The Proof Tax: Why 321 Digital Health Buyers Just Put a Price on Trust

Why this issue exists

Every issue of this newsletter since #7 has run on one rule: one buyer, one claim, stress-tested against the market. This issue breaks that rule on purpose. Instead of one executive in one guest session, the buyer voice this time is 321 of them — health plan, employer, and health system decision-makers who just told the Peterson Health Technology Institute, on the record, in a survey fielded in July 2026 and published this month, exactly how they're deciding who gets paid and who gets cut. That's not an anecdote. That's a verdict.

And the verdict matters more right now than almost anything a single executive could say in a Zoom room, because its landing in the same month venture capital quietly hit a new high-water mark for digital health funding, CMS rolled out a ten-year payment model that bakes this exact buyer behavior into federal law, and the employer benefits market started actively RFP-ing a third of its existing vendor contracts out the door. The claim, the stress test, and the provocation of this issue all point at the same uncomfortable fact: the money hasn't dried up. The patience has.

Curb moment: Three hundred twenty-one buyers just signed the same confession. This issue is the stress test on all of them at once.

THE CLAIM: WE'LL PAY, BUT ONLY FOR RESULTS WE CAN PROVE, AND WE'RE NOT GROWING THE PIE TO DO IT

PHTI surveyed 321 people who make or heavily influence digital health purchasing decisions, 106 at health plans, 115 at employers, 100 at health systems, in July 2026. Here's what they said, in aggregate, about how they actually buy.

First, the money for proof is real. 57% of all purchasers are currently running performance-based contracts, including 68% of employers and 55% of health plans. Where those contracts exist, they're not symbolic, 85% put at least a quarter of total contract value at risk against performance, and 31% put more than half at risk. Cost of the solution and proven track record are now the top factors across every purchaser type when picking a vendor, ahead of relationships, ahead of brand, ahead of feature lists.

Second, this is the part most founders skip past — 96 to 97% of purchasers say they expect a digital health solution to improve outcomes, expand access, and strengthen their own financial performance, all three, simultaneously. That's not a wish list. That's the rubric every vendor is being graded against, whether they know it or not.

Third, the ceiling is real too. A majority of purchasers (56%) plan to only maintain their digital health spending over the next 12 months, not grow it, a reversal from 2024, when half the market was still planning to increase. Digital health portfolios are shrinking in breadth: only 9% of purchasers now run 11 or more unique solutions, down from 23% in 2025. And cost of the solution remains the top reason purchasers say they'd switch vendors in 2026, with lack of meaningful engagement rising into every segment's top three for the first time.

Put the three findings together and the claim writes itself: purchasers aren't walking away from digital health. They're done paying for potential. They'll pay for proof, with real money at risk, inside a budget that isn't getting any bigger, which means every dollar a vendor wins now comes directly out of some other vendor's renewal.

Curb moment: The market didn't shrink. It just stopped paying for the pitch and started paying for the receipt.

Stress-Test #1: VENTURE CAPITAL IS STILL BETTING ON GROWTH THE BUYERS JUST SAID ISN'T COMING

If 321 purchasers controlling budgets across health plans, employers, and health systems are telling PHTI they're holding spending flat, you'd expect the venture capital pouring into digital health to be cooling off to match. It isn't, and the mismatch is the story.

Rock Health's Q3 2026 market update puts U.S. digital health startup funding at $3.7B across 114 deals for the quarter, with $11.1B deployed year-to-date across 358 deals. CB Insights' own Q3 count runs even higher, $7.7B across 546 deals globally, with mega-rounds (Neko Health's $700M Series C, Ortet's $500M launch round among them) accounting for 56% of total dollars for a third straight quarter. Generative AI and AI-enabled workflow infrastructure was the single fastest-growing investment category in the quarter.

Two data sets, two different counting methods, same signal: investors are still writing checks sized for a market that's expanding, into a buyer base that just told an independent research institute, in writing, it's holding the line. Mega-round concentration makes this worse, not better, fewer companies are absorbing more capital, meaning more of this money is chasing the same flat, maintenance-mode budget PHTI just described. Someone's math is wrong. Either purchasing data undersells how much decision-makers will flex for a truly differentiated product, or a lot of term sheets signed this quarter are pricing in TAM growth the buyer side has already quietly retired.

Curb moment: VCs are still funding a growth story. The people holding the checkbook just told a research institute they're done growing the checkbook.

