Aggregator business models

What drives the choice and the evolution of business models aggregators adopt

“Our differences are the real treasures.” ― TemitOpe Ibrahim

Of all the business models that tech has enabled, aggregators are the most prominent ones. They have come to dominate the industries they are in. While some of them might look similar on the surface, they have evolved very differently, and continue to do so, based on the competitive/industry dynamics. While I am sure there might be a few different interpretations of the term aggregator, here I am sticking with the way Ben Thompson defines them — direct relationship with users, zero marginal costs and decreasing acquisition costs. The focus of aggregators is to serve demand as opposed to, say platforms, that serve the builders.

My attempt here is to articulate the business models I have seen aggregators adopt in their goal of serving the buyers/demand, and how aggregators might think about choosing a business model over the other.

The 3 business models

• Model 1 — PLUG into supply: Plug into supply sources that connect with the actual suppliers
• Model 2 — BUILD supply: Build supply through direct relationships with suppliers
• Model 3 — BECOME supply: Become a supplier to other aggregators

🔌 Model 1 — PLUG into supply

This model involves plugging into sources of supply which, in turn, connect with the actual suppliers. An example would be Amazon or eBay leveraging Shopify or BigCommerce to let sellers sell on their sites without engaging with the sellers directly. This is in contrast to Amazon or eBay having some sellers directly use Amazon or eBay’s UI/APIs to sell.

For this model to be an option, the industry needs a certain maturity and standardization; e-commerce is a perfect example of this. Most of the products we buy have UPC codes to uniquely identify each SKU. These UPC codes have become the common language across the different systems, almost like a standard that creates trust in the value chain. eBay doesn’t need to ask Seller XYZ what she is selling because the seller already told Shopify that she is selling product ABC, and eBay understands what that product is.

As the demand for e-commerce has increased, the number of demand channels that a seller could sell through has also increased e.g. smaller e-commerce aggregators such as GOAT for sneakers, Rebag for luxury handbags, etc. So, it makes much more sense for sellers to start engaging with these channels through BigCommerce or Shopify, which become the source of truth for inventory management. In essence, e-commerce industry has matured into making this plug & play model a very feasible option for any future aggregators such as Facebook Shops.

This model makes the supply available on a platter to anyone who can build demand, breaking part of the moat aggregators such as eBay built through years of supply acquisition. That said, eBay still has some differentiated supply which leads to defensibility, and Facebook Shops is building differentiated supply as well. However, generally speaking, supply acquisition in e-commerce is much easier now.

This plug & play model also an option in mature but less standardized industries such as hotel booking. Only some hotel booking aggregators actually engage with the hotels directly. TripActions, a corporate travel management startup, integrates with the sources that connect to the hotels, sometimes through multiple systems. However, unlike e-commerce, hotel booking is a more rudimentary version of plug & play. Aggregators such as TripActions still have to do a lot of heavy lifting to reconcile the hotel and room inventory across multiple sources because there is no unique code for a particular room in a particular hotel. The common language, unlike UPC in e-commerce, in the hotel industry is primitive.

⚒️ Model 2 — BUILD supply

This model involves the aggregator building the supply base from ground up through a direct relationship with the end supplier e.g. eBay with its sellers, Uber with its drivers, DoorDash with its couriers and Thumbtack with its professionals. Any early aggregator in any industry would end up adopting this model because there doesn’t exist a ready supply source to plug into.

Another scenario in which aggregators opt for this model is when they are bringing differentiated supply to established markets such as Lodging. This is what Airbnb did when they made homes/apartments available as stays. Airbnb had to build direct relationships with the hosts because these hosts were not used to opening up their places for stays before Airbnb came in.

There are also scenarios when the supply is available as plug & play but the aggregator’s value proposition for the demand rests on it going direct to supply. As example of this is HotelTonight. While there existed plenty sources to plug into thousands of hotels without the need to directly engage with the hotels, HotelTonight chose not to go that route. It negotiated proprietary deals with hotels and gave them an extranet to upload prices that were only to be found on HotelTonight. These last minute deals to stay at a “hotel that night” is what got the demand interested; they couldn’t find these on Expedia or Priceline (Booking.com).

