Twilio's Growth Fell From 48% to 22% in Four Quarters as Customers Simply Sent Fewer Messages

A customer runs a two-factor login for its app on Twilio. Every text message it fires off costs a fraction of a cent, and in a good year the app sends billions of them. Then the economy tightens, the customer trims its marketing blasts, throttles its notification volume, sunsets a feature — and Twilio's revenue from that account falls the same week the usage does. No renewal date to wait for, no minimum commitment to burn through. The meter simply slows. Multiply that across more than 280,000 active customer accounts² and you get the defining number of Twilio's 2022: revenue growth that stepped down, quarter after quarter, in near-lockstep with what its customers stopped spending.

The official story is that Twilio hit a growth cliff in 2022 — a sudden fall from grace. That is not what the filings show. What they show is four discrete steps down, each one a fresh reading of a meter that customers were turning down. Twilio's year-over-year revenue growth went 48%, then 41%, then 33%, then 22% across the four quarters of 2022.¹²⁴ Not a cliff. A staircase — and the company was walking down it one usage cut at a time.

The meter is a feature until it runs backwards: consumption pricing bills exactly what customers use, which is glorious on the way up and merciless on the way down

Twilio's whole appeal was that you paid only for what you used. No seat licenses, no annual platform fee waiting to be justified — you sent a message, you paid for the message. In a boom, that pricing model is a rocket: as customers scale, Twilio scales with them automatically, and its Dollar-Based Net Expansion Rate — the money existing customers spend this year versus last — reads well above 100% without a single new deal signed. In the year ended March 2022, that expansion rate was still 127%.¹ Every existing account was, on average, a growing account. That is the magic of the meter: growth comes to you.

But the meter has no memory and no floor. A subscription company that sells annual seats has a moat against a downturn built into its billing: a customer who wants to spend less has to wait for the contract to lapse, and even then the seats churn out slowly. Twilio's customers didn't have to wait for anything. When budgets tightened, they simply sent fewer messages — and the revenue impact landed the same quarter, unbuffered. This is the trade Twilio made when it chose pure usage pricing over layered subscription minimums: it accepted a top line wired directly to customer behavior, in both directions. The staircase down in 2022 was that wiring, working exactly as designed.

Why one round of layoffs was never going to be enough: a cost cut sized against a slowdown that keeps sliding gets outrun by the slide

In September 2022, Twilio cut 11% of its workforce, expecting $70 million to $90 million in related charges. CEO Jeff Lawson told employees the company had 'grown at an astonishing rate over the past couple years,' that 'it was too fast, and without enough focus,' and took 'responsibility for those decisions.'⁶ Notice what he did not say: he did not blame the pricing model or the macro. He framed it as an execution problem the company owned. But here is the trouble with sizing a layoff to a slowdown that is still sliding — you cut against the growth rate you can see, and by the time the charges clear, the meter has ticked down again.

The September cut was calibrated against a business still growing 33% in Q3.³ By the time Q4 landed at 22%,⁴ the cost base was again too heavy for the revenue trajectory beneath it. So in February 2023 — just five months later — Twilio announced a second round, cutting roughly 17% of its workforce, about 1,500 people, against the 8,992 employees it had reported to the SEC as of September 30, 2022.⁷ Two corrections, not one clean one. That is what it looks like when your revenue is chasing your customers' usage down a staircase: every step you size your costs to has another step below it you haven't seen yet.

Wasn't this just the macro — the same thing that hit everyone?: the downturn was real, but the pricing model chose how directly and how early it arrived

The fair objection is that 2022 was brutal for the whole software sector, and blaming Twilio's pricing model is too tidy — everyone slowed, usage-priced or not. True. The macro was the shock. But the pricing model was the transmission line: it decided how fast and how directly the shock reached the top line. A seat-based peer felt the same tightening customers, but its contracts absorbed the first blow; Twilio's meter did not. And the deceleration was arguably sharper than the headline let on — Q3's 33% growth included $34.8 million from the Zipwhip acquisition, and Twilio's own organic figure was a point lower at 32%.³ Strip out M&A and the usage decay was cleaner and steeper.

The honest counter runs the other way too: this wasn't purely a 2022 event. Twilio's expansion engine had been cooling for years. The Dollar-Based Net Expansion Rate was 141% in Q2 2019 and had already fallen to 132% by Q2 2020⁸ — long before the downturn is popularly dated. So 2022 was a continuation, not a first crack. The pricing model didn't cause the slowdown. It made sure Twilio felt every degree of it, immediately, in the number Wall Street watches most.

Twilio's meter did in 2022 exactly what it did in every good year before it: it reported, faithfully and instantly, what customers were actually doing. On the way up, that honesty looked like genius — growth arriving on its own. On the way down, the same honesty looked like a staircase, each step a quarter, each quarter a fresh reading of demand that had cooled since the last one. The company crossed $1 billion in quarterly revenue for the first time in that final quarter of 2022⁴ — and grew the slowest it ever had while doing it. That is the deal you sign when you charge for what people use: you never have to guess how your customers feel about the economy. They tell you, in the only language a meter speaks, the same week they decide.