1. Lyft’s take rate (net revenue/bookings) of 25%+ is surprisingly high. For the core ridesharing business, Lyft’s bookings are the total amount of money collected from riders. However, Lyft doesn’t keep all of the bookings as revenue — the company pays out a wage to the driver, so Lyft’s net revenue is (fare quoted to rider)-(wage paid to driver). For pure marketplaces, we often see 5–15% of bookings (or GMV) as net revenue — in our analysis of eight public marketplaces, we saw a median take of 14%. In 2018, Lyft saw a ~27% revenue take rate, up from 18% in 2016.
Lyft’s take rate is increasing for two reasons — the company is: (1) giving riders fewer discounts, and taking a higher cut from drivers; and (2) paying less in driver incentives (bonuses to encourage them to drive). Lyft sees these trends as industry-wide — as ridesharing has matured, all platforms have lowered deals to get riders and drivers to engage. Lyft forecasts this take rate will increase, as they see more room for improvement in incentives. They also expect their bike/scooter business to grow (this revenue was non-material in 2018), which is 100% net revenue — Lyft owns and operates the vehicles.
While Uber’s financials are not yet public, the company’s take rate last year was reportedly close to (or upwards of) 25% — suggesting high take rates may be systemic to rideshare. In a great piece on marketplace KPIs, Accel’s Andrei Brasoveanu referenced the ability to charge a higher take if your marketplace is a “key distribution channel.” This is certainly the case for rideshare, as it would be almost prohibitively difficult for a driver to spin up their own network, particularly in contrast to a marketplace like Etsy where the seller could list items on their own website, at physical marketplaces, or in stores.
In some ways, it therefore seems fair Lyft and Uber charge a higher take — as well as acquiring customers for their drivers, they handle regulation, customer service, and insurance (we’ll get to this later!) — seller services that are a lot more intensive than those you’d find in a typical online marketplace.
2. A “mission-driven brand” is portrayed as the way to beat Uber. Lyft’s leadership has always been vocal about the importance of being a good corporate citizen, but we were surprised by how strongly this was emphasized as a competitive advantage. Frequent words/phrases in the S-1 included social responsibility (10x), authenticity (5x), empathy (4x), and mission-driven (4x).
Lyft attributed its recent U.S. market share gains (39% in December 2018, up from 22% in December 2016) to this branding resonating with consumers. Lyft has certainly benefited from Uber’s stumbles in at least a few markets (remember #deleteuber?), and the S-1 outlines Lyft’s conviction that “users are increasingly choosing a ridesharing platform based on brand affinity and value alignment.”
The S-1 also details some of Lyft’s specific actions and campaigns around corporate responsibility (present and future):
- 100% of tips go to drivers, with $750M distributed so far
- $10M donated to charities via Lyft’s Round Up and Donate program (from 700k riders)
- Free rides during natural disasters, emergency situations, and on voting days for underserved communities
- Rides are carbon neutral— Lyft is one of the largest voluntary purchasers of carbon offsets
- Partnership with cities to offer bike/scooter rides to low-income residents
- Plan to invest the greater of 1% of profits or $50M to social impact efforts annually
3. Huge insurance and incentive costs hit margins. Lyft’s gross margin is 42% (up from 19% in 2016), but the company’s EBITDA margin is -45% (up from -201.7%). These numbers are relatively low based on public marketplace comps, which typically see EBITDA margins of 5–10% — reflecting the fact that Lyft is much more complex than a pure software marketplace.
By far the largest hits on profitability are cost of goods sold (58% of revenue in 2018) and sales and marketing (37% of revenue). Within COGS, insurance costs increased by $319M in 2018, compared to only $110M for payment processing, and $75M for platform hosting. The Lyft team details insurance needs at length in the S-1 — the company provides $1M in commercial automobile liability for each driver, which is managed through their own insurance subsidiary and third-party providers.
As of EoY 2018, Lyft was holding nearly $865M in an insurance claims payment account (more than doubled from $360M in 2017). The S-1 attributes a significant increase in insurance costs in Q1 2018 due to “increased frequency and severity of claims,” though the company still appears to be somewhat cautious in estimating reserves. From 2016–2018, only 20–35% of the year’s reserves were paid out in losses.
