How Much Does Consulting Cost? Engagement Models Explained for Small and Mid-Sized Businesses

The honest answer to "how much does consulting cost?" is that the number depends far less on who you hire than on how the work is structured. Two firms with similar people can quote figures that differ by a factor of three for the same problem, because one is proposing a two-week diagnostic and the other is proposing a six-month build. Neither is wrong. They are answering different questions.

This guide explains the five engagement models you will encounter, when each one fits, how to compare quotes fairly, and what Syzygy's own engagements typically cost.

How much does consulting cost?

For a small or mid-sized business, consulting ranges from a few thousand dollars for a short diagnostic to six figures for a multi-month implementation, and the model you buy determines most of that spread. Hourly work is cheap to start and expensive to finish. Fixed-fee projects are predictable but only as good as the scope behind them. Retainers and fractional roles cost less per month than a hire but add up over a year.

The mistake most buyers make is comparing headline numbers. A $4,000 quote and a $40,000 quote are not competing bids unless they are promising the same thing. Before you can judge price, you need to understand what each structure is actually selling.

What are the main consulting engagement models?

There are five, and each allocates risk between you and the consultant differently.

Hourly or time-and-materials

You pay for hours worked, usually against an estimate. This is the simplest model and the one that puts the most risk on you: if the work takes longer, you pay more. It suits genuinely uncertain work where nobody can define the deliverable yet, and short advisory conversations where a fixed fee would be silly.

A low hourly rate with an open-ended scope is the most expensive way to buy consulting.

The trap is that hourly rates are easy to compare and scope is not, so buyers anchor on the rate. A consultant who bills less per hour but needs twice as many hours to get to the same place costs more, and you will not find out until the invoice arrives.

Fixed project fee

You pay a set amount for a defined deliverable: an audit report, a working prototype, a system in production. The consultant carries the risk of overruns, which is why fixed fees are only offered when the scope is clear. This is the right model for most first engagements, because it forces both parties to write down exactly what "done" means. A good fixed-fee proposal lists deliverables, assumptions, what is out of scope, and what triggers a change order.

Retainer

You pay a recurring monthly fee for a defined amount of capacity, access, or support. Retainers make sense once something is in production and needs steady attention: iteration, monitoring, small improvements, and questions that come up as the business changes. They also make sense for advisory access, where the value is being able to call someone senior without negotiating a new contract each time. Look for a clear statement of what the monthly fee includes and how unused capacity is handled.

Fractional role

A fractional arrangement puts a senior person in a defined leadership seat, such as head of technology, for a fraction of a full-time schedule. It is a retainer with accountability attached: the person owns outcomes, not just deliverables. It costs a fraction of a full-time executive's fully loaded compensation and suits companies that need executive-level judgment on strategy, vendors, and risk, but not forty hours a week of it. We cover when this beats hiring in our guide to the fractional technology partner.

Value-based or outcome-based

You pay a share of the measured result, sometimes with a smaller base fee. In theory this aligns everyone perfectly. In practice it requires a baseline you both trust, a metric neither party can game, and a measurement window long enough to be meaningful. For a first engagement, those conditions rarely exist. Value-based pricing works best as a later stage, after a fixed-fee project has established what "value" means in your business and how it will be measured.

Which model fits which situation?

Match the model to how well you can define the outcome and how long you need the relationship.

• You do not yet know where AI would help: buy a fixed-fee diagnostic. Its whole purpose is to define the scope of everything that follows.
• You know the process and want proof it can be automated: buy a fixed-fee prototype with explicit success criteria.
• A prototype has passed and you want it in production: buy a fixed-fee or milestone-based implementation with a defined integration scope.
• Something is live and needs to keep improving: move to a retainer sized to the pace of change you expect.
• You need ongoing strategic judgment across the whole technology estate: consider a fractional role.
• Scope is truly unknowable and the work is short: hourly is fine, with a cap.

The sequence from diagnostic to prototype to implementation to retainer is not an accident. It mirrors how a well-run engagement is phased, and each step is priced so that stopping is cheap.

How do you compare consulting quotes?

Compare cost per unit of outcome, not the headline number. To get there, force every bidder to answer the same seven questions. We call this the Like-for-Like Quote Comparison.

