The other one does not.
I think you can guess which one is which.
For the record - I'm just as skeptical as many others are about tech valuations, but there is a legitimate reason to value companies they way we do, it's just not always a "its tech, so it has to be valued at BLAH", there's a lot of nuance to these businesses.
Mazda?
Because according to their SEC filings Duolingo has pretty limited growth potential, increasing losses (or in other words, negative gross margins), and they spend a significant amount of money to attract each new paying customer. They only approach 80% gross margins if you use fantasy unicorn accounting instead of GAAP.
> Because according to their SEC filings Duolingo has pretty limited growth potential,
The global market for direct-to-consumer language learning is large, growing, and shifting online. According to HolonIQ, total consumer spend on both online and offline language learning represented a $61 billion market in 2019, and will grow to $115 billion in 2025, implying a CAGR of 11% over this period. Online language learning is the fastest-growing market segment, projected to grow from $12 billion in 2019 to $47 billion in 2025, representing a CAGR of approximately 26% over this period, and to comprise 41% of total consumer spend on language learning in 2025. We believe that growth in digital spend will be driven in part by a shift away from offline offerings, as consumers seek more affordable, convenient, and higher quality online solutions.[0]
That's not how losses and negative gross margins work. I just double checked, gross profit for DuoLingo in 2020 was 71.55% (sorry not 80%). Compare that to Mazda's abysmal 21.7%. Cars specifically require heavy R&D (which does not go into COGS), so the comparison is a nice one since R&D for DuoLingo also requires heavy R&D.
> spend a significant amount of money to attract each new paying customer
Customers who pay subscriptions, not a one-time car cost. Every new car created requires upfront capital to build the car. SaaS CLV vs a car CLV requires WAY less costly touch-points after the initial sell.[1]
[0] - https://www.sec.gov/Archives/edgar/data/1562088/000162828021...
[1] - https://a16z.com/2014/05/13/understanding-saas-valuation-pri...
Duolingo's "gross profit" is not calculated using GAAP. It's calculated using fantasy unicorn accounting. Key giveaway: if a company's revenue "more than doubles" but it's losses also more than double, it's not actually profitable on a per-unit basis; it's simply shifting items that should be reported in COGS to other items that let it artificially inflate its gross margins. In Duolingo's case, it's actually worse: while revenues doubles, losses are on track to quadruple (they reported a $13.5 million loss for just 2021Q1, compared to roughly $16 million for all of 2020).
I've worked for enough startups going public that I can spot when a startup is playing games with its financials. Duolingo is one of them. This time next year, Duolingo will be looked on as 2021's Groupon.
These are hilarious businesses to cherry pick because they explicitly used fuzzy accounting for the nature of their industry - which was selling coupons that may or may not be used. Their numbers were irregular because they accounted for coupons that may have never been actually used, but were allocated on their books as such. DuoLingo is NOTHING like that. People pay a subscription for a service that uses software as its delivery mechanism. COGS is simple for them -> hosting + customer service. COGS is very different for Groupon and LivingSocial.
> Key giveaway: if a company's revenue "more than doubles" but it's losses also more than double, it's not actually profitable on a per-unit basis; it's simply shifting items that should be reported in COGS to other items that let it artificially inflate its gross margins.
You have no idea what you're talking about and clearly have never actually seen financials of a business like DuoLingo.
Guess when NetSuite's most profitable year was? That's right, 2009, during one of the biggest recessions in the world's history, and it's because they fired their growth engine (S&M) and just pumped out cash.
https://a16z.com/2015/05/15/a16z-podcast-why-saas-revenue-is...
:) My list of clients when I was still working at a firm include many tech companies ranging in size from vastly unprofitable startups to multi-billion dollar public companies, in industries including robotics, hardware, SaaS, PaaS/IaaS, biotech, media tech, aviation, and clean tech.
A number of YC startups were clients, including some that are still actively discussed on these forums today. At least one of the Groupon, LivingSocial, etc., cohort was a client...
DuoLingo is NOTHING like that. People pay a subscription for a service that uses software as its delivery mechanism. COGS is simple for them -> hosting + customer service. COGS is very different for Groupon and LivingSocial.
DuoLingo's hosting costs should be low. But content development costs should be included in COGS (being pedantic: COGS doesn't apply to non-inventory business activities, so we should be using the term COR instead). Content development and customer service aren't cheap.
And of course, gross margins are irrelevant if you're constantly blowing past your revenue numbers by overspending on "product development" (i.e., the tech side). History is littered with companies who claimed to have great "gross margins" but died because they overspent on "product development."
Please do tell.
> DuoLingo's hosting costs should be low. But content development costs should be included in COGS (being pedantic: COGS doesn't apply to non-inventory business activities, so we should be using the term COR instead). Content development and customer service aren't cheap.
I agree with you here, except for specifics. Do we know how much content costs? (ps, once you create content for a language it can scale infinitely - as opposed to having teachers in a classroom who can't). We don't know how much customer service costs / personnel they have, BUT it is part of COGS (as per their definition) so its impossible to qualify that hosting "should be lower" unless you know their specific hosting costs.