Marketing BS: Online Travel Agencies and Platform Economics

Good morning everyone, 

Today’s essay considers how previously learned lessons from online travel agencies can be applied to the current battles over platform economics. 

In this week’s interview, I speak with Joshua Kanter about marketing careers, recruiting, and more. 


Platform Economics 

You can’t visit a news site these days without stumbling across an article about platform economics. There is A LOT going on right now with platform economics and advertising networks — with significant amounts of money and power at stake. 

Here are some of the major stories: 

All of these stories are related. To better understand the connections, let’s take a look at the online travel industry. 

Expedia vs 

When I worked at Expedia (about a decade ago), the company was competing with for a stronger position within the European hotel market. And Expedia was struggling. 

I led a “special task force” to better understand why Expedia was losing so badly. There was a long list of ways that Expedia and Booking contrasted on strategy, but there was a particularly notable one: Expedia engaged with hoteliers in a totally different way than Booking.

When you book a hotel with an online travel agent (OTA) like Expedia or Booking, the agent takes some sort of margin and then passes the rest on to the hotelier. But there are different ways to structure deals, and Booking approached things very differently than Expedia. 

There were three key distinctions:

  1. Pay now vs. pay later. Booking allowed consumers to pay when they checked out of the hotel; Expedia charged people when they booked the hotel.

  2. Merchant model vs. agency model. Booking allowed the hotelier to “own the customer” by charging the credit card and then passing Booking’s share back to the OTA. In contrast, Expedia charged the credit card and then sent the hotelier their share. 

  3. Margin requirements and sort order methodology

Expedia executives believed that Booking was succeeding in Europe because of points 1 and 2. The internal hypothesis was based on the assumed preferences of consumers and hoteliers. Because consumers generally prefer to “pay later,” Booking had a higher conversation rate than Expedia (which compelled consumers to “pay now”). Plus, the idea that hoteliers preferred the “agency model” (i.e., the hotelier owned the customer) helped drive Booking’s higher penetration of smaller hotels (especially because Europe had a much larger percentage of independent hotels than the US).

For Expedia, there was a clear problem: adopting a “pay later” structure and shifting to an “agency model” would cost the company A LOT of money. Expedia had BILLIONS of dollars in cash sitting on hand from the 30+ days between the time they collected consumers’ upfront payments to the time they sent the hotelier their share. 

Expedia felt trapped. The company felt confident that changing to “pay later” and the “agency model” would increase hotel supply and improve consumer conversion rate. BUT, would these changes more than make up for the significant cash flow hit?

My task force investigated the problem. First, we determined that Expedia’s hypotheses about consumers and hoteliers were partially correct. Although many customers did prefer to pay when they checked out, almost half of them opted to pay when they booked their stay — even with no discount or incentives. On the hotelier side, many independent hotels experienced difficulty integrating with Expedia’s merchant model (especially in markets where Expedia was not a major presence). In other words, Expedia’s merchant model simply wasn’t worth the trouble for them. 

Most significantly, though, I found that point 3 from the above list — margin requirements and sort order methodology — was a far more important factor than most executives believed. 

Share Marketing BS with Edward Nevraumont

OTA Sort Order 

When a customer uses an OTA to search for a hotel in a specific city, they are shown a list of properties from that market. The online process for booking a hotel room gives people a lot of options; they can filter for price, location, amenities, etc. In reality, though, most people don’t use those tools; they just look at the list of hotels and examine them one by one, starting at the top. As a result, the hotels near the top of the “default sort order” are far more likely to be booked by travelers. 

Online consumers are generally lazy. To a disproportionate degree, they will click or consider or book whatever information appears at the top of a page. This behavior explains why Google has become one of the most valuable companies on the planet.

For OTAs, the fact that most consumers stick with the default sort order created an opportunity — OTAs could play with margin requirements and sort order methodology.

As for HOW to manipulate margins and sort orders, Expedia and Booking pursued very different strategies. 