The second largest drag on margins is sales and marketing — which is where driver referral and rider incentive (discount) costs are allocated, which represented 36% of all sales and marketing spend in 2018. Lyft paid out $296.6 million in targeted rider and driver incentives in 2018, almost double the 2017 amount of $155.6 million (which was up only slightly from $141 million in 2016).
4. Lyft has invested heavily in AV — and is closer behind Uber than they may seem. Uber’s work in autonomous driving has been heavily publicized, especially during the company’s drawn-out lawsuit with Waymo over the acquisition of Otto. But Lyft is also betting big on autonomous vehicles, which were referenced more than 100 times in the S-1. Lyft, through a partnership with Aptiv (formerly nuTonomy) has completed nearly 35,000 autonomous rides (accompanied by a human safety driver) in Las Vegas.
Of Lyft’s 4,700 employees, “several hundred” work in the company’s AV lab in Palo Alto, the Level 5 Engineering Center. Lyft plans to build its own AV system and partner with third-party developers who can build on Lyft’s Open Platform. Last year, Lyft disclosed its $72M acquisition of computer vision startup Blue Vision Labs, as well as its five-year agreement with an auto manufacturer to build AV tech — both of which are referenced in the S-1.
As a result of these efforts, Lyft hopes to serve a “portion” of rides via AV in the next five years, and the majority of rides via AV in the next ten years. Uber, by contrast, has pursued a more aggressive strategy in deploying autonomous tech — after acquiring Otto for a reported $700 million in 2016, the company started AV tests in four North American markets (at least one without a DMV permit). Uber originally planned on deploying a driverless car service by EoY 2018, but has scaled back with new CEO Dara Khosrowshahi. Similar to Lyft’s projections in the S-1, Khosrowshahi stated in early 2018 that Uber’s ability to serve the majority of a city’s rides autonomously is still 10–15 years out.
5. Bike and scooter revenue is “not material” now — but likely will be soon. You may remember Lyft’s $250M acquisition of Motivate, the largest bikeshare company in the U.S., in late 2018. According to the S-1, the acquisition gave Lyft 200 new employees, as well as a commitment to invest $100 million into the New York bikesharing ecosystem over the next five years. Based on the frequency of mentions in the S-1 (“bikes and scooters” comes up 105 times), providing short-distance EV transport is crucial to Lyft’s mission of enabling multimodal “transportation as a service.”
Lyft did not disclose bike and scooter revenue, beyond stating that it was “not material” for 2018 — which is unsurprising given the acquisition took place in November. However, Lyft’s disclosure in the S-1 that there is “no material difference” in active revenue per ride between rideshare and bikes/scooters gives us some more insight into this business line. Because Lyft has a 28% take rate on rideshare and a 100% take rate on bikes and scooters, this implies the average rideshare booking is ~3.5x the average scooter booking.
Lyft does not outline revenue per ride in the S-1, but we calculated it at $2.80 given the rides taken and revenue earned figures for 2016–2018. Since Lyft’s bike/scooter share services have only been deployed since Q4 2018, this implies the average bike/scooter booking is ~$3. At a $1 cost to unlock and $0.15 per minute to ride, this implies an average ~13 minute ride.
This is relatively in-line with data that has been publicly disclosed by Bird, Lime, Jump, and other major e-vehicle operators, though we don’t know what Lyft’s vehicle utilization looks like compared to competitors. In the S-1’s, Lyft’s listed concerns about bike and scooter operations include: reliance on and potential quality issues from third-party manufacturers, uncertain weather-related seasonality, and “negative public perception” around e-vehicle launches in new markets.
Lyft did not disclose specific plans for expanding the bike and scooter service, beyond growing a pilot that connects riders to local transit or a Lyft shared ride. The company estimates the useful life of a scooter as 12 months for depreciation purposes (which seems somewhat optimistic, given a recent report suggesting that the average lifespan is closer to one month), and as of EoY 2018 had $23M in value of scooters not yet deployed. In total, the company listed $68M of bike and scooter fleet acquisition costs for 2018.