We have updated the Uber numbers following the submission of their IPO paperwork (Uber Technologies S-1 filing).
In its much anticipated S-1 filing to IPO, Lyft revealed a wealth of financial detail. Finally, one of the most debated business models in venture capital is opening up to outsiders. And as Uber released selected figures in the run-up to its own IPO, it’s now possible to compare the two companies in more detail than ever before. More importantly, we can see what the path to profitability might look like.
The black numbers are either reported by the company’s or implied from those reported numbers. The grey numbers are estimated by Dealroom based on quarterly trends, benchmarks for similar companies and some reasonable guesswork.
Insight #1: Lyft grew revenues 2x faster than Uber, but…
Lyft grew revenues 100% to $2.2 billion, ahead of market expectations. Meanwhile, Uber grew revenues “only” 42% to $11.2 billion. Part of Lyft’s high revenue growth was driven by a jump in the take-rate (from 23% to 27%), which could be temporary. It’s better to look at gross bookings, where Lyft grew 76% YoY compared with 44% YoY for Uber. That’s a far less dramatic difference, especially taking into account the 5x difference in size.
Still, in Q4 Uber growth slowed down a bit further: Uber’s gross bookings increased 37% YoY and revenue increased 24% YoY.
Insight #2: Winning market share in the U.S.
Lyft has been gaining market share in the U.S. and claims 39% market share as of December 2018 by number of rides in the U.S.. Data analytics firm Second Measure, which analyzes credit card data, estimates Lyft’s market share at 29% by sales, compared with 69% for Uber. Momentum seems to be in Lyft’s favor. Part of this could be attributable to Uber’s PR problems (e.g. the #DeleteUber campaign).
The company said it had 18.6 million quarterly active riders as of December 2018. For the full year Lyft counts 31 million active riders. By comparison, Uber reports 91 million *monthly* active riders, globally.
Insight #3: average Lyft driver ≠ average Uber driver
Especially from the supply side (driver side), the two companies are quite different. The majority of Lyft’s drivers “drive in their free time to supplement their income.” Uber riders are mostly full-time. As a result, the average Lyft driver earns $4.3K per year, a number that’s been increasing. Meanwhile, the average Uber driver earns $16.7K per year, a number that’s been decreasing. The average Lyft driver does 1 ride per day (326 per year), whereas the average Uber driver does 5 rides per day (1,773 per year). Uber has only 2x more drivers than Lyft, despite Uber being 5x bigger (3.9 million vs 1.9 million drivers).
Also, Lyft’s revenue per ride is $13 per ride, whereas Uber’s $10 per ride. This is likely due to Uber’s presence in more densely populated areas, and its global footprint, which includes emerging markets. Lower revenue per ride isn’t better or worse per se. What matters more is overall efficiency. But it’s interesting to note nonetheless.
Insight #4: roughly similar take-rates, trending upwards for Lyft, down for Uber
Both companies have roughly similar take-rates, i.e. the percentage net revenues of gross bookings. However, both numbers are heavily distorted. Lyft has eScooter and other services, where net revenues equal bookings. We therefore cannot conclude too much from this. The below chart from Lyft’s S-1 shows the increase of Lyft’s take-rate. In theory, it could go up for a long time, driven by eScooter services (near 100% take-rate) and in the longer-term autonomous vehicles (also potentially 100% take-rates, but still very far away). In the very long-term this could drastically change the profitability of the business.
Insight #5: path to profitability
Back to the near term. Lyft’s margins have improved dramatically since 2016, as the following chart shows. Total costs now are 136% of revenues, down from 367% in 2016. But how many years will it take to become profitable?
Lyft’s biggest expense item, similarly to Uber, is Cost of Revenue, which were $1.2 billion or 55% of revenues in 2018 (down from 100% in 2016). This primarily consists insurance premiums, regulatory costs, payment processing fees (partly Stripe) and hosting fees (it pays AWS $100 million per year!).
The next biggest expense item is Sales & Marketing, which has fallen dramatically from 133% of revenues to 36% in 2018, despite the aforementioned market share gains. In absolute terms it still increased to $0.8 billion.
In 2018 Lyft seems to have acquired roughly 13 million new riders (grew from 19 to 31 million riders, assuming almost no churn) and acquired roughly 0.6 million new drivers. Therefore it costs about $70 to acquire one new rider. Net revenue per rider is roughly $70 too, but the contribution per rider is roughly half of that, so it takes about two years to recoup these Sales & Marketing costs (in the absence of more precise data). While two years would be way too long for most other businesses, Lyft’s customers are extremely sticky.
The below chart shows the rides by cohort, demonstrating that activity by cohort actually increases, rather than decreases. We can therefore assume that Sales & marketing costs as % of sales could reduce dramatically in the long-term, to less than 10%.
The remainder of costs are General & Admin, Operations & Support and Research & Development. Together these comprise of about 45% of revenues. Without having seen any equity research yet, we’ve used the available cohort data and some estimates to see what a path to profitability could look like.
It will be interesting to see the equity research which has had the benefit of management briefings. But it seems likely that Lyft could reach profitability in 2020 (perhaps even late 2019) if it wants to.