Usage Pricing Can Liquidate the Customer Base
Executive summary
Harvest's new pricing model keeps a seat charge and adds usage fees tied to active projects, clients, tasks and invoicing. The company describes that structure as flexible: customers pay for the activity they use. Recent customer reports show a less comfortable result. Some long-tenured accounts have received renewal quotes several times their previous bills, and Harvest's own help center illustrates how a Flex bill can rise from $44 for seats to $404 after usage is added.
The mainstream interpretation is that this is an aggressive SaaS price increase. The deeper issue is what the meter values. Projects, clients, tasks and invoices are not merely units of computing consumption. They are accumulated business structure, much of it created over years. Pricing those records can turn customer tenure and migration difficulty into a billable input.
That is especially consequential under Bending Spoons, which acquired Harvest in 2025. Its June prospectus describes an acquisition model built around established customer bases, high return hurdles and deep monetization changes. It also reports that 48% of subscription revenue across its portfolio came from customers at least five years old, while net revenue retention remained below 100%. The contrarian thesis is that abrupt monetization can liquidate the very customer-base asset the acquisition model depends on. A large first-year revenue lift can still destroy long-duration value if it converts switching costs into switching projects.
Market context
Harvest is a mature work system for time tracking, project reporting and invoicing. The product says it is trusted by more than 70,000 businesses. For many customers, the relationship is not a lightweight monthly app subscription. Years of projects, client names, billing history, task structures, integrations and staff habits sit inside the account.
Its current pricing page lists Teams from $9 per seat per month on an annual plan, or $11 monthly, and Enterprise from $14 annually or $17.50 monthly. Those visible prices are starting points. Harvest says additional invoices, projects, clients and tasks are billed according to use.
The help center makes the mechanism clearer. Teams and Enterprise can be billed as Flex, which combines a seat rate with usage charges based on the prior period's active projects, active clients, active tasks and invoicing, or Unlimited, which adds a fixed usage fee. Usage charges begin at the second paid renewal for new accounts. Harvest gives the example of a bill rising from $44 to $404 when usage appears after a seats-only first invoice.
Recent reporting gives the change a sharper edge. Subscription Insider, citing BBC reporting published August 20, says UK consultancy Salentis saw an average monthly bill move from about $130 to $2,110. After it challenged the quote, Harvest offered $1,309 a month for one year if paid upfront. A U.S. customer told the BBC its annual bill moved from $2,800 to $23,000. These are individual accounts, not a representative sample, and Harvest does not publish a standard Unlimited fee.
Bending Spoons' June 30 prospectus provides the strategic context. The company says its playbook is to acquire digital businesses, transform and optimize them, and reinvest the cash into more acquisitions. It applied acquisition hurdles of 65% levered and 25% unlevered internal rates of return to nearly all acquisitions completed from 2023 through the first quarter of 2026. Following the Harvest purchase, it redeployed about half of the team previously dedicated to Evernote to accelerate Harvest's transformation.
Findings
Finding 1
The usage meter prices customer structure, not only marginal service cost.
Usage pricing is economically clean when the bill tracks a costly unit delivered: storage consumed, messages sent, transactions processed or compute used. The customer can connect the charge to volume and, ideally, to value.
Harvest's meters are different. An active client, project or task is partly a record of how the customer organizes work. An invoice is an output of the customer's business. Those objects can correlate with value, but they do not establish equivalent marginal cost to Harvest. The model therefore resembles value-based pricing layered onto a system of record.
