When Should a Service Business Stop Taking Clients? Basecamp's Product Crossover
Short answer: Stop taking new client work when product contribution, not just revenue, can cover essential monthly costs for at least six months, no single product customer controls the result, and retention survives without founder-led onboarding. Reduce services in stages before ending them. Keep enough cash to withstand a 25% product-revenue fall and the final costs of completing or handing over client commitments.
Client work finances product development, but it also claims the exact people and attention needed to improve the product. The decision is not whether product revenue has become exciting. It is whether the product can carry the company through an ordinary bad month.
Basecamp grew from a project-management system that the 37signals design firm built for its own client work. The company says prototypes began in July 2003, David Heinemeier Hansson worked around 10 hours a week, and the product launched in February 2004. Basecamp's account appears in Shape Up. Its company history says product revenue exceeded web-design revenue roughly a year after release, after which the firm stopped design work. Basecamp describes the crossover here.
Public accounts do not provide complete monthly contribution, churn or cash figures at the decision date. Those gaps matter. Revenue crossover alone is not a safe rule for another founder.
Use the Services-to-Product Crossover Gate
The Services-to-Product Crossover Gate opens only when five conditions hold together.
| Gate | Services evidence | Product evidence | Pass condition | |---|---|---|---| | Contribution | Margin after delivery labour | Margin after support and infrastructure | Product covers essential fixed costs | | Durability | Signed backlog and concentration | Retention and recurring use | Six-month downside remains solvent | | Capacity | Hours consumed by commitments | Hours needed for product growth | Released time has a defined use | | Concentration | Revenue by client | Revenue by account and channel | No single loss breaks the plan | | Exit cost | Completion, refunds and handover | Migration or support obligations | Fully funded before stopping |
My position is that you should not stop services merely when product revenue equals service revenue. Products often have lower marginal delivery cost, but early revenue can be fragile and support-heavy. Switch when product contribution covers the operating floor and the downside cash forecast survives. This is later than the emotional crossover and earlier than waiting for zero risk.
Compare contribution, not headline revenue
Separate direct delivery labour from general overhead. A £12,000 client project requiring £8,000 of staff and contractor time does not equal £12,000 of product subscriptions with £2,000 of hosting and support. Equally, founder support cannot be treated as free.
Calculate monthly contribution for each side:
services contribution = service revenue minus delivery labour, contractors and project-specific costs
product contribution = recurring product revenue minus infrastructure, payment, support and customer-specific delivery
Then identify essential fixed costs: minimum salaries or drawings, rent, insurance, software, tax provision and legally committed spending. The product must cover these under a realistic downside, not only in the latest record month.
Use [gross profit versus net profit](/business-finance/gross-profit-versus-net-profit/) to keep contribution and overall profit distinct.
Test durability before releasing the bridge
Client revenue may be lumpy but contracted. Product revenue may recur monthly but churn quickly. Examine at least three cohorts and six months where possible. Track opening recurring revenue, expansions, contractions, cancellations and unpaid accounts.
If the product is younger, do not invent retention. Continue a reduced service base until enough time passes or keep extra runway. One large annual prepayment should be spread across the service period for decision purposes, even if the cash is already in the bank.
Stress the product case at 75% of current revenue and 120% of support cost. If essential outgoings remain covered for six months, the gate is stronger. Six months is a working guide. A business with long sales cycles or high regulatory obligations may need more.
Reduce service work in steps
Do not abandon existing clients mid-project. Stop accepting low-contribution or distracting work first, raise the minimum project size, shorten the service menu and complete commitments. This releases capacity while preserving some cash.
A practical sequence is:
- Reject new work outside the most profitable service.
- Move from bespoke proposals to one defined package.
- Limit delivery days or client slots.
- Hand over suitable work to trusted providers with consent.
- Stop new engagements after the product passes the full gate.
Customer contracts, employment, data, tax and professional obligations vary by jurisdiction. Review notice, completion, refunds, intellectual property and handover with qualified advisers before ending services.
Worked example: StudioFlow Projects
StudioFlow is a three-person web studio with a project-management product. Monthly figures are:
| Measure | Client services | Product | |---|---:|---:| | Revenue | £14,000 | £11,200 | | Direct labour and variable cost | £8,400 | £2,300 | | Contribution | £5,600 | £8,900 |
Essential fixed costs are £7,100 a month. Product contribution currently covers them by £8,900 - £7,100 = £1,800. At 75% of product revenue, revenue falls to £11,200 x 75% = £8,400. Assume variable product cost falls only to £2,000 because support remains high. Downside contribution is £8,400 - £2,000 = £6,400, which is £700 below fixed costs.
