Every time-and-materials contract contains a quiet transfer. The firm does not know how hard the work will be, so the client agrees to pay for finding out.
That arrangement is so normal in IAM consulting that nobody reads it as a transfer anymore. Discovery takes longer than scoped: billed. The legacy tenant hides a decade of undocumented exceptions: billed. The integration that should have taken a week takes five: billed. The firm carries no delivery risk because the meter is always running.
We price the other way. Our engagements are fixed-fee and outcome-based. The client buys a result, and the variance between the estimate and reality is our problem.
This is not a marketing posture. Fixed-fee pricing is only rational for a firm whose delivery method actually removes the variance. This essay is the argument for why ours does.
Why hourly billing survives
Hourly billing survives because IAM delivery has been genuinely unpredictable.
The unpredictability has a specific source. Most of the work in an identity engagement is mechanical: reading what a tenant does, rebuilding configuration application by application, testing each cutover. The mechanical work is executed by hand, and hand-executed work has wide variance. An engineer's week disappears into one badly documented integration. Nobody can see the whole tenant, so nobody can bound the effort.
A firm that prices a fixed fee on top of that variance is gambling. The honest ones lose the gamble often enough to go back to hourly. So the industry standardized on a model where the client funds the uncertainty, and estimates are openers rather than commitments.
The variance is not a law of nature. It is a property of the method.
What changes the variance
Our method runs on control-plane orchestration. We read a tenant's configuration into a normalized model, surface the conflicts and drift for human decisions, and write the approved target state back with a full audit trail.
In an engagement run this way, up to 80 percent of per-application configuration is executed by agent orchestration. The agents do the work that used to consume engineer weeks: reading state, applying the normalized policy to each application, logging every change. Humans do the part that deserves humans: policy decisions, exception review, approval of every write.
Look at what that split does to variance. The mechanical work, which carried most of the unpredictability, becomes repeatable. Run the read twice and you get the same answer. The per-application step behaves like the fortieth run of a tested process, because it is one. What remains variable is the judgment work, and judgment work is bounded: a tenant has a knowable number of policy decisions, visible after the first normalized read.
When the variance is bounded, the gamble disappears. We can see the shape of the work before we commit to the number. So we commit to the number.
What the incentive does
Pricing the outcome changes behavior on both sides, and the changes compound.
For the firm: every hour of waste comes out of our margin, not the client's budget. We invest in tooling that makes delivery faster, because the firm captures the benefit. Under hourly billing, that incentive points the other way. Efficiency reduces revenue. It is worth being plain about that: a time-and-materials firm earns more when your migration goes badly.
For the client: scope conversations get honest. You are not approving timesheets, you are holding us to a result. The change-order theater that pads hourly engagements has no oxygen, because added complexity inside the agreed outcome is ours to absorb.
For both: the discovery phase stops being a profit center. We want discovery to be fast and complete, because we own what it finds. The normalized read makes the tenant legible up front, which is exactly when legibility protects the client too.
There is one more alignment worth naming. We are services only. We do not resell Okta or Auth0 licenses, so no recommendation we make is shaped by partner margin on a product. The fee is the entire commercial relationship. One number, attached to one outcome.
The test a buyer should apply
Any firm can put the words fixed fee in a proposal. The argument behind the words is what to interrogate.
Ask what delivery method removes enough variance to make the fixed price rational. If the answer is a staffing plan, the firm is gambling. A firm that gambles on your migration will be reading the contract for exits when the gamble goes wrong. If the answer is a mechanism you can see, with a normalized read of your own tenant as evidence, the price means something.
Fixed-fee pricing is not generosity. It is what a delivery method that actually works lets you charge. The method came first. The pricing is the proof.