That can be legitimate. Vendors are not required to price at cost, and a customer managing hundreds of active engagements may receive more value than one managing ten. The operating problem is predictability. The public page does not display standard unit rates for every usage dimension or a standard Unlimited fee, so an outside buyer cannot reconstruct the full bill from the visible seat price. Existing customers must use account-level simulation tools.
| Billing element | Public description | Economic object being priced | Operator risk |
|---|---|---|---|
| Seat | $9-$17.50 per seat per month, depending on plan and billing cycle | User access | Familiar and forecastable |
| Active project | Usage above included allowance | Customer workflow structure | Activity cleanup changes the bill |
| Active client | Usage above included allowance | Breadth of customer book | The vendor participates in portfolio complexity |
| Active task | Usage above included allowance | Process granularity | Detailed use can become a penalty |
| Invoicing | Usage above included allowance | Customer commercial output | Bill can scale with monetized activity |
| Unlimited | Seat charge plus fixed usage fee | Account-wide right to avoid metering | Full public comparison is unavailable |
Source: Harvest pricing page and help center, accessed August 24, 2026. Units: U.S. dollars per seat per month and qualitative billing dimensions. Method: direct transcription and Blackrock Research classification of the object being priced. Limitations: included allowances and account-level quotes can vary; the table does not estimate Harvest's marginal costs or claim that every customer faces an increase.
The strategic danger is that the meter can feel retroactive. A customer may have created years of projects and tasks under a seat-based bargain, then discover that the accumulated operating structure affects the renewal price. Even if only current-period active items count, changing the unit after the workflow is embedded alters the meaning of the installed base.
Finding 2
Large renewal increases can make migration economically rational.
Switching costs create pricing power only within a range. Data migration, staff retraining, workflow redesign and integration repair all discourage churn. But a sufficiently large price increase turns those same costs into a one-time project with a measurable payback.
The reported Harvest cases illustrate the threshold effect. Salentis' quoted monthly bill rose by about 1,523%, from $130 to $2,110. The revised $1,309 offer was still roughly 907% above the prior level. The reported U.S. account rose about 721%, from $2,800 to $23,000 a year. Those percentages are Blackrock Research calculations from reported nominal bills; currency units should not be compared across customers.
| Reported account | Prior bill | New or revised bill | Calculated increase | Evidence status |
|---|---|---|---|---|
| Salentis, initial quote | $130/month | $2,110/month | 1,523% | Customer report cited by BBC and Subscription Insider |
| Salentis, revised one-year offer | $130/month | $1,309/month | 907% | Customer report cited by BBC and Subscription Insider |
| U.S. customer | $2,800/year | $23,000/year | 721% | Customer report cited by BBC and Subscription Insider |
| Harvest help-center example | $44 billing period | $404 billing period | 818% | Company-provided illustration of seats plus later usage |
Source: Harvest help center and Subscription Insider's August 22 report citing BBC reporting from August 20. Period: customer renewal experiences reported in 2026; Harvest illustration current on August 24. Units: reported U.S. dollars and percent change calculated as (new-prior)/prior. Limitations: the customer cases are not audited, representative or directly comparable; the help-center example illustrates mechanics rather than a legacy renewal; discounts, plan features, usage and billing periods can differ.
The proper test is not whether the increase covers migration cost in month one. It is whether the present value of staying exceeds the cost of moving. A customer quoted an extra $20,000 a year can justify a meaningful implementation budget. Competitors can subsidize import, preserve familiar data structures and turn the incumbent's price notice into their acquisition event.
Finding 3
The acquisition model depends on long tenure, but monetization can shorten it.
Bending Spoons' portfolio economics make the tension visible. The company has grown rapidly through acquisitions, with revenue rising from $387 million in 2023 to $1.31 billion in 2025. Yet its reported net revenue retention was below 100% in each disclosed period: 93% in 2023, 91% in 2024, 95% in 2025 and 94% in the first quarter of 2026. Bending Spoons notes that NRR can fluctuate significantly, often because of monetization initiatives.
At the same time, the portfolio is unusually dependent on mature relationships. In the first quarter, 48% of subscription revenue came from customers with at least five years of tenure, including 28% from customers with at least ten years. Revenue-weighted average subscriber tenure was 8.0 years.
| Portfolio measure | 2023 | 2024 | 2025 | Q1 2026 |
|---|---|---|---|---|
| Net revenue retention | 93% | 91% | 95% | 94% |
| Revenue | $387m | $671m | $1.31bn | $601m |
| Subscription revenue from customers with tenure >=5 years | - | - | - | 48% |
| Subscription revenue from customers with tenure >=10 years | - | - | - | 28% |
| Revenue-weighted average subscriber tenure | - | - | - | 8.0 years |
Source: Bending Spoons prospectus dated June 30, 2026. Period: calendar years 2023-2025 and quarter ended March 31, 2026. Units: percent, U.S. dollars and years. Method: direct transcription; Q1 revenue is quarterly and should not be compared with full-year totals as if it were annual. Limitations: metrics cover the consolidated portfolio, not Harvest alone; pre-acquisition customer data are estimated for the NRR calculation; acquisition mix drives a significant share of revenue growth.
