Chegg has bought time

The old engine is shrinking faster than the new one grows

The academic library once spread the cost of an answer over many paying students. Its weakness is that the same question can now lead a student to a different interface or a free substitute. The annual filing identifies AI and search as risks without isolating their individual causal effects. The operating problem can be analyzed without pretending the revenue loss identifies which alternative caused it.

Academic Services revenue fell from $543.615 million in 2024 to $308.254 million in 2025, and Skilling also declined. That lost revenue is the funding problem I would size before underwriting the new businesses. Annual filing.

Managing the legacy for cash can be rational if investment follows its remaining economic life. Funding a new service from that shrinking stream is harder: the old business may contract before the new one produces sufficient contribution. The strategy therefore needs a payment schedule and a sequence of observable customer gains. A new growth label does not lengthen the funding runway. ^S5

A larger share can disguise a small gain

In Q2, Skilling rose from $17.217 million to $17.534 million while Academic Services fell from $87.903 million to $34.315 million. Skilling’s share rose from 16.4% to 33.8%, chiefly because the denominator contracted. The $0.317 million gain is the progress; the larger percentage share is mostly the damage around it. ^S2

The first quarter had shown 9% skilling growth, alongside management discussion of enterprise and channel partnerships. That was a better growth rate, but one quarter cannot carry the transition. I would follow incremental dollars and contribution through renewal before calling a changed mix a turnaround. ^S6

The successful endpoint need not recover the company’s former scale. A smaller operation can create value if delivery and shared costs fit the contribution it earns. That is a more plausible early hurdle than revenue replacement, but it is not an excuse to ignore declining dollars. The contribution and liquidity models below measure how much of the new business is needed under explicit cost assumptions.

Scroll horizontally to inspect the figure →

Which outcome does the payer buy?

Management described an employability service connecting job search, matching and coaching with its learning products. It reported more than 10,000 beta users and six new partners, including OpenSesame and Dale Carnegie, with launches planned for the second half of 2026. These are distribution and product-development milestones. The disclosure does not attach recurring collections or margins to them. ^S3

An employer buying better performance or retention and a learner seeking a new job have different objectives. Learning, coaching, matching and credential presentation can each support them, with different costs. Partner launches identify a route to a payer; collections and renewals determine the recurring economics. The useful commercial record connects the promised outcome to incremental collections, delivery cost, partner compensation and renewal.

Chegg’s survey methodology reports 200 responses from 1,408 invited graduates, about 14.2%. The 43% efficiency and 41% work-quality figures are respondents’ self-reports, excluding non-completers and a counterfactual. I’d use those experiences to design a paid-outcome test. ^S8

Less investment explains the free-cash-flow improvement

First-half free cash flow rose from $3.791 million to $9.461 million while operating cash flow fell from $19.686 million to $14.161 million. The reconciliation is simple: an $11.195 million reduction in cash capex more than offset the $5.525 million operating-cash-flow decline. Under Chegg’s definition, free cash flow is operating cash flow less cash purchases of property and equipment.

At the prior cash-capex level, H1 2026 operating cash would have left −$1.734 million of free cash flow. I read the improvement as a lower investment bill, with the right long-run budget still dependent on the smaller product portfolio.

The reduction may be sensible for a smaller company. Chegg capitalizes eligible content and internal-use software costs, however, so its capex includes investment in the digital product as well as equipment; the filing also describes asset abandonments as products were streamlined. Sustained low spending needs to support reliable delivery and retention without deferring a development bill that later returns. ^S1

A lower development budget can be exactly right for a smaller portfolio. Its sustainability appears later in delivery, retention and the cost of replacing aging content or systems. If those outcomes hold, less investment improves the transition. If they deteriorate, the cash saved was partly borrowed from the future. The current FCF bridge identifies the spending change but cannot decide between those interpretations.

