Before deciding whether to charge by hour, project or month, explain what changes for the client when they hire you. The billing unit organises the relationship, but cannot answer that question alone. A service can carry a monthly price while remaining a sequence of tasks with little connection to the buyer's objective.
When can hourly billing make sense?
Hourly billing can make sense when effort varies, the intervention is still being defined or the client needs specialist capacity for a period. Define what counts as work, how time consumption is tracked and when additional time needs approval. The client should understand the purpose of the hours purchased and how they support a decision.
In a hypothetical example, a company requests help investigating an undocumented process. Defining the full implementation in advance may be difficult. A limited analysis phase, with a time ceiling and expected conclusion, allows discovery without promising a fixed project on the basis of insufficient information about the work.
The risk appears when hours become the proposal's only explanation. Clients can compare hourly rates while missing differences in experience, decision quality or supervision requirements. Explain how the work will be directed and what it will allow the client to decide at the end of the intervention.
You also need to consider non-billable time, preparation and available capacity. An apparently comfortable hourly rate may not sustain the business if few hours are actually sold. Internal calculations should reflect the real service, including work the client does not directly see during the engagement.
To prepare your next step, explore Service pricing and profitability.
When is a defined deliverable preferable?
A defined deliverable makes sense when you can describe the output, required materials and acceptance criteria with reasonable clarity. The client buys a bounded intervention while you take responsibility for organising execution. Scope changes need explicit treatment so that a fixed price does not conceal unlimited work or unclear expectations on either side.
A page, diagnosis or configuration can be a deliverable, but its name is not enough. Specify what will be checked, which revisions are included and what information depends on the client. Two proposals for a page can involve very different research, content and validation work before delivery is complete.
Pricing a service when execution becomes faster should be explained through scope and responsibility. If you execute faster without losing quality, efficiency changes your cost but does not erase the work of deciding and verifying what is delivered.
Charging for a deliverable does not mean promising a commercial outcome outside your control. You can commit to a functioning page and agreed criteria; that should not become a sales guarantee dependent on the offer, traffic and the client's subsequent follow-up of enquiries.
What justifies charging for ongoing support?
Ongoing support justifies recurring billing when decisions, checks or interventions remain necessary over time. Define the responsibility retained, the cadence and response conditions. A monthly fee should correspond to agreed work or availability, even when the number of visible tasks changes between periods and the client does not receive a new asset every month.
Imagine an operation changing campaigns and offers throughout the month. Results analysis, prioritisation and coordination with the team may require continuity. In that scenario, the client is buying support with an identifiable function, rather than simply a new file delivered at the end of each month.
Distinguish reserved availability from unlimited requests. Explain priorities, timings dependent on others and when a new project requires a separate agreement. If the need disappears, the relationship should be reviewed against what still makes sense rather than assuming the original arrangement must continue unchanged.
Client retention in services depends on this match between expectations and delivered value. Continuing to charge because a fee has always existed, without explaining the current responsibility, makes an honest discussion about the future relationship much harder for both parties.
Can you combine models within one relationship?
You can combine models when each corresponds to a distinct phase or responsibility and the transition is understandable to the client. A diagnosis can precede a defined project, followed by support if a recurring need exists. The combination should reduce uncertainty and clarify work without creating overlapping charges for the same responsibility or task.
In the hypothetical undocumented-process example, initial analysis might produce a map and recommendation. The company then decides whether to implement. Ongoing support only enters when there is a continuing function, such as checking operation and adapting rules when the underlying process changes after implementation.
Do not make that sequence compulsory for everyone. One client may only need diagnosis because an internal team can implement it. Another may arrive with a clearly defined problem and require little discovery. The design should follow the context while preserving enough information for you to accept responsibility for the work.
Show in the proposal where each phase ends and what enables the next one to begin. Avoid presenting a succession of small charges that clients only discover as they proceed. Initial clarity lets them compare alternatives and make a decision without artificial surprises later in the relationship.
What changes when you use AI and automation?
AI and automation can change execution time, cost and task distribution, but the model should remain connected to your commitment to the client. Review preparation, validation and maintenance before concluding that the service became cheaper. Faster work in one stage does not demonstrate that all responsibilities have disappeared or that total costs have fallen equally.
A faster first version may require review or additional information before it becomes usable. Automation can reduce repetitive work while creating monitoring or exception-handling needs. Include those elements in internal analysis rather than comparing only the duration of the task that has visibly become shorter.
When rethinking services with AI, distinguish a change in method from a change in scope. If the commitment remains the same, explain how you assure quality. If the intervention changes, present its new value and limits as clearly as in the original proposal.
Do not retain fictional hours to justify a price or invent complexity that no longer exists. Use efficiency to reconsider capacity, quality and the offer itself. The relationship becomes stronger when the client understands the work they are actually buying and the responsibility that accompanies it.
How do you choose without complicating the proposal?
Choose the model that best clarifies the client's decision, the team's responsibility and how variations in work are managed. Check whether you can explain what is included, when it ends and what happens when a new request appears. Then confirm that expected revenue covers costs and capacity while leaving room for uncertainty in delivery.
Test the explanation against a typical case and a plausible exception. If an extra revision, delayed materials or changed objective creates uncertainty about who pays and who decides, the service needs clarification. Resolving those situations before selling prevents arguments no billing unit can settle by itself.
Track the first contracts with attention to effort and client questions. Adjust future proposals using that evidence while respecting existing agreements. The best model is one you can fulfil and explain, aligned with a real need and the way the work actually happens within your business.