At Expedia, sort order was determined via a negotiation between their market managers and the individual hotels. Expedia market managers tried to sign up as many hotels as possible, with as much of their inventory as possible, at the highest possible margin for Expedia. For the hotels, it was usually in their best interest to sign up with Expedia, but they (obviously) wanted to pay Expedia the lowest margin possible. 

On the topic of inventory, hotels wanted to strike a balance between filling rooms and maximizing margins. If a hotel had an otherwise-empty room, then selling it through Expedia was almost always the best choice (even though the hotel had to share some margin with Expedia). But if the hotel could sell a room on their own, — WITHOUT Expedia — they would be happier to keep all of the margin for themselves. 

The booking balance was further complicated by the seasonality of the tourism industry. During a peak season, like Cannes for the film festival, hotels regularly sell all their rooms. As such, most hotels did not want to give Expedia ANY inventory — why share the margin bounty with the OTA? During low seasons, on the other hand, the hotel wanted Expedia to sell as many otherwise-empty rooms as possible. Better to sell rooms with Expedia’s margin than not sell them at all.

Expedia’s market manager and the hotelier both knew what the other wanted, and they would negotiate. The market manager would push for higher margin and more inventory (especially for peak season); the hotelier would push for lower margin and better sort position (particularly during low season). The process was time consuming, but from my observations it seemed like both sides enjoyed the negotiations. Plus, the final results were never too unexpected. (Hotels willing to give up more inventory and margin would receive better sort results. Hotels that held out would get pushed to the bottom).

Booking used a completely different model to determine margins and sort order. The company did not employ as many market managers. Moreover, they did not negotiate on margins. Booking allowed the hotels to set whatever margins they wanted (above a minimum level). From there, Booking shifted the sort order to ensure the hotels that were both popular AND high margin were placed at the top of their listings. If a hotelier wanted more volume from Booking, they did not need to negotiate with anyone; hoteliers simply went onto Booking’s platform and raised the percentage of margin they agreed to pay. Almost magically, the hotel’s sort position improved and they sold more rooms via Booking. 

When peak season arrived, hoteliers did not need to pull the property off of the platform; they just went to Booking’s platform and reduced the margin to the minimum level. Hotels enjoyed a lot of flexibility with this system — they could even adjust margins on a daily basis. 

What was the net result of Expedia and Booking’s different approaches? Expedia had significantly higher average margins per hotel, but actual realized margin was about the same; Booking was sending the majority of their travelers to the subset of hotels that were paying “Expedia-like” margins. Meanwhile, Booking had far, far MORE hotels in Europe than Expedia. Hoteliers found it much easier to get set up with Booking’s platform; in addition, hotels could see that they were paying the minimum margin (so the hotels did not feel they were “paying too much”).

There is one key takeaway from this situation: both companies had a way of shifting purchase behavior towards the highest margin “products.” Expedia accomplished this through negotiation, and Booking did it with an algorithm.

After my team analyzed the problem, we had a few more questions: 

  • Was Expedia (and Booking) being aggressive enough about margins? 

  • Were there any hotels that would have given up even MORE margin in exchange for higher volume — but could not make this happen within the existing systems?

To answer the question, Expedia ran a test by adding advertisements. In addition to the “organic” listings, Expedia allowed hotels to bid on “cost-per-click” ads. This option allowed hotels to pay Expedia to get more volume. In theory, this option should not be needed; during negotiations with the market manager, the hotelier could have agreed to give up more margin in exchange for better sort position. But negotiations are never perfect. People withhold information or overpromise results. Perhaps the hotelier does not want to divulge their maximum allowable margin. Or maybe the market manager does not deliver on a better sort order.  

By offering advertisements, Expedia gave hotels a more transparent way to reveal how much margin they were willing to give up. Even if hoteliers were happy with their market-manager-negotiated deal, advertisements provided a new option to continually fine-tune their situation by trading more volume for less margin. 

Essentially, advertising becomes a tool for platforms to discover the REAL price that suppliers are willing to pay in order to access the platform’s customers. Both Booking and Expedia found a way for hotels to volunteer how much margin they were willing to give away. Booking did it directly; Expedia used advertising to do it indirectly.

