A considered process, senior-led from start to finish.

Five phases is the shape. What follows is what actually happens inside them, because a process diagram tells you the order of events and almost nothing about whether the work is any good.

Listen — and write down the awkward parts

The first conversation is thirty minutes and produces no document. We want the shape of the workforce, what is straining, what has already been tried, and what would have to be true for this to be worth your time. We also want the constraints nobody volunteers: the contract that names the employing entity, the union agreement, the transaction eighteen months out, the system somebody has already promised the board would be replaced.

Those constraints decide more outcomes than any comparison does, and they surface in conversation or not at all. If we discover one in month two it usually means restarting.

Frame — settle structure before anyone quotes

This is the phase most processes skip, and skipping it is why decisions get reversed. On the PEO and ASO side it means resolving the employer-of-record question on its own terms before a single rate is discussed. On the platform side it means establishing the real shape of the workforce — how much is hourly, how many jurisdictions, how many genuinely awkward pay rules — because that determines which capabilities are load-bearing and which are marketing.

The output is a written frame: what is being decided, what is explicitly not in scope, and which two or three variables the answer actually turns on. Everything after this is measured against that document.

Evaluate — comparison or conviction, stated plainly

Which of the two you get depends on the market, not on what is easier to sell. PEO and ASO is genuinely fragmented, so we compare: shortlist, normalize every proposal onto one basis, and model them against your real payroll for a full year including seasonality and bonus cycles. Platform selection is not fragmented in our view, so you get a considered recommendation and an explicit invitation to pressure-test it.

Either way you see the reasoning, not just the conclusion. A recommendation you cannot interrogate is a recommendation you cannot own.

Decide — and be told when the answer is no

You make the call. Our job is to make sure the trade-offs are fully in view first, including the ones that make our own recommendation less attractive. A meaningful share of our evaluations end with advice to change nothing, and that outcome is priced into how we work rather than treated as a failure.

Sustain — the phase that decides whether any of it worked

Most of the value in these decisions is realized or lost after the signature. So the same people stay through implementation readiness, cutover sequencing, and the reviews afterwards. When a renewal arrives, when you open in a new state, when the platform underperforms its promise, the person picking it up already knows the history.

Engagements go wrong in predictable ways, and most of them are avoidable from your side rather than ours. Three things make the difference.

One person who can decide

Not a committee, and not a coordinator relaying questions to a committee. Someone with the authority to settle a question and the calendar space to settle it this week. Where that role is vacant, every phase stretches and the cost lands on both sides.

The real data, not the tidy version

Current rates, actual claims experience, the true headcount by state, the pay rules as they are practised rather than as the handbook describes them. We are not auditing you. But an analysis built on the tidy version produces a recommendation that fits a company you do not have.

Willingness to hear an inconvenient answer

Sometimes the finding is that the problem is upstream of the thing you asked us to evaluate — a staffing gap, an implementation done badly, a process nobody owns. That is a harder conversation than a vendor comparison, and it is usually the more valuable one.