Disruption vs. Enablement: Aggregators

“…we don’t make money when we sell things. We make money when we help customers make purchase decisions.”

- Jeff Bezos

This is part 3 of a 7 part series. If you haven’t done so already, you can read the first two posts here and here.

If the industry is plagued by the abundance of choice, customers will shop around and look for guidance or advice in seeking out the best product for them. In these cases, an aggregator might be the best strategy or what is often called the “Kayak for X” model.

The nuances between aggregators and marketplaces might seem like I’m splitting hairs, however I personally define an aggregator as a two-sided marketplace in which the aggregator acts as an intermediary between buyers and sellers, using data to route the buyer to the best seller for their needs. Both marketplaces and aggregators match buyers and sellers, however marketplaces control the supply (and as a result are liable if there is an issue) while aggregators pass the customer on to the supplier to fulfill their purchase.

To use a real world example, take ZocDoc for instance. ZocDoc aggregates doctors with their availability, ratings and network of insurers for patient booking. If a patient is misdiagnosed by a doctor discovered and booked on ZocDoc, you would have an issue with that specific doctor, not ZocDoc as a company, making it an aggregator. On the flip side, if you were to purchase a product on eBay and the wrong product was shipped, you would take that issue up with eBay who, as a marketplace, would be responsible for taking care of the issue.

This works best when the industry’s economics allow incumbents to withstand a startup commoditizing the market and therefore these startups can acquire new customers for them and improve the experience while everyone enjoys the profits of the new distribution channel.

There also needs to be enough existing companies providing products and/or services that are on equal or close to equal footing. To clarify, the market doesn’t need to be entirely fragmented but there needs to be enough supply available so that you’re not at the mercy of an 800 pound gorilla incumbent.

Unlike in the marketplace example, the products/services provided by these companies needs to consist of enough differentiation where customer preference, needs, wants, etc. do matter. The goal for these businesses is to use technology to provide guidance and information in order to “commoditize” supply to such a degree so that the customer can make an informed and empowered purchase decision.

This makes sense for certain industries, however there is a fine line between a market that fits within the above criteria and one in which the current products are vastly different and confusing to an end user in which case commoditization becomes far more difficult and a trusted third party human advisor, broker, agent, etc. still might be the preferable route for customers. I wrote about this dynamic recently when we announced our investment in Indio, a SaaS company building tools for the small commercial insurance market (more on these models in a later post in this series).

The aggregator model worked for Kayak because travel agents benefitted from information asymmetry, not because customers couldn’t make sense of various airline and hotel packages. At the same time, companies like Kayak and their competitors were made possible by two technological advances — Global Distribution Systems (GDS’s) and the Internet. GDS’s are a centralized data warehouse where all of the airlines pushed their inventory data in order to best coordinate all supply information including routes and pricing. This allowed Kayak and other companies in the space to displace traditional travel agents by building a front end on top of these systems that users could transact on.

Typically, markets that are most attractive for aggregators are those in which there is a significant barrier to entry in competing directly, but whereby technological advances have created a catalyst for an intermediary to aggregate incumbents, related metadata, and consumer preferences (as was the case with GDS’s).

Industries where aggregators work best should (at least in theory) be one with fragmented supply and product differentiation, but not to the point where by consolidating and commoditizing it further just leads to more confusion and a difficult experience for customers to navigate.

The sweet spot for aggregators might seem difficult to understand but you can think of it as a graph where the y-axis is commoditized supply and the x-axis is differentiated supply. The “sweet spot” on the line represents a market with the highest likelihood of success in an aggregator business.

Aggregator Sweet Spot

Now, even if you’ve found that perfect sweet spot, the other thing to watch out for is markets in which supply (and any metadata associated with it) is easy to access. Given that capital is abundant and it’s cheaper and easier than ever to start a technology company, the defensibility in starting one of these companies (even if you’ve found that perfect sweet spot) in today’s market lies in customer acquisition (often capital), a very risky investment to make at the early stages. I use the example of Hipmunk at the end of this post as a case study of a company with the right idea but one that couldn’t gain market share due to incumbent brand and spending power. In these cases, there can still be successful companies but identifying the attributes of the successful ones is harder and the barrier to entry is typically around brand.

