TripAdvisor’s Click-to-Uber Deal
An Educated Guess at the Revenue Model Driving the API Partnership
Uber opened its API to third-parties in September and there are a number of high profile partnerships integrating its ride hailing capabilities. Integration within hugely popular apps, like OpenTable (restaurant reservations), Hinge (dating), and Starbucks (retail), provides Uber immense exposure and is incentivizing developers to use the API by offering new Uber users ride credit and (probably) some level of revenue sharing.
In an era when a public API can be a major driver of growth, and a startup’s best business development move, I wonder, how valuable are these partnerships for Uber? For their partners? And how much did it cost Uber? I decided to play with some numbers, pulling my base statistics from news coverage, SEC filings, and a few extrapolations, to reverse engineer the Uber-TripAdvisor partnership to better understand the way these deals are structured and their potential value. I’ll take you through my process and methodology and then present key learnings. The full excel model is linked to at the end of this post.
The Fun Part — Methodology and Process
To quantify the value of the partnership, I had to dig deeply into the potential market exposure TripAdvisor offers Uber, and how valuable that market might be. On the TripAdvisor side, the key variables are their monthly unique users, and the percent of those users who would hail an Uber from within TripAdvisor. On the Uber side, the key numbers are its average revenue per ride, average lifetime value (LTV) for new users, and how many active users it already has. With these numbers, I can project how much potential revenue would be generated by TripAdvisor using Uber users, and how valuable signing up a new Uber user might be. The only problem: not all of these numbers are available. Fortunately, TripAdvisor breaks down their monthly mobile visitors in their financial disclosures to the SEC. Here were the mobile numbers TripAdvisor provided for the last two years:
Over two years the growth rate is right around 80%, but growth is decelerating. I plugged in these numbers, determined the monthly growth rate, and projected a declining growth rate over five years (the rate of growth over the past two years has been declining by around 35%). Here’s how I estimated TripAdvisor’s monthly mobile visitors will look going forward:
Clearly, not all of these users will be hailing an Uber when they sign in. This was the most difficult aspect of this relationship to model. First, how do the numbers of monthly unique viewers translate into actual usage of the TripAdvisor product? And then, how often do people who are using TripAdvisor end up at one of the places they browsed? And how often do they travel to the business immediately after using TripAdvisor’s app or mobile site?
Unfortunately, TripAdvisor does not publish the statistics of clicks for directions, which answers many of these usage questions.
But Yelp does.
I set about determining the relationship between the total number of Clicks-to-Directions that Yelp has on mobile (according to presentations given to their investors) compared to their monthly unique mobile visitors (given in numbers filed with the SEC). For Yelp, the number of monthly clicks for directions as a percentage of monthly unique mobile users hovers around 16%. (In other words, if you’re a mobile Yelp user, there’s a 16% chance you used the click-to-directions function within the last month.) I used this as my starting point.
Click-to-Uber would probably be used much less than click-to-directions; people need directions much more often than they actually need rides. I arrived at a usage rate per monthly unique of around 1% for TripAdvisor monthly uniques that are active Uber users. This is based on the fact that Uber users tend to use Uber more at night (and when they’re drinking), that many destinations might be within walking distance or within easy public transport, and that an Uber might not be available in the time or might be too expensive for the TripAdvisor user’s taste. I then estimated that a non-Uber member would be about 40 times less likely to sign-up and use the app than someone familiar with the service, even assuming that Uber was even available in that location (getting someone to sign up for an account is notoriously more difficult than getting someone to use a service they’re familiar with.)
I also estimated the overlap of TripAdvisor and Uber by looking at their markets and estimated that the overlap probably was about 60% in June of 2014, increasing quickly (the overlap increasing at about 2% per month, so that by August of 2016, Uber and TripAdvisor will have 100% overlap). I took out of contention the TripAdvisor users that Uber does not yet serve. Clearly TripAdvisor won’t offer Click-to-Uber until the service exists!
Then, I estimated out how many Uber active users there are and projected how many there will be. To do this, I needed to calculate how Uber’s active users would grow since an active user would be more likely to Click-to-Uber. However and conversely, as Uber’s membership numbers grew from other sources, those potential clickers wouldn’t earn TripAdvisor any referral fees.
Luckily, Valleywag leaked some of Uber’s metrics last year and Prosperity Analytics conducted some nifty analysis to determine Uber’s monthly active users. Ballparking it, they estimate Uber had about 1.06 million global monthly active users in December 2013. Here’s how these number probably look going forward (assuming a high initial growth rate of 214% that shrinks by about 20% annually — the growth rate and decline are in part based on FutureAdvisor’s analysis of U.S. growth, but I also took into account that Uber has been focusing on growth worldwide).
Uber’s Average Ride Fare
Both FutureAdvisor, and online investment management site which performed an analysis of their customer’s car-share spending habits, and Valleywag’s leaked numbers from November calculate the average ride fare that Uber makes per ride. The FutureAdvisor numbers put it at $21. Valleywag’s numbers peg it at about $26. This discrepancy can be explained by Uber’s fare drop after the numbers were leaked. But its worth noting that Uber dropped its fares again after the FutureAdvisor numbers were calculated. Both the Valleywag leaked numbers and FutureAdvisor numbers are also for all Uber fares, but requesting a ride through TripAdvisor automatically takes you to UberX.
Taking the fare drops and the fact that you’ll be driven by an UberX into account, the average UberX fare is likely to be around $14.32. Uber takes 20% of that, meaning Uber’s revenue on that fare is $2.86. Aswath Damodoran’s excellent analysis on Uber’s valuation estimates that Uber’s profit is about 40%, so we can estimate that Uber makes about $1.15 per UberX ride.
