zero gravitySOLUTIONS
← All insightsENGINEERING PARTNERSHIPS / WORKED FORECAST

Software budget forecasts that make uncertainty visible.

QUICK ANSWER

Forecast software delivery costs by separating actual spend, remaining work, and optional scope. Explain the assumptions behind the estimate and the reasons it changes. Use the forecast to compare next steps, rather than presenting an uncertain scope as a guaranteed total.

Inside this article 6 sections

A software budget can be numerically precise and still be difficult to trust. If the forecast does not explain unfinished work, unresolved dependencies, and the cost of decisions still being made, the apparent precision hides the uncertainty rather than managing it.

A useful delivery forecast connects money to a defined scope and a current understanding of the work. It separates spending already incurred from the estimated cost to finish, then explains what could move that estimate. This article uses a hypothetical project to show the structure; the amounts are illustrative, not Zero Gravity rates or a quote.

Separate actual spend, remaining work, and optional scope

Begin with an agreed scope boundary. In the example, a team is implementing an account import feature, including validation, review, submission, and operational support. A proposed automated enrichment step is optional. If the forecast quietly includes that option in some updates and excludes it in others, the client cannot compare the numbers meaningfully.

Suppose recorded spend is €48,000. The team estimates €26,000–€37,000 to complete the agreed scope. The resulting forecast is €74,000–€85,000, excluding optional enrichment. This range should be tied to named assumptions about validation rules, integration behavior, and release work. It is not a statistical confidence interval unless the estimation method actually establishes one.

ILLUSTRATIVE FORECAST / SAME SCOPE, VISIBLE ASSUMPTIONS
Spent to date€48,000 — recorded actuals
Remaining scope€26,000–€37,000 — current estimate
Forecast at completion€74,000–€85,000 — actuals plus remaining work
Optional enrichmentSeparate decision — excluded from this forecast
Arithmetic example: €48,000 + €26,000 = €74,000; €48,000 + €37,000 = €85,000. The range is an estimate, not a guaranteed bound.

Explain the uncertainty through specific drivers

“There may be unexpected work” is true of many projects and does not help a buyer choose an action. A more useful note says that the import API’s duplicate behavior is undocumented, that the next test will establish whether reconciliation is needed, and that reconciliation is the main difference between the lower and upper estimate.

Separate uncertainty about effort from uncertainty about scope. The team may understand the required behavior but not yet know how much refactoring it takes. Alternatively, the business may not have decided whether invalid rows should block the whole import. The first problem needs technical evidence; the second needs a product decision. Calling both “engineering risk” conceals who can reduce them.

Use a short assumption register rather than a large catalogue of hypothetical disasters. Each entry should identify the assumption, its effect on remaining work, the owner of the next investigation or decision, and when the estimate will be revisited. Remove assumptions that have been resolved so the register reflects current uncertainty.

Show why the forecast changed

A new forecast should explain its movement from the previous one. Perhaps actual implementation took less effort than expected, while a confirmed requirement added a reconciliation screen. Show those changes separately. Otherwise, favorable and unfavorable movements can cancel numerically and hide important information about the project.

For the example, imagine that a dependency test removes €4,000 of anticipated recovery work but an approved audit requirement adds €6,000. The net change is €2,000, yet the business should still see both decisions. One reflects reduced uncertainty; the other reflects added scope. This distinction matters when evaluating future estimates and deciding what to build next.

Keep scope changes linked to the decision that authorized them. A shared change record can describe the requested outcome, the estimate impact, the effect on delivery, and the chosen option. It need not become an elaborate approval bureaucracy, but the client should be able to reconstruct why the expected total moved.

Do not confuse available capacity with guaranteed output

In a dedicated-team engagement, agreed capacity and rates may make the recurring engineering cost predictable. They do not make the amount of finished product equally predictable. Holidays, review dependencies, unclear requirements, and work discovered during implementation can affect throughput even when the team size is unchanged.

Report capacity, completed outcomes, and remaining scope together. If a critical reviewer is unavailable, explain which work can proceed and which decision remains blocked. More assigned tasks do not necessarily resolve a constrained review or release path.

Separate engineering costs from cloud infrastructure, AI usage, subscriptions, and other external charges. Forecast variable services using workload assumptions and an owner who can inspect the actual usage. A model API bill should not appear as an unexplained surprise at the end of an otherwise transparent engineering invoice.

Agree on the cost record you will receive

Before commissioning delivery, ask for a sample report that separates approved scope, actual spend, estimated remaining work, and decisions that have no estimate yet. Use one currency and an explicit treatment of external costs. A report should let the client connect a movement in the forecast to a changed assumption, a new requirement, or completed work. It should also make the person who can approve the next commitment visible.

The small program below demonstrates the arithmetic with hypothetical hours and rates. Its known work has a bounded range, while an unresolved provider dependency remains outside that range. This is deliberate: adding a precise total around known items should not conceal an unknown that could change the delivery. Changing the input file lets your team inspect how the range moves without depending on a spreadsheet formula hidden in a presentation.

For a scoped engagement, agree on the acceptance boundary and how a material change will be priced before work expands. For a dedicated team, agree on available capacity, included roles, reporting cadence, and the budget approval point. We can use the same record through either arrangement so financial visibility follows the actual work rather than being recreated at invoice time.

Use the forecast to choose the next commitment

A good budget conversation presents choices. The team might complete the agreed import flow first, defer enrichment, or spend a bounded amount testing the uncertain API behavior before committing to the full implementation. Explain what each choice buys in terms of product capability or reduced uncertainty.

Keep cash timing separate from this delivery estimate where necessary. An invoice schedule and an estimated cost to finish answer different questions. The worksheet here supports delivery planning; it does not determine accounting treatment or financing decisions.

Review the forecast at meaningful decision points and at the engagement’s agreed reporting cadence. Our dedicated development teams approach includes shared visibility into work and costs. The operating agreement helps define those reporting responsibilities, while the AI cost worksheet explores the separate economics of an automated workflow.

FROM INSIGHT TO ACTIONDedicated development teams

Keep delivery decisions connected to the budget.

Our dedicated team engagement can include an agreed effort record, transparent assumptions, and a forecast review that explains changes before the next commitment.

Explore dedicated teamsDiscuss your team

Start with your goals.
We’ll work through the next step together.