The Evaluation Was Already Happening

In 2019, GLP-1 medications represented $1.43 per member per month in pharmacy spend for the average self-insured employer. By 2024, that number was $24.59.

One drug class. Seventeen times the spend in five years.

Most companies weren't building GLP-1 strategy into their vendor decisions two years ago. The vendors that address it today — with clinical wraparound, adherence support, risk stratification before a single prescription is written — either didn't exist or weren't on anyone's shortlist. The category formed faster than the evaluation process could follow it.

That's the gap worth paying attention to. Not which vendor won. How the decision got made, and what the starting point was when it did.

The decision you already made

Before we talk about what's coming, it's worth sitting with something that already happened.

Over the last five years, nearly every mid-market employer added a behavioral health point solution. The category grew fast, consultants recommended it broadly, and most companies made a reasonable decision with the information they had available.

Here's the uncomfortable part: the information they had was incomplete. Not incidentally — structurally.

Claims data codes depression in a typical commercial population at around 5–6%. The actual share of a workforce dealing with a diagnosable depressive episode in any given year runs 18–30%. That's not a rounding error. It's the difference between what the data showed and what was actually true about the population — and it means any vendor that evaluated your population from claims data alone was working from roughly a third of the real picture. The company that selected them was working from the same third.

This isn't a vendor failure. The vendors didn't hide anything. It's a starting-point failure — the baseline used to frame the problem was itself incomplete. The same pattern repeats across other conditions, and across a growing share of care that happens entirely outside claims: cash-pay therapy, carved-out benefits, medications that never feed back into the main data stream.

The result is a stack of decisions made against a partial picture, with outcomes measured against that same partial picture. It's internally consistent. It's also structurally misleading.

The problem is structural, not situational

GLP-1 spend and the behavioral health data gap aren't two isolated examples. They're symptoms of the same underlying issue: evaluation that begins before the full picture is assembled, in an environment that punishes incomplete pictures.

The mid-market self-insured employer sits in a particularly exposed position here. Claims volatility absorbs directly into the operating budget. Stop-loss renewals are running 12–18% this cycle. Fiduciary duty is real and active — not theoretical. And most companies are navigating decisions of genuine consequence without the dedicated analytical infrastructure that larger buyers take for granted.

That's the environment every vendor evaluation happens inside. And the frameworks most companies are using weren't built for it.

The symptoms are recognizable: vendor engagement that never reaches the levels promised at signing, ROI claims that don't reconcile to the broader cost picture, a creeping sense that it's impossible to tell which spend is working and which is just inertia. Roughly 41% of employers report they have too many point solutions — and the honest follow-up question, which of these is actually delivering, and against what baseline?, rarely has a clean answer. Because the baseline was never clean to begin with.

What it looks like to start from the right place

Before we evaluate a single vendor, we build a complete picture of what a buyer in that segment actually looks like — the cost structure, the risk profile, the regulatory environment, the gaps their own data can't show them.

We call these Buyer Lens profiles.

A Buyer Lens isn't a market overview or a vendor shortlist. It's a calibrated model of what a specific type of plan actually looks like before any vendor enters the conversation. We've built them across the segments where vendor evaluation decisions carry the most consequence — self-insured mid-market employers, commercial health plans, Medicare Advantage plans, and Medicaid programs, among others. The underlying dynamics are different in each. The discipline is the same.

In the self-insured mid-market Lens, the behavioral health gap isn't a surprise — it's a modeled input. We don't discover it during evaluation; we account for it before evaluation begins. Same for pharmacy trajectory, high-cost claimant concentration, the categories where spend is saturated versus where genuine white space exists.

When a vendor says “we'll find your hidden risk,” the Buyer Lens is what lets us ask the right follow-up: found from what starting point, against what baseline, validated how?

The starting point shapes the outcome. A framework built on a partial picture will keep producing partial answers.

The next one is already forming

The GLP-1 story didn't start in 2024. It started years earlier, when the spend was small enough that no one was building evaluation criteria around it. By the time it was obvious, most companies were already behind.

The behavioral health gap didn't appear recently either. It was always there, underneath the claims data, waiting for someone to ask what the data wasn't showing.

The next version of this story is already forming somewhere in the landscape. A drug class, a condition category, a fiduciary exposure moving faster than the procurement cycle. The companies that will handle it well aren't the ones who happen to land on the right vendor. They're the ones who built a starting point rigorous enough to ask the right questions before the vendor conversation begins.

The evaluation was already happening. The question is what it was built on.