1. What exactly is delivered? A list, not a paragraph. "An assessment" is not a deliverable; "a ranked list of five opportunities with estimated hours saved per week for each" is.
2. Who does the work? Senior people who scope it, or junior people who inherit it? Both can be fine, but the price should reflect the answer.
3. What is assumed about us? Data access, staff availability, existing systems. Assumptions that turn out false become change orders.
4. What is out of scope? Integration, training, and licenses are the usual omissions.
5. What happens after delivery? Warranty period, handover, documentation, and the price of ongoing support.
6. Who owns what? Code, prompts, data, and accounts should be yours. If they are not, the price is understated.
7. How will success be measured, and by whom? A bidder who cannot answer this is selling effort, not results.

Once each quote has answered all seven, divide the price by the outcome you actually care about: hours returned per week, quotes turned per day, or cases handled without escalation. The cheapest quote per unit of outcome is usually not the cheapest quote.

Compare quotes on cost per unit of outcome, not on the headline number.

What hidden costs should you budget for?

Your own team's time is the largest cost that never appears on a proposal. Discovery interviews, prototype testing, and training all take hours from people who have other jobs. Plan for it rather than being surprised by it.

Other costs that live outside the consulting fee:

• Software and usage fees for the AI services and tools the solution runs on, which continue after the engagement ends.
• Data cleanup, when the process you want to automate depends on records nobody has maintained.
• Integration work on the systems the solution needs to talk to, especially if your ERP or CRM vendor charges for API access.
• Change management, meaning the management attention required to get people to change how they work.

A responsible consultant will estimate these for you. An unusually low quote often means they have been left out.

Where Syzygy fits

Syzygy's engagements follow the diagnostic-first sequence described above, and the pricing is structured so that each step is a small bet. AI Audit & Prototyping engagements start at $3,000 and typically run two to four weeks. Design & Prototype engagements typically run $6,000 to $8,000 over four to six weeks. Implementation is scoped per project and typically runs eight to twelve weeks or more, depending on integration depth. Ongoing Partnership retainers start from $2,000 per month. Every quote states deliverables, assumptions, exclusions, and how results will be measured, so you can run the seven-question comparison against it. Details are on the pricing page.

A worked example

Consider a 25-person engineering consultancy that wants to cut the time its principals spend assembling proposals. Three firms respond.

The first quotes hourly at an attractive rate with an estimate of "roughly 60 to 100 hours." The second quotes a fixed $7,500 for a working prototype that drafts proposals from past projects, with a two-week test period and a written accuracy target. The third quotes $35,000 for a "proposal automation platform" with a twelve-month contract.

Run the seven questions. The first firm cannot state a deliverable, so its real cost is unknowable; at the top of its own estimate it exceeds the second quote with no defined result. The third firm is selling a build before anyone has proven the approach works, and its deliverable list turns out to include a license the consultancy would pay for regardless of results. The second firm has defined "done," put the overrun risk on itself, and made stopping cheap if the prototype disappoints.

The second quote is the right purchase, not because it is the middle number, but because it is the only one that can be compared to an outcome.

Where this goes wrong

• Anchoring on hourly rate. The rate is the least informative number on any proposal.
• Buying a build before a diagnostic. Skipping the cheap step to save time is the most common way to overspend.
• Signing a long contract for an unproven approach. Twelve-month commitments belong after a prototype has passed, not before.
• Ignoring your own costs. If your team cannot free up the hours, the engagement will run long no matter how it is priced.
• Treating value-based pricing as free. It is not free. It is a different bet, and it requires measurement you may not have yet.
• Choosing on price without reading the assumptions. Assumptions are where the change orders hide.

The bottom line

Consulting cost is a function of engagement model and scope, not of anyone's rate. Start with a fixed-fee diagnostic, move to a prototype with explicit success criteria, then to an implementation, then to a retainer, and price each step so that stopping is cheap. Compare quotes by forcing every bidder to define deliverables, assumptions, and measurement, then divide by the outcome you care about. If you want a clear view of what an engagement should include before you request quotes, start with our guide to what an AI consultant actually does.