Stress-Test #2: FEDERAL POLICY JUST MADE THE BUYERS' BEHAVIOR THE LAW OF THE LAND

If performance-based, outcomes-at-risk purchasing were just a private-market trend employers and health plans invented to squeeze vendors, you'd expect federal policy to be lagging behind it, or silent on it entirely. It's not. It's catching up fast, in the government's largest payer.

CMS's Innovation Center launched the ACCESS model this year, a ten-year, voluntary payment model running 2026 through 2036 that replaces fee-for-service billing for technology-enabled chronic care with fixed, outcome-aligned payments tied directly to whether a patient hits defined health targets. Eligible conditions include hypertension, diabetes, heart disease, kidney disease, musculoskeletal pain, and behavioral health, nearly the identical clinical footprint PHTI's purchasers named as their top focus areas. Participants have to prove three capabilities to qualify: continuous digital patient engagement, outcome measurement tracking, and quarterly clinical integration updates. That's not a private payer's contracting preference anymore. That's now baked into a decadelong federal reimbursement structure. This should concern vendors who still treat "performance-based contracting" as a scary clause a tough procurement officer added to make their life difficult. It isn't a negotiating tactic. It's where the entire system, private and public, payer and employer, is converging at the same time, for the same reason: nobody wants to pre-pay for healthcare technology anymore.

Curb moment: Employers and health plans didn't invent outcomesbased payment to be difficult. CMS just spent a decade of federal policy confirming they were early.

Stress-Test #3: THE SAME BUYERS DEMANDING PROOF ARE ALSO CUTTING VENDORS WHO CAN'T SUPPLY IT — AND QUIETLY BUILDING AROUND THE ONES THEY KEEP

If purchasers are genuinely raising the bar on proof rather than just complaining about cost, you'd expect two things to show up in parallel: active vendor consolidation, and health systems starting to build some of this capability themselves rather than wait on a vendor roadmap. Both are happening, and together they box vendors in from two directions at once.

On consolidation: 75% of large employers now manage four or more digital health point solutions at a median annual maintenance cost of $580,000, and 74% report high "point solution fatigue." 61% say their point solutions can't demonstrate verifiable, claims-based ROI — which tracks exactly with PHTI's finding that cost savings is the single area where vendors most often fail to hit contracted performance metrics (27% overall, rising to 31% among health plans). 51% of large employers are actively planning to RFP or replace their health and wellbeing vendors specifically to prioritize simplicity and provable outcomes. That's not a budget freeze. That's an active cull.

On the build side: health system technology leaders' willingness to simply wait for an EHR or point-solution vendor to ship a needed feature dropped from 52% to just 22% in a single year, as AI-assisted development quietly reopens the "build it ourselves" option CIOs had mostly abandoned a decade ago. Epic and Oracle together now hold roughly 65% of acute-care EHR market share and are racing to make the EHR itself, not a bolt-on point solution, the delivery layer for ambient documentation, decision support, and patient engagement. Every feature Epic or Oracle absorbs natively is a point solution that no longer needs to exist.

The sales job market is already pricing this in. Digital health employers are restructuring sales and customer-success teams around AI-native platforms, cutting roles built around manual onboarding and point-solution demos, in the same quarters they keep reqs open for anyone who can speak payer contracting, claims reconciliation, and performance-based deal structuring fluently. A vendor reading PHTI's survey and thinking "I just need better outcomes data" is only half right. You also need a reason to exist that a consolidating employer won't RFP away and a feature set the platform of record hasn't already decided to build.

Curb moment: Buyers aren't just asking vendors to prove it. A growing number have decided it's faster to build it, or fewer, than keep paying someone else to prove it.

The Provocation

Here's where I push back, not on PHTI, whose data is solid, but on how nearly everyone in this industry is reading it.

The consensus take on this survey will be "show buyers the ROI." That's table stakes advice dressed up as insight, and it misses what the data is actually screaming. Look again at where vendors fail most: cost savings, the one metric purchasers increasingly weigh above almost everything else. Now look at why: 45% of purchasers cite difficulty integrating vendor data with their own claims and utilization systems as their single biggest barrier to extracting value from a performance-based contract. 40% say vendors report outcomes from favorable subpopulations instead of the full eligible group they're contractually supposed to cover. It isn't that vendors lack proof. PHTI's own companion research on AI adoption found the exact same structural problem one layer up, AI is getting bolted onto fee-for-service plumbing that rewards transaction volume, so it dutifully increases volume, prior-auth cycles and all, because that's the incentive it was handed. The technology isn't failing to produce outcomes. The technology is producing outcomes the buyer's own broken data infrastructure can't verify, reconcile, or trust.