HotelTonight’s promise to the hotels was that it allowed them a “safe way” to improve occupancy i.e. improve occupancy without negatively impacting their overall yield (margin). It also promised them high quality traffic (travelers) who could become loyal customers. It took a more hands-on approach in choosing the hotels it sold in app, even sending professional photographers to photograph hotels and building a perspective on each hotel. More on its approach here.

This model obviously is hard because the aggregator needs to acquire both demand and supply at the same time, but that it also what makes aggregators opting for this model a bit more defensible. For example, its not possible for an aggregator to come in and use the existing supply sources to offer hotel deals better than the ones Expedia and Booking offer unless the aggregator starts engaging with the hotels directly.

⏩ Model 3 — BECOME supply

This models involves the aggregator itself becoming a supplier to other aggregators. This is not the model aggregators start with but is one they might evolve into. I have seen it happen in at least two scenarios. The rationale in both the cases is for aggregators to make the most of their supply knowing that they don’t have a very strong demand pull. This lack of demand pull could be isolated to company-specific reasons or just a reality of the industry the aggregator is in.

An example of the former is eBay launching “eBay Buy APIs” in 2016 to enable apps to let their users buy items from eBay without the need for the user to visit eBay to complete the purchase or even know that the item is coming from an eBay seller. Though not very successful, these APIs have still generated $1BN in GMV for eBay. eBay did this because it was growing slower than the e-commerce market and needed to retain its hard-earned supplier base, some of which is differentiated i.e. sells minimally outside eBay. Amazon hasn’t done any such thing because it has the demand pull and doesn’t need to dilute its margins and competitive positioning by offering Buy APIs.

An example where the industry structure mandates this model is hotel booking. Both Expedia and Booking Holdings (Priceline) expose their supply to apps via APIs. They do that because travel is an infrequent activity, happening a couple of times a year, making it hard for aggregators to build brand loyalty. Customers acquired for one travel occasion might need to be reacquired for the next one. This is why OTAs (Online Travel Agencies) such as Expedia and Booking one of the largest ad spenders on Google. While OTAs make less revenue on an outside transaction (i.e. on an app that uses their Buy APIs) compared to the revenue on a direct transaction (i.e. directly on their site), they can use these outside transactions to negotiate lower prices and higher commissions with their suppliers (hotels). These higher commissions apply to all transactions, direct or not, making the Buy APIs overall margin accretive. In general, increased demand keeps the flywheel spinning i.e. enables the aggregators to maintain a strong hold on supply, which allows them to continue serving the demand better (lower prices and more availability).

The continuous evolution

Industries and companies are constantly evolving. Aggregators might adopt multiple models from start, or start with one model and grow to adopt other models as well. Expedia and Booking started with Build Model, and now have also adopted both the Plug and Become Models. Airbnb also started with the Build Model and might have to consider the Become Model as it faces increasing competition from both Expedia and Booking doubling down on alternate accommodations.

While Plug Model is a lower effort/risk way to test demand for a new app/concept, it isn’t the most margin efficient because of the revenue share for each transaction. If an aggregator becomes successful, they might be driven to move partially from the Plug Model to the Build Model. The need to own differentiated supply and build a moat might be another reason to want to move towards the Build Model.

On the contrary, investing too much into the Build Model in an industry where new aggregators, driven by the Plug Model, are constantly trying to take a share of demand, might lead to one missing key demand trends. This could result in a weakening, instead of strengthening of the aggregator’s moat.

Choosing the model to invest in is, therefore, a complex equation of short and long-term profitability, state of the industry and competition, and really understanding the role the differentiation of one’s supply plays (or can play) in driving demand to them as an aggregator. Hope this article served as somewhat of a guide of how to think about choosing a business model as an aggregator at a point in time, and in the future.

P.S. Opinions in this article are my own and don’t reflect the opinions of any of the companies mentioned

🐦 Twitter @setparth