Nerdy, which trades as Varsity Tutors, is the largest listed consumer tutoring marketplace in the United States. In its 2025 accounts it spent 33.6% of revenue on sales and marketing.
That is the best number in the company's history. It has never spent less.
It is also the third consecutive year the company lost money and the second consecutive year revenue fell. In February 2026 it told investors it was winding down its schools business to concentrate on consumers, and guided the coming year to roughly breakeven on flat-to-lower revenue.
A company gets to that position by doing everything the playbook says. What the filings show is that the playbook has a ceiling, and Nerdy has now found it.
Seven years of the same shape
Everything below comes from Nerdy's own annual filings.
| FY | Revenue | Gross margin | S&M as % of revenue | GAAP net income |
|---|---|---|---|---|
| 2019 | $90.5m | 65.9% | 42.0% | — |
| 2021 | $140.7m | 66.8% | 46.5% | −$3.8m |
| 2022 | $162.7m | 69.4% | 45.6% | −$35.4m |
| 2023 | $193.4m | 70.6% | 35.4% | −$40.2m |
| 2024 | $190.2m | 67.5% | 37.7% | −$42.6m |
| 2025 | $179.0m | 58.0% | 33.6% | −$39.9m |
Cumulative net loss from 2021 to 2025: −$161.9 million.
Read the two middle columns together, because that is where the argument lives. Sales spend fell from 46.5% to 33.6% of revenue. That looks like discipline, and some of it was. But the ratio only improved after revenue stopped growing, and gross margin fell to 58.0% in the same year, its lowest in the series. The company got more efficient at buying customers at exactly the point it had fewer customers to serve and less margin on each of them.
There is no year in this table where volume fixed the distribution cost. That is the finding.
Four things cap the model, and all four are visible in the filings
Distribution cost never amortises. Seven straight years between 33.6% and 46.5% of revenue on sales and marketing. In a business where acquisition is a fixed cost of doing business rather than an investment in a durable base, the line does not come down as you grow. It comes down when you shrink.
Gross margin is capped by human supply. Nerdy's gross margin has run between 58.0% and 70.6%. The ceiling is set by what tutors are paid, and Nerdy has been increasing its investment in tutor pay and incentives to hold retention. You cannot expand margin and improve supply quality at the same time. Every point of margin you take comes out of the person delivering the lesson, and they can leave.
Demand is episodic. Tutoring is bought for an exam, a term, or a crisis. Nerdy's answer was to convert transactions into subscriptions — higher-frequency "Learning Memberships." It worked on the metric it was aimed at. In the fourth quarter of 2025, membership revenue was $41.6 million, or 85% of the total, from 33.2 thousand active members, with average revenue per member at $364, up 21% year over year. Revenue per member rose by a fifth. Total revenue still fell for the year. The subscription did not create demand that was not there. It extracted more from the demand that was.
The institutional escape hatch depends on somebody else's budget. This is the part worth reading twice. Nerdy had a funded, evidence-backed high-dosage tutoring product for schools. In the fourth quarter of 2025 that segment signed 56 contracts worth $4.1 million of bookings, down 11% year over year, which the company attributed to "federal and state funding delays and the related impact to high-dosage tutoring contracting and program start dates."
So it left. Nerdy is winding down Varsity Tutors for Schools and exiting First Tutors, cutting about $11 million of annual fixed cost, in order to "concentrate the company's capital, product development and management attention on its consumer learning business."
That is a public company saying, in a filing, that the institutional market it built for could not be converted on a schedule it could finance.
What the good quarter actually says
Nerdy's fourth quarter was its best in two years, and the detail is worth having.
Revenue was $49.1 million, up 2% and the first growth since the second quarter of 2024. Adjusted earnings before interest, tax, depreciation and amortisation came in at positive $1.3 million, the company's first positive quarter on that measure. Gross margin, excluding a one-off software write-off, was 66.8%. Cash stood at $47.9 million. The net loss for the quarter was still $14.0 million.
Then the guidance: revenue of $180 to $190 million for 2026 and approximately breakeven adjusted earnings.
Put plainly, a company with roughly $179 million of revenue reaches breakeven by exiting a segment and holding revenue flat. The model works. It works at a size that does not return a venture fund.
Europe ran the same experiment
GoStudent was Europe's best-funded tutoring marketplace, and its numbers are a compressed version of the same story.
| Date | Valuation | Prosus carrying value |
|---|---|---|
| 2021 | €1.4bn | — |
| Early 2022 | €3.0bn | $226m |
| 2023 | €2.1bn | $160m |
| June 2024 | €903m ($969m) | $68m |
Losses went from €89 million in 2021 to €221 million in 2022. Three rounds of layoffs between 2022 and 2024 removed roughly half of about 2,000 employees. The company had raised €675 million since 2016, almost a third of it from Prosus. It announced profitability in 2024, after the cuts.
Two companies, two continents, the same resolution. Both reached profitability by becoming smaller. Neither reached it by growing into its cost structure.
Why this lands harder in this region
Across nine listed education companies, only two ever spent under a fifth of revenue on sales and marketing, and those two are the only ones that made money. Duolingo spent 12.1% in 2025 and earned $182.4 million before tax. At the other end, 2U spent 39.3% in 2023 and lost $317.6 million in that year alone. Coursera spent 33.8% in 2025 and lost $51.0 million; Udemy spent 41.3% and made $3.8 million.
Nerdy sits at 33.6%, on the wrong side of that line, in the most capital-hungry category on it.
Consumer tutoring is also one of the most-attempted models in this region. When MAGNiTT last published a ranking of the region's most-funded EdTech companies, in November 2021, the top two were Abwaab in Jordan at more than $27 million raised and Noon Academy in Saudi Arabia at more than $21 million — both social-learning and tutoring businesses. That was the shape of the regional category then, and much of it still is.
The difference is the financing available to sustain it. In the twenty-four months to August 2026, exactly one disclosed MENA EdTech round exceeded $10 million: ULA's $28 million Series B. Every other disclosed round came in at $7.4 million or below. Nerdy funded seven years of forty-per-cent sales ratios through public markets. GoStudent funded its version with €675 million of private capital. Neither of those options is on the table for a tutoring company in Amman, Cairo or Riyadh.
Consolidation has already started here, from the other direction. Abwaab bought Egypt's Apex Education in January 2026, at an undisclosed price. Nerdy and GoStudent arrived at contraction after raising hundreds of millions. Regional companies are being asked to arrive there first.
What a founder should do about it
Put sales and marketing as a share of revenue on the first page of every board pack. Not customer acquisition cost, which you control the definition of. The ratio, from the accounts, next to the same ratio for the four public comparables above.
If a human delivers the lesson, model 58 to 70 per cent gross margin and defend anything higher. Nerdy's ceiling is the category's ceiling, and it fell as the business matured, not as it grew.
Test whether your demand is episodic before you price a subscription against it. Nerdy's membership push raised revenue per member 21% and did not stop revenue falling. If customers buy for an exam, a subscription changes the billing, not the behaviour.
Do not treat schools as the fix for consumer economics. The company with the best-resourced attempt at exactly that has just written the wind-down into a filing, and the reason it gave was the budget cycle, not the product.
Ask what the business looks like at $10 million of revenue. Nerdy reaches breakeven at roughly $179 million after eliminating a segment. Work out where your version of that point is, and then ask whether the capital available in this region gets you there.
Nerdy's 33.6% is the number a founder should carry out of this. It is the best distribution ratio a consumer tutoring marketplace has ever posted at scale, it took seven years and a shrinking business to reach, and it still is not low enough.