Scroll horizontally to inspect the figure →

Positive cash flow matters, including the costs already paid

Chegg generated operating cash while restructuring. Starting from a $2.723 million GAAP net loss, depreciation, stock compensation and other adjustments produce $26.644 million before changes in operating assets and liabilities. Those changes consume $12.483 million, leaving $14.161 million of operating cash flow. After $4.700 million of cash capex, free cash flow is $9.461 million. ^S2

Management also disclosed $14.4 million of first-half severance payments. Paying those costs while keeping operating cash flow positive is meaningful progress. The payments are already in the reported cash flows; subtracting them again would make the same restructuring twice as expensive on paper. ^S3

Adjusted EBITDA and collections answer different funding questions. The operating-cash bridge already includes restructuring payments and the movement of liabilities, so those payments cannot be deducted a second time. Stock compensation is noncash in the bridge but still costly to owners. Following the complete cash schedule gives a better view of what can fund the next product commitment than selecting the most favorable earnings subtotal.

The stress model carries earnings, cash conversion and capex through successive steps. It can test a cost plan without pretending to reconstruct GAAP net income or Chegg’s internal budget.

Debt retirement leaves a reserve to fund

On September 1, Chegg confirmed repayment of its remaining convertible notes and no outstanding debt. Its approximately $38 million of cash and investments was explicitly June 30 pro forma for the repayment. It was not a newly observed September cash balance. ^S4

The arithmetic starts with $72.314 million of June cash and short-term investments and deducts $33.860 million of note principal. That gives $38.454 million before subsequent operations and other movements. The stress model uses management’s rounded $38 million reference so the same repayment never appears twice.

Investment maturities and sales helped increase the cash line during H1. They converted one liquid asset into another; they did not create operating value. Cash and investments together are the relevant funding pool.

Debt retirement removes a principal deadline; the zero-coupon notes supply no recurring coupon saving to this model. Leases, taxes and operating liabilities remain. The appropriate starting resource is cash and investments together, with the original pro forma repayment applied once. A later operating decision needs an updated balance and payment schedule before making a new commitment.

The next academic decline sets a steep replacement hurdle

Take the reported Q2 revenue bases and assume another 30% Academic Services decline in a comparable quarter. That removes $10.2945 million. Skilling would have to grow 58.7% from its $17.534 million base simply to hold total revenue flat.

Contribution replacement can be harder. Under the table’s assumed 70% academic and 45% skilling contribution margins, Skilling needs 91.3% growth to replace the contribution lost. More delivery-intensive revenue demands more growth for the same cash before shared costs.

The model assumes 70% and 45% contribution margins. Equal margins make revenue and contribution replacement coincide; efficient delivery or lower shared costs can reduce the growth burden. This is the strongest operating countercase to a simple replacement ratio: a smaller business can work. It wins only when the full remaining cost base fits the contribution actually collected.

The formula is revenue lost divided by the skilling base, multiplied by the academic-to-skilling margin ratio for contribution replacement. I would use that relationship to set an investment hurdle, then replace the assumed margins with observed cohort economics as they become available.

How much time does the reference cash buy?

The four-quarter stress test starts with the $38 million reference and the midpoints of August management guidance for Q3: $43.5 million of revenue and $1.5 million of adjusted EBITDA. Guidance calibrates the starting point; every subsequent model quarter is a hypothetical step. ^S3

I allocate that revenue to $26 million of Academic Services and $17.5 million of Skilling. At the same 70% and 45% assumed contribution margins, implied shared cash operating costs are $24.575 million: $26 million × 70% + $17.5 million × 45% − $1.5 million. The first quarter therefore reconciles before the operating paths diverge.

Each quarter deducts $0.5 million of cash-conversion drag and $3 million of cash capex. The first also pays an assumed $1.2 million of residual severance, roughly the June restructuring liability. These payment amounts and dates are author inputs; the first-half severance and September debt repayment receive no second deduction.

The paths omit seasonality, acquisitions, new borrowing, litigation recoveries, buybacks and investment-value changes. Repeating a weak-quarter calibration is deliberately demanding. Management’s expectation of meaningful second-half free cash flow remains its outlook; this model tests how the reserve behaves when declared operating conditions persist. It assigns no probabilities and forecasts no calendar-quarter balance.

