Chegg Cut 81% of Its Staff. MasterClass Halved Its Own. Ask Where Their Traffic Used to Come From.

Chegg had 3,191 employees at its 2022 peak. It ended 2025 with 595.

That is an 81% collapse, delivered in two rounds eight months apart: 248 jobs in May 2025, then 388 more in October 2025. The 595 figure is Chegg's own, filed with the SEC as of December 31, 2025.

MasterClass tells a quieter version of the same story. Across multiple rounds of layoffs the company cut its workforce by more than 50%, from over 600 people to roughly 300. CEO David Rogier blamed "the worsening macroeconomic environment" for the first 20% cut in June 2022.

Guild is the third. It raised $175 million at a $4.4 billion valuation in June 2022, a round led by Wellington Management with Oprah Winfrey participating. Secondary market activity in 2024 implied a valuation between $1.46 billion and $1.98 billion, a 55-66% cut from that peak. Growth had cooled to 6%.

Three companies. Three different products. One shared dependency.

They weren't selling the same thing. They were buying traffic the same way.

Course Hero belongs on the list too: a $380 million Series C in December 2021 at a $3.6 billion valuation. In March 2023 it ran its first layoffs in 17 years, cutting 15% of staff.

What every one of these businesses had in common was not curriculum, price point or format. It was a distribution model that worked for two decades: publish free educational content, rank for "how do I do x," convert the traffic into paid enrollments.

It is the same model sold to every course creator on the internet. Write the blog post. Rank the blog post. Capture the email. Run the funnel.

5W's EdTech AI Visibility Index, released in June 2026, found that the legacy SEO winners — Chegg, Course Hero, Quizlet, Study.com — that dominated Google rankings for educational queries for a decade now sit outside the top 15 for AI citation share. Chegg ranks #19, Quizlet #17, Course Hero #21. The index tracked citations across 60+ prompts run through ChatGPT, Claude, Perplexity and Google AI Overviews in Q1 2026.

The brand at the top is Khanmigo, Khan Academy's AI tutor, at 18.5% citation share. A nonprofit.

5W founder Ronn Torossian described the category as one where "brands that have ranked first on Google for a decade are sometimes nowhere in the answer box."

Now the number nobody sends you

Here is where it stops being a story about edtech unicorns and starts being a story about your funnel.

Similarweb data shows citations appearing in US ChatGPT prompts rose from about 1.6% in June 2025 to roughly 6.8% by May 2026. Small, but quadrupling. The interesting part is the spread underneath the average.

Travel and hospitality prompts carry a citation about 23% of the time, the highest rate Similarweb tracked. Automotive sits around 20%. Professional Services sits under 4%.

Professional Services is where coaches, consultants, agency owners and most course sellers actually live. A separate read of the same Similarweb data puts education at 4.8%, with retail at 13.5% and sports at 10.7%.

Similarweb's explanation is blunt: categories where users compare specific, checkable options — a flight, a car, a hotel — get cited noticeably more often than categories built on general advice.

General advice is the entire product category of the online course industry. When a model answers "how do I price my coaching offer," it does not need to check anything. It answers from memory. No link. No click. No email capture.

One honesty note, because the data deserves it: Similarweb warns that citation rates pulled back for roughly four months after an October 2025 jump, and that any single month is as likely to be one data point in a noisy climb as a new normal.

Cited and recommended are two different businesses

The instinct is to chase citations. That may be the wrong target in a sub-4% category.

In low-citation categories a brand can influence answers frequently without ever appearing as a link. ChatGPT cites only about 15% of the pages it retrieves while researching an answer; the other 85% get evaluated and dropped. Those dropped pages still shaped the answer. They just didn't get credit for it.

Which reframes what a course creator's content is for. Not a traffic asset. A training input.

The channel comparison that matters: your own blog versus everywhere else. Bain data cited in a 2026 B2B analysis found 89% of citations for unbranded B2B questions come from third-party sources, not the brand's own website. Reddit, YouTube, LinkedIn, Wikipedia and Forbes rank among the most-cited sources across AI platforms, with ChatGPT specifically favoring Wikipedia, Reddit and editorial outlets.

The owned-media playbook that built Chegg is the one being devalued. The earned-media playbook nobody in the course world wanted to run — press, podcasts, YouTube, Reddit threads, being written about rather than writing — is the one feeding the answers.

The demand didn't go anywhere

Buyers still want to be told what to buy. G2 found that half of B2B software buyers now start research in an AI chatbot more often than Google — 29% in April 2025, 51% by March 2026.

And when AI does send someone, the traffic is good. AI-referred retail traffic converted 42% better than non-AI traffic in March 2026. A year earlier the same traffic converted worse.

Meanwhile the old channel hasn't died, it has been absorbed. Similarweb measured 494 million ChatGPT users worldwide between March and May 2026, and 95% of them also used Google, which is folding AI into its own interface through AI Overviews and AI Mode.

Chegg's second-quarter 2026 revenue was $51.8 million, down 51% year over year, with guidance pointing to $43-44 million in Q3. 5W's read on why: Chegg's growth engine was Google routing students to paywalled study pages at $19.95 a month, ChatGPT did the same thing free, and then AI Overviews captured the traffic before any click.

Nobody at Chegg shipped a worse product in 2025 than in 2021. The pipe closed.