How Software Consulting Is Priced, and What Each Model Rewards

Most conversations about consulting cost focus on the number. The number matters less than the structure, because the structure determines what the engagement drifts toward when nobody is watching.

Four models dominate. Each rewards something different.

Fixed fee

Defined scope, defined deliverable, defined date, one price.

What it rewards: finishing. Once the price is set, every extra day costs the firm margin, so the engagement is pulled toward a conclusion. For an assessment — an architecture review, due diligence, a readiness check — this is almost always the right structure, because you want a decision by a date, not an ongoing exploration.

The risk: scope arguments. If the boundary is vague, you will spend week three debating whether something is included. Fix this by specifying deliverables as physical artefacts rather than activities. "A ranked risk register with effort estimates, a target-state diagram, a 90-day plan" is enforceable. "A comprehensive assessment" is not.

Time and materials

A day rate against an estimated range.

What it rewards: continuation. Not dishonestly — nobody is padding hours — but the natural gravity of the arrangement is toward more work rather than less, because the engagement ends when someone decides to end it.

Appropriate when scope genuinely cannot be known upfront: an unfamiliar legacy system, an investigation whose shape depends on what the first week finds. A warning sign when scope obviously could be defined and the firm prefers not to commit.

The protection: a not-to-exceed ceiling with a mandatory checkpoint at 50%, where continuing requires an explicit decision rather than a default.

Retained advisory

A fixed monthly amount for a defined slice of senior attention.

What it rewards: availability. You are buying the ability to ask a good question on a Tuesday and get a considered answer, which is genuinely valuable and hard to buy any other way.

This works well after an assessment, when the value has shifted from a document to a stream of decisions. It works poorly as a first engagement, because there is no shared context yet and the retainer becomes a slow, expensive discovery phase.

The failure mode: silent under-consumption. Three months in, nobody is calling, and the invoice keeps arriving. Set a floor — a scheduled monthly session that happens whether or not anything is on fire — or the retainer quietly becomes a subscription to nothing.

Outcome-linked

Part of the fee contingent on a measurable result.

What it rewards: whatever you measured, which is the entire problem.

It sounds like perfect alignment and rarely is, because outcomes depend on execution the consultant does not control. If the fee depends on a delivery date, and your team is the one delivering, you have created an argument for later rather than an incentive now. Honest firms are cautious about this structure for exactly that reason, and the caution is a good sign rather than a lack of confidence.

It works in narrow cases where the metric is unambiguous, attributable and short-horizon — a specific cost reduction, a measured latency target on a system the consultant is actually changing.

The clause that matters more than the model

Whichever you pick: contract and price the assessment separately from any implementation.

The structural conflict in this industry is that the assessment fee is small and the implementation contract that follows is large. Any firm doing both has a standing financial reason to find more work than exists. Hiring better people does not remove an incentive.

Separating the contracts does. It costs nothing, and no reputable firm will object.

What a fair price looks like

Day rates vary so much by region and seniority that quoting them is useless. Proportionality travels better:

A fixed-fee assessment costing more than about 3% of the decision it informs is hard to justify. One costing less than about 0.5% is probably not buying enough senior attention to be worth reading.

Judge the fee against the size of the decision, not against a day rate. A two-week review before a modest platform investment sits at one end; full due diligence on an acquisition sits at the other, because the decision is larger.

Frequently Asked Questions

Which pricing model is best for an architecture review?

Fixed fee, with deliverables itemised as physical artefacts. You want a decision by a date, and fixed fee is the only structure that pulls toward finishing.

Is time and materials always a bad sign?

No. It is appropriate where scope genuinely cannot be known — unfamiliar legacy systems, investigations that branch. It is a warning sign only when the scope clearly could have been defined and the firm chose not to commit.

How do I stop a retainer from becoming a subscription to nothing?

Set a floor: a scheduled monthly session that happens regardless, with a standing agenda. Retainers fail through silent under-consumption far more often than through overuse.

Should I ask for outcome-based pricing?

Only where the metric is unambiguous, attributable to the consultant's own work, and measurable within the engagement. Otherwise you have created a future argument rather than a present incentive.

How much should a technical assessment cost?

Between roughly 0.5% and 3% of the decision it informs. Outside that band in either direction, ask why.

Can we negotiate the discovery phase down?

Yes, and you should ask. Much of the descriptive work in an assessment — dependency mapping, version drift, dead code — is now automated and takes hours rather than weeks. If a firm still quotes a three-week discovery at full rates, ask what share is automated and how the fee reflects it.

---

The full guide — five engagement types, deliverables, SOW clauses and the 30-day test — is here: Software Development Consulting Services: What You Actually Get for the Money.

TechCirkle builds products and runs assessments alongside custom software development. Get in touch for a scoping call.