The “master stroke” though for aggregators comes after there is significant demand side growth. As discussed above, aggregators typically thrive in industries where there is a significant barrier to entry for upstarts. As we’ve seen with digital content aggregators specifically, these companies can use the data they’ve compiled on their users and leverage their economies of scale with suppliers to create their own supply while their suppliers are forced into commoditization because the aggregator owns the consumer and therefore suppliers must tailor the product or service to their specifications. This is Ben Thompson’s observation which he calls Aggregation Theory and below are his thoughts on where aggregators are headed:

More broadly, I suspect many industries will in the long-run face this sort of one-two punch: the first aggregators will focus on better connecting the consumer to the supply that exists, but their domination won’t be complete because they are competing with their suppliers. Later, though, will come Version 2.0 aggregators, which not only capture consumers but also completely modularize supply that is often exclusive to their platform. For example, Google aggregated newspapers, but Facebook ultimately won the attention war by developing its own supply of content. Netflix aggregated video content, but its real long-term value is in the creation of its own; YouTube has gone in that direction from the beginning. The degree to which incumbents fend them off will be a function of just how much differentiation for the end user they provide.

This end game for aggregators aptly sums up why finding the perfect market to build these businesses can be extremely profitable and defensible over the long term. By starting out in a market where significant barriers to entry exist in building a full stack marketplace, aggregators can actually back into the very same model over time, taking advantage of better unit economics and a sustainable long term competitive advantage to become the market’s king maker at a much greater scale than the products and services they first aggregated, changing the entire structure of the industry.

Case Study:

Worked: SeatGeek

Funding: $103M

Outcome: Active

• Is fragmentation low? Yes. There are a number of marketplaces and brokers that sell tickets online however they have the same inventory and it’s difficult for the end user to make sense of the various fees, quality of seats and reliability of these sites. There also aren’t hundreds or thousands of relevant ones and therefore it is possible for SeatGeek to aggregate them.
• Are there enough existing players on equal footing? Yes. While StubHub is the 800 pound gorilla in this industry, other existing players have the same inventory and sometimes better inventory. As a result, even if StubHub has more tickets and the capital to acquire customers, they don’t own the tickets and therefore the industry’s elastic demand doesn’t allow them to control pricing in a way that can kill SeatGeek or other competitors.
• Is supply mostly a commodity? Yes, tickets are tickets and therefore SeatGeek can aggregate various ticket sites on their platform.
• Is there enough differentiation where customer/business needs/wants/preferences matter? Yes. Each ticket has two important pieces of data associated with it — the price and the location. Both of these data points are part of a consumer’s purchase decision and therefore the value SeatGeek brings is in scoring the value of each ticket.

Takeaway: SeatGeek fits quite well into the sweet spot mentioned earlier by providing a points system that “commoditizes” the ticket supply so users can make an informed and empowered purchase decision. They now also have their own marketplace so users can sell their tickets directly through SeatGeek, providing them with their rating system on the list price so that sellers can have a high likelihood of selling their tickets. This puts SeatGeek well on its way to achieving success as one of Ben Thompson’s version 2.0 aggregators.

Didn’t Work: Hipmunk

Funding: $55M

Outcome: Failed to unseat incumbents such as Kayak, Priceline, Expedia, Orbitz, etc., acquired by Concur Technologies in September 2016 for an undisclosed price.

• Is fragmentation low? Yes and no. There are many different routes flown at airports all over the world every day and there are 5000 airlines but there is a long tail here and the industry isn’t so fragmented that a pure marketplace is possible given the high barriers to entry in starting an airline.
• Are there enough existing players on equal or close to equal footing? Yes. The airline industry is one in which incumbents compete aggressively on price, service and other amenities to gain market share but it’s extremely difficult for any of their advantages to be sustainable over the long term given how easy it is to copy one another and the elastic demand of customers searching for flights.
• Is supply mostly a commodity? Yes. As discussed above, the Global Distribution Systems commoditize supply and therefore make it easily accessible to online aggregators.
• Is there enough differentiation where customer/business needs/wants/preferences matter? This aspect of the online travel aggregator industry is the one in which I believe Hipmunk underestimated. There is a small number of preferences associated with each flight including — price, flight time, layovers, and amenities. While these preferences were enough for Kayak, Orbitz, Travelocity and Priceline to gain significant market share in the earlier days of the consumer web, the “agony score” (based on duration and number of stops) and “ecstacy rating” provided by Hipmunk wasn’t useful for enough people to build a large, standalone competitor to incumbents. As a Hipmunk user, I personally enjoy their user experience and find these attributes useful, but for the broader market it isn’t useful enough to warrant a shift in consumer mindshare, especially when you consider the foothold incumbents have in the market and the war chest they have at their disposal to acquire customers and build their brand.

Takeaway: It’s worth mentioning that Hipmunk is by no means a failed company. The product is very much still live under Concur and the purchase price was undisclosed, however they weren’t able to gain enough traction to unseat or threaten incumbents, mainly because it was easy for incumbents to replicate their best features and only a small portion of the market found the additional value they provided in rating flights to be a feature of utmost importance.