This is the contrarian point. Long tenure is not merely proof of loyalty. It is an asset whose value depends on the expected duration of future cash flow. Raising price can improve near-term revenue from survivors while reducing the useful life of the cohort. If migration pain was the main reason a customer stayed, an extreme renewal can finance the search for an exit.
Implications for operators
Subscription operators should price installed-base changes against customer lifetime value, not first-year uplift. Model price realization, contraction, cancellation, migration assistance and support demand by tenure and account complexity. A ten-year customer deserves a longer observation window because the foregone cash flows are larger.
Usage units should be explainable before purchase. Publish the meter, allowances, reset rules and representative bills. Give customers exportable usage histories and alerts before they cross thresholds. A calculator hidden behind an authenticated account may help an administrator, but it does not create market-wide price comparability.
Acquirers should treat retention capacity as finite. Cost reduction, product investment and monetization may all be valid parts of a transformation, but simultaneous shocks can exhaust the goodwill embedded in an acquired brand. Sequence changes and preserve a credible lower-cost path for customers whose use does not require the premium feature set.
Competitors should prepare for price-led migration windows. Build importers, reconciliation reports, parallel-run tools and contract credits before a rival's renewal cycle peaks. The most valuable sales message is not a lower sticker price; it is a bounded switching plan with verified data completeness.
Customers should calculate the option value of portability before the notice arrives. Maintain regular exports, document integrations, know retention requirements and estimate migration effort. A system of record is less dangerous when leaving it is a rehearsed process rather than an emergency project.
Risks & open questions
Public evidence cannot establish Harvest-wide churn, realized average price or account-level profitability. The reported increases may be concentrated among unusual configurations, and the company may retain most customers through Flex optimization, Unlimited discounts or negotiation. Higher prices could also fund product improvements that increase value.
The thesis would weaken if Harvest's cohort data show stable logo retention, net revenue retention above 100%, low support escalation and sustained product usage after the renewals. It would also weaken if customers reduce active objects without operational harm or consolidate other software spend enough to offset the higher bill.
The opposite evidence would be a wave of exports, downgrades and cancellations among long-tenured accounts, especially where competitors publicize migration programs. Bending Spoons' future reporting may not disclose Harvest separately, so independent measures such as app usage, web traffic, customer reviews and competitor imports will remain imperfect proxies.
There is also a fairness question that public pricing alone cannot answer: whether the metered dimensions accurately track customer value across business models. A consultancy with many small projects can generate less economic value than a firm with a few large ones. A count-based meter can therefore tax workflow shape rather than willingness to pay.
Appendix / methodology notes
This report reviewed Harvest's public pricing and help-center documentation available August 24, 2026; Bending Spoons' June 30 prospectus; Subscription Insider's August 22 account of BBC reporting; and the BBC URL identified by that article. Harvest's plan mechanics and Bending Spoons' portfolio metrics are primary-source facts. Customer renewal figures are attributed reports and are not treated as representative market data.
Percentage increases were calculated from the cited old and new nominal bills without adjustment for taxes, discounts, currencies, added features or account usage. The chart-ready tables should be read as a pricing-mechanism and case-study comparison, not a distribution of Harvest customer outcomes.
A conclusive follow-up would require Harvest-level cohorts segmented by legacy plan, tenure, seat count and activity: quoted and realized price change, logo retention, gross and net revenue retention, contraction, support contacts, feature adoption, data exports and cancellation reason at 30, 90, 180 and 365 days after renewal. That dataset is not public.