Now, we switch gears and look at what Uber makes per customer over the customer’s lifetime. Here I used Kissmetrics’ useful framework for calculating LTV.
Here are the Kissmetrics formulas:
Simple: Periods per year x Average customer value per period x Years
Traditional: Years x (Periods per year x Average Spend x Avg. Use Per Period x Profit Margin)
Custom: (Periods per year x Average customer value per period x Years x Profit Margin)(Customer Retention Rate / (1 + Discount Rate – Customer Retention Rate)).
And Kissmetrics advocates using an average across these three formulas. Plugging in what I calculated before plus a few numbers from Prosperity Analytics and from FiveThirtyEight’s great analysis (the retention rate (54%), the number of uses per month (2), and the profit margin (40%)), yields an average LTV of $195.58. (Note that the LTV number itself is pretty close to Prosperity Analytic’s calculation of $187, though they use a different LTV formula, customer lifespan, and discount rate. Tomato, tomahto.)
With Uber’s LTV and an estimate of their revenue per ride calculated against the number of TripAdvisor users (in overlapping areas) and their conversion rate, I could then estimate how many more rides and users will be added because of this partnership, and what they’re worth to Uber. Here’s a snapshot of what it all looks like put together (though obviously we keep going about 54 more months to the right). Dive into the excel if you want to look at all the working parts.
Tweet This: Uber revenue is about $2.86 per UberX ride, LTV is around $196 — how much have its partners negotiated for?
Key Learnings: TripAdvisor Deserves a Cut of User-Signups, the Value Driver of this Deal
For Uber, the value of having a TripAdvisor member sign up for Uber for the first time absolutely dwarfs the value of having an Uber member use Uber because it’s integrated into the TripAdvisor mobile experience. That’s because the LTV for a user-signup stands at around $196, whereas Uber’s revenue for a ride is around $3.50, and its profit is around 40% of that.
As a result, even assuming that a TripAdvisor user will be about 40 times more likely to use the Click-to-Uber feature if she already has an Uber account than she will be to sign up and use it if she has never used the service, Uber still will achieve most of their value through new signups — as a result, so should TripAdvisor.
TripAdvisor’s Referral Fee
TripAdvisor negotiated a great discount for their users — to receive up to $30 for their first Uber ride (rather than the $25 they get if they sign up through a friend). I hope that TripAdvisor also negotiated a good referral fee for itself too. $20 would make sense — Uber has often provided a new signup with a free ride up to $25 and gives a $25 referral fee to a friend, for a total of $50: totaling this ($30 for its users, $20 for itself) should have been TripAdvisor’s negotiating goal.
Even with such a high a referral fee, Uber is garnering around 90% of the value. This makes sense because the service is largely provided by Uber. Uber has to contend with the costs of launching and maintaining the service in each city, the constant legal wrangles and political battles, customer service issues, and signing up drivers. TripAdvisor only has to implement an API.
In terms of the per-ride fee, I determined that Uber makes about $2.86 per ride in revenue and a profit of about $1.15. My model assumed that TripAdvisor might ask for $.50 of that, but it could be much lower. As the summary indicates, however, this number could be $0.00 and TripAdvisor would still have captured nearly the same revenue [89,856/(89,856+4,478)=95%].
Impact on Revenue
Ultimately this partnership will only have a slight impact on revenue going forward, representing about 1% of total revenue for each company. But for TripAdvisor, this revenue will be much easier than revenue that has required thousands of sales calls to local businesses. For Uber, it’s just another way to quickly add users to build its empire and undercut its competitors.
Other Strategic Value
I wasn’t able to put some of the strategic benefits into the model because they were harder to quantify. I didn’t want to give them short-shrift though, so I’ll include them here.
Improving the user-experience for TripAdvisor customers while simultaneously becoming a larger part of how they decide to patronize a business, i.e. helping “close the circle” on their transaction with that business is one of them. TripAdvisor can take these metrics (how often people took an Uber to a business page featured on TripAdvisor) and show businesses that TripAdvisor quite literally drives customers to their businesses. Being able to show this should help the sales-team close more deals — the more traffic a business gets on TripAdvisor, the more customers they get delivered to their door. The Click-to-Uber improve TripAdvisor’s product by providing its users the benefit of being able to do more from the TripAdvisor app or mobile site itself. This improvement in the user-experience, like all improvements in user-experience, should help TripAdvisor acquire and retain users.
Destroying the Competition
It’s clear that Uber itself values growth highly; it also values beating out the competition. One way of asking what this relationship might be worth to Uber might be to put it a different way: How much would Uber pay prevent Lyft from having this partnership? Lyft desperately needs to catch up (Uber is adding riders six times faster in the U.S. and probably even faster abroad); Uber likes nothing more than destroying its competition.
Targeting a Much-Sought-After Marketing Demographic
TripAdvisor also provides another benefit that not many partners can boast: studies have shown that people on vacation are more likely to try new products and services and break old habits. (See my article on Peek’s potential partnership with Lyft for more about this). TripAdvisor users are made up of exactly this crowd. With this partnership, Uber could capture people from cities that don’t yet have Uber. People love to come home from vacations and gush about their experiences — wouldn’t Uber love if its service was one of those experiences?
At the end of the day, the partnership has significant monetary and strategic value. It provides a non-negligible profit for TripAdvisor, with relatively low cost, and the added benefit of improving its product, and helps Uber grow faster, its main strategic objective. Not everything can be modeled…but the modeling sure is fun!