Everyone wants to play (and pay) 

Over the years, bricks-and-mortar retailers have offered many ways for suppliers to advertise with them — paying to be featured in flyers or shelf position or endcaps. Building a digital ad network takes this idea to the next level and allows the same “price discovery” to happen. A decade ago, Expedia and Booking (and Google) recognized that allowing suppliers to set their maximum allowable margins could yield impressive results. 

Today, more and more platforms are realizing how much money they’re leaving on the table— unless they implement strategies that allow suppliers to volunteer more margin. For example, Amazon is a massive distributor that provides high value to suppliers who want access to customers. If those suppliers wanted more volume from Amazon, they could reduce their prices offered on the platform and/or increase the margins they are willing to give to Amazon. But when Amazon launched their ad network, they discovered that most suppliers were willing to give up more margin — a lot more. 

Apple is a different story. If a developer wants to list their product on the App Store, they must agree to hand Apple 30% of their sales revenue. In 2020, Apple earned $64 billion in revenue through the App Store — at almost 100% margin. With all of the antitrust legal pressure on Apple, their “always 30%” margins are likely to be reduced; that change would have a meaningful impact on the company’s total profit. 

But there is a way for Apple to replace the money they will lose on margin from the App Store. The company is going to offer the same “premium placement” for apps that Expedia did for hotels.

Microsoft has already fired the first salvo, reducing their PC gaming store fees from 30% to 12%. Note, however, that Microsoft is not just giving up the other 18%. There are ads placed on the top of Microsoft’s search results. Now that developers are only paying 12% to Microsoft, every click they get is worth more than it was before. So what do the developers do? They bid more per click. Microsoft has reduced the minimum payment required, but that does not mean their average payment will decrease to the same degree.

Spotify pays all artists the same amount per play (based on total revenue collected and then divided by total play time). Last November, Spotify introduced the ability for artists to pay for “priority” status. Priority songs receive placement in Spotify-created playlists and greater algorithmic consideration for personalized recommendations. In exchange for priority status for a song, the artist must agree to a lower payout per stream. This new option is Spotify’s version of sort order — and it’s likely to net higher revenue for the company. 

Let’s get back to Apple. If legal action or competitive pressure forces Apple to reduce their App Store fees from 30% to 12% (or lower), then how will developers respond? Based on the example from Microsoft, many developers will be willing to pay an additional 18% (or more) for advertisements that will increase their volume of sales. 

Using an ad platform to make up for reduced App Store fees only works if all app-related advertising flows through Apple. As such, Apple is not only building their own advertising platform, but they are also limiting the ability of rivals to manage effective ad platforms. Apple’s recent push into “app privacy” only applies to “third-party tracking.” So, apps cannot share metrics with Facebook (in order to optimize their ad spend on the Facebook platform), but they CAN share metrics with Apple to optimize their ad spend on the Apple platform. 

Any developers willing to spend money on advertising might still try Facebook (especially if ads on the social network have performed well in the past). But if Apple’s new privacy rules reduce the effectiveness of Facebook ads — which is likely, given the restrictions on data tracking — then don’t be surprised if developers shift their ad spend from Facebook to Apple. 

At the end of all this, Apple will end up just where they started (or maybe a little ahead). Facebook will lose out. Apps will perform a little worse (they lose significant ground on the Facebook platform, but make up some of that on the Apple platform).

There is a key point to remember about platform economics: when there is margin to be made (and there is a lot on zero-marginal-cost apps), then the market will “bid up” the advertising to eat into that margin. Then it’s just a matter of which player in the stack gets to keep the surplus.

When platforms are launching, the surplus goes to the suppliers. Then the suppliers figure out how to better advertise, which shifts the surplus to the effective advertising platforms. Finally, the platforms get smarter and find ways that the suppliers can “keep” the surplus on the platform.

Keep it simple,


Edward Nevraumont is a Senior Advisor with Warburg Pincus. The former CMO of General Assembly and A Place for Mom, Edward previously worked at Expedia and McKinsey & Company. For more information, including details about his latest book, check out Marketing BS.

Share Marketing BS with Edward Nevraumont