Why we retired the broker model in HCM and workforce management.

For most of this practice’s history, we did what the market expected of an advisor. A company in the mid-market would decide it needed a new HCM or workforce-management platform, and we would help them run a comparison — shortlist the vendors, sit through the demos, score the responses, negotiate the terms. It is the model the entire category is built around. It is also, we have come to believe, the wrong model for the companies we work with.

The bake-off optimizes for the wrong thing

A competitive evaluation feels rigorous. Everyone in the room can point to the scoring matrix and the parallel demos and say the decision was made carefully. But rigor of process is not the same as quality of outcome. What the multi-vendor bake-off actually optimizes for is defensibility — the ability to justify the choice after the fact — not fit. And in a segment where the platforms have largely converged on capability, the differences that matter are not the ones a demo surfaces.

The features that get demonstrated are the features that demo well. The things that decide whether an implementation succeeds — the quality of the implementation team, how the platform behaves under the specific shape of your workforce, what breaks in month four — do not appear in a scripted demo. So a buyer runs a process that looks thorough and learns very little that predicts the outcome they actually care about.

What convergence actually looks like

It is worth being precise about that claim, because "the platforms have converged" is the sort of statement that can be a substitute for thinking. What we mean is narrower than it sounds. Twenty years ago the difference between two systems was whether a thing could be done at all. Today, in this segment, the difference is almost never capability. It is how much configuration effort a capability costs you, how the system behaves at the edges of your particular rules, and what the vendor’s delivery organization is actually like to work with.

All three of those are real and they matter enormously. None of them is legible in a demo, because a demo is a controlled environment running the vendor’s own data. You are watching a rehearsed performance of the easy case. The hard case — the seventh-consecutive-day premium, the call-in guarantee, the three accrual rules nobody can justify but nobody will remove — is precisely what a demo is constructed to avoid encountering.

So the bake-off compares platforms along the dimension where they are most alike, using the one method guaranteed to conceal the dimensions where they differ. That is not a rigorous process with a few blind spots. It is a process pointed at the wrong target.

Weeks of demos are a tax on the buyer

There is also a cost that no one puts on the scorecard: the buyer’s time. Running a real bake-off in this segment consumes weeks of senior attention — the CFO, the head of HR, IT, sometimes the CEO — spread across parallel vendor tracks that mostly repeat each other. That time is the scarcest resource in a growing company, and the bake-off spends it on a comparison whose conclusion, in our experience, is knowable in advance by someone who has already done the work.

What we were actually selling

There is an uncomfortable version of this worth saying out loud, because it is the part that made the change necessary rather than merely interesting.

The bake-off is good business for an advisor. It is legible to a board, it generates artifacts — matrices, scorecards, a written recommendation — and it takes weeks. Those weeks are not the advisor’s to spend: they come out of the payroll lead, the HR lead and somebody from finance. When we looked honestly at the last several evaluations we had run, the conclusion at the end was one we could have written at the beginning. The weeks in between had gone into assembling the evidence that would let everyone feel the conclusion had been earned.

That is not fraud. Every client got a defensible decision and most of them got a good one. But it is a service performed on process rather than on judgment, and once you notice that about your own practice you have to either change the practice or stop claiming to sell judgment.

What we do with the time instead

Removing the comparison did not shorten the engagement as much as you might expect. It moved the work.

What replaced it is the part that was always undersupplied: auditing source data before anyone commits to a timeline, deciding which of the client’s own policies are load-bearing and which are merely old, naming an internal owner with genuine authority to decide, sequencing the cutover so that two full parallel payroll cycles survive contact with the schedule, and planning the change so managers meet the new system before it is live rather than through a launch email.

None of that is glamorous, and none of it is what a buyer believes they are shopping for. It is also, reliably, where the difference between a good outcome and an expensive one gets decided. Moving our hours from the selection to the implementation is the single change that most improved what our clients actually end up with.

Conviction is not the same as a conflict of interest

The obvious objection is that an advisor who recommends one platform starts to look like a reseller. We take that seriously, which is why the structure matters. On our PEO and ASO work, where the market is genuinely fragmented and the right answer is client-specific, we stay vendor-neutral and comparison-driven. On the platform question, where we have reached a settled conclusion, we say so plainly — and we tell any buyer who wants to run their own comparison to go do it and come back if our recommendation still fits.

That is the test of whether conviction is honest: are you willing to lose the engagement to the buyer’s own diligence? We are. We would rather be your second conversation and your right answer than your first conversation and your compromise.

What would change our mind

A conviction that cannot be falsified is a position rather than a judgment, so it is worth stating plainly what would move ours.

If a platform we do not represent began winning consistently on the dimensions we actually weight — configuration effort for complex pay rules, delivery quality, behavior in compliance-heavy multi-state environments — we would change the recommendation. If our partner’s implementation organization degraded, we would change it, and that is the failure mode we watch for most closely, because it is the one that arrives quietly. If the mid-market stopped being well served by the architecture that serves it now, we would change our position and say so here.

We re-examine the question on a deliberate rhythm rather than waiting to be surprised by it. The recommendation is a conclusion we hold, not a relationship we defend.

Retiring the broker model was not a marketing decision. It was an admission that the process we were selling had stopped adding value — and that the more useful thing we can offer a mid-market buyer is a considered recommendation we are prepared to stand behind, and to be argued out of.