The company holds £36,000 unrestricted cash. A six-month downside deficit is 6 x £700 = £4,200. Finishing client projects and refund risk require £9,000. After reserving both, cash remaining is £36,000 - £4,200 - £9,000 = £22,800.
StudioFlow could stop services and survive the modelled downside, but product contribution does not yet cover the operating floor on its own. The safer move is to keep one profitable client slot worth at least £1,500 monthly contribution for three more months while improving retention and support. These figures are illustrative, not agency or SaaS benchmarks.
Decide what the released time will buy
Stopping services creates capacity, not results. Specify the work that client delivery currently blocks and the expected commercial mechanism. Examples include reducing activation failure, improving retention, completing a self-service purchase route or contacting a defined segment.
Avoid replacing client meetings with undirected feature building. Assign weekly measures such as activation rate, support minutes per account and paid conversion. If released time does not improve the product economics after a defined period, reconsider the plan.
The wider aim is a [business that does not depend entirely on you](/operations/business-that-doesnt-depend-on-you/), not simply exchanging one form of founder dependence for another.
Set a re-entry rule
A service firm can pause new work without destroying the capability. Define when you would reopen limited services, such as product contribution falling below 90% of essential costs for two months or runway dropping below nine months.
Re-entry is not shame. It is a cash-control option. Price it high enough to justify the interruption and keep it within the product's target market where possible, so client knowledge strengthens rather than distracts from the product.
Make the crossover decision in 30 days
In week one, separate contribution for services and product. In week two, build the six-month downside and list every exit obligation. In week three, rank client work by contribution and strategic fit. In week four, stop the weakest new work and redirect named hours to one product constraint. Review monthly. Close the service gate only after product contribution, runway, retention, concentration and exit funding all pass together.
Related guides
Frequently asked questions
Is recurring revenue always safer than project revenue?
No. Recurring billing is only durable when customers continue receiving value and can pay. Early product revenue may churn quickly, depend on one acquisition channel or require hidden founder support. Project revenue can be contracted and paid in advance, although it is less predictable after the backlog ends. Compare retention, concentration, collection timing and contribution. Spread annual subscriptions across their service period when judging sustainability. The word "recurring" describes the billing arrangement, not the certainty of future cash. Model a missed renewal as well.
Should I tell clients that I am building a product?
Yes when it affects delivery, creates a conflict, uses their information or may change the relationship. Be clear about priorities and contractual commitments. Client problems can inform a product, but do not reuse confidential material, intellectual property or data without permission. Some clients may become early users, while others may reasonably worry about reduced attention. Do not promise indefinite service merely to avoid an uncomfortable discussion. Review confidentiality, ownership and conflict terms with qualified legal advice before commercialising work developed during client engagements.
What if one client funds most of the product?
Treat that as concentration risk, not product validation. The client may be a valuable bridge, but its requirements can turn the product into bespoke software and its departure can remove runway overnight. Separate what it pays for services from what independent customers pay for the product. Model the business without that client and cap custom development. Before stopping other work, build a reserve or diversify contribution. If the client owns relevant intellectual property or receives exclusivity, obtain qualified advice before selling the product more widely.
Can I sell the service business instead of closing it?
Possibly, if revenue, team, processes, contracts and client relationships can transfer without depending on you. A buyer will examine concentration, profitability, retention and change-of-control terms. Start by documenting delivery and reducing founder-only knowledge. Do not assume the sale will complete or fund the product until terms are signed and cash conditions are clear. Tax, employment, data and contractual consequences can be material, so use qualified legal, tax and corporate-finance advisers. A managed wind-down may be more valuable than a rushed sale.
How much runway should I have before switching?
Six months after funding exit obligations is a bare working minimum for a product with proven retention and short sales cycles. Nine to twelve months is safer when revenue is young, customers are concentrated or product improvements take time. Calculate runway under a downside case, not current growth. Include tax, refunds, annual bills and founder living needs. More runway does not repair weak economics, but insufficient runway forces short-term decisions. Build a [13-week cash-flow forecast](/business-finance/build-a-13-week-cash-flow-forecast/) alongside the longer product model.
Should I hire someone to keep services running?
Only if the service contribution after their full employment or contractor cost remains attractive and quality can be transferred. Document the process, test delegation and include management time. Hiring solely to avoid making the crossover decision can create a second team and more fixed cost. A defined, profitable service package may finance the product well; bespoke founder-led work usually does not transfer cleanly. Employment status, consultation, supervision and client-consent requirements vary, so obtain qualified local advice before changing who performs contracted work.
Comments (0)