That changes who wins in 2027. It won't be the vendor with the best clinical outcomes data in a glossy case study PDF. It'll be the vendor whose sales and implementation team showed up already fluent in the buyer's claims feed, their eligible-population definitions, their reconciliation cadence, because that vendor is the only one whose "proof" the buyer can actually audit without hiring data team they don't have. Most digital health sales organizations are still trained to sell outcomes. The buyers in this survey are telling you, in plain language, that they're actually buying auditability. Those are not the same skill, and almost nobody selling into this market right now has been taught the difference.

Curb moment: Everyone's telling founders to bring better proof. The buyers are telling you they can't even read the proof you already have.

Commercial Corner: What This Means for Your Sales Motion

Lead your next pitch with your data integration plan, not your outcomes deck. If you can't show, in the first conversation, exactly how your reporting reconciles against a payer's or employer's existing claims and utilization data, you're pitching proof the buyer structurally cannot verify and 45% of them will tell you that's their top barrier before you even ask.

Price your contract for a flat budget, not a growing one. With 56% of purchasers planning to hold spend steady, your growth has to come from displacing an existing vendor's renewal, not from a buyer's expanding discretionary pool. Know exactly which incumbent you're replacing and why your cost structure beats theirs on the metric purchasers weigh most.

Build your eligible-population and engagement definitions before the buyer asks. 44% of purchasers are actively narrowing who counts as "eligible" by diagnosis, and "active use of the platform", not registration, is now the dominant definition of engagement. A vendor who still reports enrollment numbers as a win is answering a question from two years ago.

If you're selling into a health system, ask whether your use case is one the CIO is already considering building in-house or absorbing into Epic or Oracle. Willingness to wait for a vendor roadmap collapsed from 52% to 22% in a year, that same willingness-to-wait is the willingness-to-build you're now competing against.

Track CMS's ACCESS model rollout in your target market even if you don't sell to Medicare FFS directly. It's a decade-long signal of where outcome-aligned payment is heading systemwide, and buyers who adopt its reporting discipline for one program tend to apply it to every vendor relationship they touch.

Curb moment: Founders keep building a better product. The buyers in this survey are telling you to build a better auditor.

SigCom Interpretation

Fourteen issues in, this newsletter has shown hospital VPs relabeling AI to dodge a budget gate, independent practices buying on gut feel because they have no gate at all, and pharma counting partnerships instead of outcomes. This issue is different because, for the first time, it's not one voice describing one organization's quirks. It's 321 buyers, independently surveyed, converging on the same three behaviors at once: pay for proof, hold the budget flat, and cut anyone who can't supply both. What the surrounding market data adds is the part the survey alone can't show you: the capital markets haven't caught up to this discipline yet, federal policy already has, and the buyers' own data infrastructure is often the real reason "proven ROI" dies in committee, not because the outcomes weren't real, but because nobody on the buyer's side could independently confirm them. That's not a sales objection you overcome with a better case study. It's a structural gap in the buyer's own house that most vendors never think to help fix.

The founders and sales leaders who come out ahead in this next cycle won't be the ones with the loudest outcomes claims. They'll be the ones who treat the buyer's broken data plumbing as part of the product they're selling because 321 purchasers just told an independent research institute, on the record, that proof nobody can verify isn't proof at all. It's just another pitch, and they're done buying pitches.

Curb moment: The market didn't get harder to sell into. It got harder to bluff.

Poll: WHERE ARE YOU EXPOSED ON THE PROOF TAX?

I've pitched outcomes data my buyer had no way to independently verify against their own claims or utilization systems

I don't know whether my target account is planning to grow, hold flat, or cut its digital health budget next year

I've lost a renewal to "cost of the solution" after being told our outcomes were strong

I've never asked a health system buyer whether they're considering building my use case in-house

The Lab Note

Next issue returns to one voice, one claim, tested the same way this newsletter has run since issue #7. If you're an executive, health plan, employer, health system, or government, with a claim about how your organization actually decides who gets paid, I want to talk to you. If you're a founder or sales leader: when a buyer asked you to prove an outcome, could your data actually survive their audit or did the conversation quietly end there?

Drop a comment or send a DM.