. The $18bn in taxes collected and remitted is not revenue either (Airbnb, n.d.).
• Cost structure: the page does not say. From the activities I would guess platform development, payments and trust programmes; that is inference, with no figures.

Where do the two canvases differ?

The last row is easy to miss: Airbnb's page counts what hosts earn, not what Airbnb earns. Which cells are empty tells you where you are guessing.

Göcke and Meier list four characteristics of a platform: multi-sidedness, network effects, the chicken-and-egg problem and a tendency towards winner-takes-all or winner-takes-most (Göcke and Meier, 2021, pp. 89-91). Airbnb meets the first, the chapter's own example. For the airline, in my reading, the sources show no second side, although AirAsia's 2022 release says it has become a one-stop-shop platform with ride-hailing, food delivery, logistics, fintech and aircraft engineering (AirAsia Newsroom, 2022); the word "platform" in a press release does not say whether those lines have two sides. How Airbnb solved the chicken-and-egg problem at launch is in the post on platforms and the chicken-and-egg problem.

What should a homestay owner do with the two canvases? (invented example)

The example is mine. A homestay owner in Hoi An has 10 rooms, takes bookings through Facebook and the phone, and is weighing a listing on a booking platform. Draw the current canvas, then redraw it with the platform: channels gain a route, customer relationships gain the platform's reviews standing between you and the guest, key partners gain the platform itself, and costs gain whatever the platform charges (that depends on the platform, and I have no source).

You appear on two canvases. On Airbnb's, a host is a customer segment; on the homestay's, the platform is a channel and a partner. In my reading, since a platform needs both groups (Osterwalder and Pigneur, 2010, p. 21), ask how much the platform needs hosts in your area before you accept its terms.

For every cell marked unknown, write down who can answer and what to ask, for instance what share of the room price the fee takes, then do the arithmetic before deciding: the canvas will not do it for you. This suits you if you need to describe or compare two options on one page. It does not suit you if you need to decide about money: the nine blocks hold no figures and no rivals.

Going deeper: what I read, and what I could not check

For Osterwalder and Pigneur (2010) I read the free 72-page preview, which holds the definition, the nine blocks and the canvas page; I did not read the multi-sided platforms pattern chapter or the strategy section. The file is dated 2009. Page numbers are those printed in the preview; I have not compared them with the Wiley 2010 edition. For the 2005 paper I read the authors' copy hosted by KTH, whose pagination differs from the journal's, so I give no page numbers. For Göcke and Meier (2021) I read the full open-access chapter.

For the two companies I have only their own accounts and one interview, with no financial statements. The two AirAsia sources disagree on the debt at purchase: RM40 million according to the airline, RM190 million in liabilities according to The Star, and neither explains why. The Star does not say which entity or profit measure its RM1.59bn covers; the 16 per cent is a ratio of two press figures. The Star's figures are from 2017 and 2018; I have not checked how the airline stands now.

This post does not show that the canvas improves decisions, and both cases are companies that survived. Which block holds each model together is my reading.