Project management for small software development shops
How a small dev shop can choose between sprints and fixed scope, estimate with a buffer, and handle change requests, with worked arithmetic for a quote and a sprint capacity.

A small software shop has two common ways to sell work: a fixed scope for a fixed price, or an open-ended stream of sprints. Each needs different management, and mixing them up is how projects drift. This guide compares the two, shows how to estimate with a buffer, and gives a simple way to handle changes.
Sprints versus fixed scope
Fixed scope. The client buys a defined set of features for an agreed price. You plan the whole thing up front, then deliver against the plan. The client gets certainty on cost. You carry the risk of overrun if the estimate is too low.
Sprints. The client buys a team's time for fixed periods, often two weeks. At the start of each sprint, you and the client choose what to build next from a prioritized list. Scope changes freely, but the cost per sprint stays the same.
A rough rule of thumb for choosing:
- Fixed scope suits a clear spec, a hard budget and a delivery that can be accepted or rejected.
- Sprints suit an unclear spec, a product that will evolve, or a client who wants to steer.
A hybrid is also possible: a short paid discovery to define the scope, then a fixed-price build, then monthly sprints for improvements.
Estimating with a buffer
Estimates often fail in a familiar way: they cover the work you can see and miss the rest. The fix is to list tasks small enough to estimate, then add a buffer for the unknowns.
Say you are quoting a client portal. Break it into five tasks:
- Login and accounts: 8 hours.
- Dashboard: 13 hours.
- Settings page: 5 hours.
- Document upload: 20 hours.
- Email notifications: 10 hours.
Sum: 8 + 13 + 5 + 20 + 10 = 56 hours.
Add a 25% buffer: 56 x 0.25 = 14 hours. Total 56 + 14 = 70 hours.
At $110 per hour: 70 x 110 = $7,700.
Show the client the task list and the total. Keep the buffer inside the price rather than as a separate line, unless you want to refund unused buffer. Then check your estimate after delivery. If you keep landing at 80 hours on jobs you estimated at 70, raise the buffer.
Planning a sprint
A sprint needs a capacity number so you do not overload it. Capacity is not team size times days. It is the productive hours.
Say two developers, a two-week sprint of 10 working days, and 6 productive hours per person per day, after meetings, reviews and email.
Capacity = 2 x 10 x 6 = 120 hours.
Now reserve some of it. Keep 20% for bugs and support: 120 x 0.20 = 24 hours. That leaves 120 - 24 = 96 hours for planned work.
Pull tasks from the backlog until their estimates add to 96 hours, and stop. Track what you actually completed each sprint. After three sprints you will have a number that is better than a guess.
Handling change
Changes are normal. The risk is in absorbing them silently. Use the same three steps on both fixed scope and sprints.
- Write it down. One sentence on what the client wants.
- Estimate it. Hours, and the effect on the end date.
- Decide. The client approves, swaps it for something else, or drops it.
Say the client asks for an export-to-spreadsheet feature, which you estimate at 12 hours. At $110 per hour: 12 x 110 = $1,320. On a fixed-scope job, that becomes a change order for $1,320. On sprints, it goes into the backlog and the client chooses what it displaces.
Keep a change log with the date, the request, the estimate and the decision. It saves arguments at the end.
Track the plan against actuals
Whatever model you use, track hours against the plan weekly. For a fixed-price job, compare hours used with the 70-hour budget and the percentage of features done. If 50% of the hours are used and 30% of the features are done, you know early.
HireOrSell offers project dashboards, Gantt charts with drag-and-drop scheduling, milestones, budget tracking and role-based permissions for team management, which cover the planning and tracking described here. If you are choosing among tools, the field is mapped in best free time tracking tools for small teams, and a comparable phase-by-phase budget for design work is in time tracking for web design agencies.
Review your estimates after delivery
An estimate improves only if you compare it with the result. When a build ends, write down for each task the estimated hours and the actual hours. Use the totals to work out your own error rate.
In the portal example, say the actual hours were 64 against a 56-hour base estimate. 64 / 56 = 1.14, so the work ran about 14% over the base. A 25% buffer was enough, and you could consider trimming it. If the actual had been 74 hours, 74 / 56 = 1.32, and the buffer would have been too small. Keep one line per job in a sheet and the right buffer for your shop will emerge from your own data.
Keep the rhythm simple
A small team does not need a heavy process. A weekly plan, a mid-week check, a Friday demo and a short retrospective every second week are enough. Write down what you will not do, as well as what you will.
Next step
Take your next quote and write the tasks down at the five to twenty hour level. Add up the hours, add your buffer, and compare the total with what you would have guessed. Keep both numbers and check them when the job is done.
Keep reading
- What is realization rate in a service business?Realization rate compares what you bill, and what you collect, with the value of your time at standard rates. See the formulas and a worked example with each step of the loss.
- The real cost of one week of scope creepA worked example of one week of scope creep on a fixed-fee website project, with the extra hours, the cost, the lost margin and the effective hourly rate calculated.
- Time tracking for web design agenciesHow a web design agency can track time by design phase, count revisions, and log the handoff to development, with a worked 90-hour budget and an overrun example.
In HireOrSell
Ready to run it all in one place?
Projects, clients, time tracking and invoicing in one platform.