Last month, Scott Stevenson, co-founder and CEO of legal AI startup Spellbook, took on X to expose what he called a “huge fraud” among AI startups: inflating the revenue numbers that companies report.
“The reason so many AI startups are crushing revenue records is because they use dishonest metrics, backed by the world’s largest funds and misleading journalists for PR reporting,” he wrote in a tweet.
Stevenson is not the first to claim that annual recurring revenue (ARR), a metric historically used to sum up the annual revenue of active customers under contract, is being manipulated beyond recognition by some AI companies. Certain aspects of the ARR scam have been the subject of multiple other news reports and social media posts.
But Stevenson’s tweet seemed to particularly touch a nerve within the AI startup community, garnering more than 200 reshares and comments from prominent investors and a number of founders, and garnering several headlines.
“Scott at Spellbook did a great job of highlighting some of the nefarious practices of some companies,” Jack Newton, co-founder and CEO of legal startup Clio, told TechCrunch, adding that the post brought much-needed recognition to the topic, referring to YC’s Garry Tan’s explanatory post on appropriate revenue metrics.
TechCrunch spoke to more than a dozen founders, investors, and startup finance experts to assess whether ARR inflation is as prevalent as Stevenson suggests.
Indeed, many of our sources, speaking on condition of anonymity, acknowledged that ARR fabrication in public filings is common among startups, and that investors are often aware of the exaggeration.
There is no actual profit yet
The main obfuscation tactic is to simply call it ARR, replacing “contracted ARR”, also known as “committed ARR” (CARR).
“Sure they are reporting CARR as ARR,” said one investor. “When a startup does something in a certain category, it’s hard not to do it yourself to keep up.”
ARR is a well-established and trusted metric from the cloud era that indicates the total revenue of a product where usage, or payment, is measured over time. Accountants do not formally audit or approve ARR because Generally Accepted Accounting Principles (GAAP) focuses on past revenue already collected, rather than future revenue.
ARR was intended to represent the total amount of signed and stamped sales (usually multi-year contracts). (Today, this concept tends to go by another name: remaining performance obligation.) On the other hand, the term “revenue” usually refers to money that has already been collected.
CARR is considered another way to track growth. However, it is a much more complex metric than ARR because it counts revenue from signed customers who have not yet been onboarded.
One venture capitalist told TechCrunch that he has seen companies where CARR is 70% higher than ARR, even though a significant portion of contract revenue never actually materializes.
Bessemer Venture Partners (BVP) wrote in a 2021 blog post that CARR “builds on the ARR concept by adding committed but not yet active contract amounts to the ARR total.” But importantly, BVP says startups are supposed to adjust CARR to take into account expected customer churn (the number of customers who leave) and “downsells” (customers who decide to reduce their purchases).
The main problem with CARR is calculating revenue before a startup’s product is implemented. If implementation is slow or unsuccessful, the client may cancel during the trial period before all or part of the contract revenue is collected.
Multiple investors told TechCrunch that they have direct knowledge of at least one high-profile enterprise startup that has reported more than $100 million in ARR, but only a fraction of that revenue is currently coming from paying customers. The rest is due to contracts that are not yet in place, and in some cases the technology may take a long time to implement.
A former employee of a startup that regularly reported CARR as ARR told TechCrunch that the company counts at least a year of effectively free pilots as ARR. The company’s board, which includes venture capitalists from major funds, was made aware during the lengthy pilot program that proceeds from the final payment portion of the contract were being counted toward ARR, the person said. The Board also recognized that a customer could cancel before paying the full contract amount.
The obvious problem with using CARR and calling it ARR is that it is much more “gullible” than traditional ARR. If startups don’t realistically consider churn and downsells, CARR can be high. For example, a startup could offer a deep discount for the first two years of a three-year contract and count the entire three years as CARR (or ARR), even though the customer might not continue paying the higher rate in the third year.
“I think Scott[Stevenson]is right, and I’ve heard all kinds of anecdotes,” Ross McNairn, co-founder and CEO of legal AI startup Wordsmith, told TechCrunch about ARR misstatements. “I talk to VCs all the time, and they say, ‘We’ve got some choppy, choppy standards.'”
Most cases are not that extreme. For example, an employee at another startup described a discrepancy where marketing materials listed ARR as $50 million, when the actual number was $42 million.
However, this person insisted that investors had access to the company’s books and that they accurately reflected the low amount. Some startups and their investors are comfortable using public metrics at their disposal to play fast, the people said, because AI startups are growing so fast that an $8 million difference is seen as a rounding error that can quickly add up.
Another, more problematic “ARR”
There is another problem with all these public ARR declarations. Founders sometimes use another measurement with the same “ARR” acronym and a similar name: Annual Run Rate Revenue.
This ARR is debatable because it estimates current earnings for the next 12 months based on earnings for a specific time period (quarter, month, week, or day, etc.).
Since many AI companies charge based on usage and results, the method of calculating annual run rate ARR can be misleading since revenue is no longer locked into a predictable contract.
Most people interviewed for this article said that ARR overstatement of any kind is by no means a new phenomenon, but startups are becoming much more aggressive amidst the AI hype.
“There’s more incentive to do this because valuations are higher,” Michael Marks, founding managing partner at Celesta Capital, told TechCrunch.
In the age of AI, startups are expected to grow much faster than ever before.
“Going from 1 to 3 and 9 to 27 is not fun,” Hemant Taneja, CEO and managing director of General Catalyst, said on the 20VC podcast last September, noting that startups are traditionally projected to reach multi-million dollar annual ARR. “You have to go from 1 to 20 to 100.”
The pressure to show rapid growth has led some venture capital firms to favor, or at least ignore, startups that present inflated ARR numbers to the public.
“VCs are definitely in this because they’re motivated to create a narrative that they have a clear winner. They’re motivated to get their companies covered in the press,” Stevenson told TechCrunch.
Mr. Newton, who last fall valued legal AI startup Clio at $5 billion, argues that venture capitalists are aware of ARR misstatements but are often silent. “There are investors who turn a blind eye to their companies inflating their numbers because it looks good from the outside,” he told TechCrunch.
What VCs actually think
Other investors who spoke to TechCrunch said there was no reason for VCs to reveal the exaggerations.
By turning a blind eye to public announcements about soaring ARRs, VCs are effectively helping to make their portfolio companies king. When a startup publishes high revenues, it is more likely to attract top talent and customers who believe the company is the undisputed winner in its category.
“Investors can’t blame it,” the VC told TechCrunch. “Everyone has a company that monetizes CARR as ARR.”
Still, anyone familiar with the industry’s intricacies would find it hard to believe that some of these startups actually reached $100 million in ARR within a few years of launching.
“To everyone on the inside, it seems fake,” said Alex Cohen, co-founder and CEO of health AI startup Hello Patient. “You read the headline and you’re like, ‘I can’t believe it.'”
However, not all startups are comfortable expressing growth by reporting CARR rather than ARR. They like to keep their numbers clean and clear because they understand that the public market values software companies by ARR, not CARR. These founders prioritize transparency.
Wordsmith’s McNairn said he remembers startups struggling to justify high valuations after the 2022 market correction, but he doesn’t want to set an even higher bar by overstating his startup’s earnings.
“I think that’s short-sighted. If you do something like that for short-term gains, I think you’re overinflating an already abnormally high multiple,” he said. “I think this is very poor hygiene. It’s going to come back and bite you again.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