Even the stronger transition consumes cash

The managed-decline path assumes 10% sequential academic declines, 5% skilling growth and 3% cash operating-cost reductions after the starting quarter. It ends with $25.4 million of liquidity, just $0.4 million above the assumed reserve floor, and a final-quarter earnings proxy around break-even.

Faster erosion overwhelms the same pace of cost cuts. With 20% academic declines and 2% skilling growth, liquidity crosses the $25 million floor in model quarter 3 and ends at $15.0 million. The floor implements an illustrative reserve policy.

The stronger transition combines 10% academic declines, 10% skilling growth and 5% cost reductions. Ending liquidity improves to $30.8 million. It still uses $7.2 million over four quarters under the common cash assumptions, which is why faster growth cannot be treated as an unlimited investment budget.

The sensitivities are immediate. Another $1 million of quarterly capex or conversion drag reduces four-quarter ending liquidity by $4 million. An $8 million project paid upfront reduces every later balance by $8 million before crediting any benefit. A plan that clears the reserve by $0.4 million has little room for that mistake.

Scroll horizontally to inspect the figure →

A new project has to earn its next tranche

A narrower employability intervention may deserve the first tranche. In Athey and Palikot’s randomized study, encouraging existing certificate earners to share credentials improved observed employment outcomes in the studied population, with heterogeneous effects. The intervention tested visibility after learning, not the causal return to taking a course. It supports investigating a specific information bottleneck before buying a larger content library.

The paper is not a Chegg forecast: it concerns another platform, selected learners and employment observed through professional profiles. Its useful contribution is a mechanism to test. A payer might value better application of skills, clearer evidence of them or help reaching a suitable role. Each route needs its own cohort and contribution record, including non-completion, human support, curriculum, platform and partner costs. A cheap intervention with a real outcome could improve capital productivity; a broad service without paid renewal would not.

The $38 million reference leaves $13 million above the assumed $25 million floor, so an $8 million upfront project uses 61.5% of the initial headroom and pulls even the stronger-transition ending balance from $30.8 million to $22.8 million before crediting any new benefit. I’d stage that commitment.

Eventual project value cannot pay the next bill. Against this transition’s funding constraints, the first commitment should test which outcome changes and whether someone pays enough for it. Subsequent collections and delivered contribution can justify a larger stage. Smaller scope, later payments or outside funding are ways to align the cash need with the benefit; none should be assumed available without terms.

The opportunity cost is the same transition reserve. A new project and a repurchase can each consume the funds needed to survive an academic miss. Authorization supplies no financing. Capital should go first to observable customer economics and to the product maintenance needed to keep current collections reliable, with the next tranche dependent on what the first demonstrates.

What would change my judgment

A current post-repayment cash schedule would replace the June pro forma reference. The operating evidence would be absolute skilling contribution, renewing paid cohorts and reliable delivery at a sustainable development budget. Outcomes need the right denominator: everyone enrolled or served, who completed, who responded, who paid and who renewed. A favorable graduate story and a viable paid cohort are different pieces of the transition.

The important unknowns remain contribution by service, retention, partner conversion and the development needed to maintain quality. Better evidence could make a smaller Chegg more attractive even while consolidated revenue declines. It could also show that the new service consumes cash too quickly to finance from the legacy. The positive case earns its next commitment when an identifiable customer outcome leaves enough cash to fund the next one.

The July 24 NYSE notice and shareholder-approved reverse-split flexibility keep their original status; later compliance is not inferred. ^{S7 S9} A changed share unit cannot supply transition cash. The recovery will be judged by the funded customer business and value per original share, with the public financial record independent of my disclosed work at Chegg.

The recovery still has to earn its funding

I would become more confident when Skilling’s recurring contribution funds delivery and development as the legacy declines. Debt retirement and positive cash flow buy the chance to do that. The 0.6% replacement figure keeps the next investment tied to the customer economics still to be earned.

Assumptions

• USD millions unless stated; cash-capex and free-cash-flow definitions follow Chegg’s reconciliations.
• The four-quarter model is an explicit stress test with no seasonality, not a calendar forecast.
• Contribution margins, cost trajectories and reserve floor are author assumptions; no double-counted debt repayment or severance.