The Third Model.

The mental health benefits category has tried to fix its utilization problem through pricing innovation. The story of how it has tried — and where it has stopped short — is the most important commercial argument this report makes.

For most of the past two decades, the dominant model was Per Employee Per Month. The employer paid a flat fee per enrolled employee. The vendor collected the same revenue whether 5 percent of employees engaged or 50 percent. This model survived for a long time because it was predictable for procurement, but it carried a structural problem: the vendors with the most profitable economics were the ones whose products were used least.

The category eventually responded. Starting in the mid-2010s, the largest digital mental health vendors began shifting toward Fee For Service — pricing structures where the employer pays per session, per visit, or per unit of clinical care delivered. The shift worked on its own terms. The vendors that adopted FFS publicly report 20–25%+ utilization, versus the 2–5% range that had been the industry benchmark for over a decade. But the model introduced a new problem on a different dimension. Engagement risk moved onto the employer's books. A high-utilization quarter — driven by a seasonal campaign, a layoff event, or a wave of mental health awareness — could produce an unexpected, unbudgeted bill. The buyer was now paying for engagement, but in a structure that made them quietly fear it.

The category has therefore traded one misalignment for another. PEPM aligned the buyer's budget but not the vendor's incentives. FFS aligned the vendor's incentives but not the buyer's budget. Both models force a trade-off that benefits leaders have come to accept as inherent. It isn't.

This report walks through the structural argument for both prior models and proposes a third — Per Participant Per Month — which aligns vendor incentives with engagement while keeping per-user cost bounded for the buyer. PPPM is the model Thoughtful operates on. It is also the model that two major US benefits consultancies have told us, in private conversations, they expect to become more common in the category over the next several years, with particular momentum in the mid-market segment where the math of PEPM has always been most punishing.

The report does not argue that current vendors are operating in bad faith. It argues something narrower: the two pricing models the category has so far adopted both force a structural trade-off that no longer needs to be made. A third model exists. The buyers who shift to it early will have better unit economics, better data, and better vendor relationships than the buyers who wait.

Mental health benefits utilization has been a consistent disappointment to the people who pay for them. The numbers tell a stable story across two distinct pricing eras.

In conversations with HR leaders, the utilization data shows up the same way again and again. A leader who bought a benefit they believed in. Quarterly reports showing single-digit engagement. A renewal cycle where the data didn't change. A growing certainty that the benefit wasn't working — paired with uncertainty about what to do about it that didn't look like ripping out something promised to employees.

Their contract is satisfactory, but utilization is low — about 5 percent or less, which we were told is standard. Newer EAPs we looked at were eight times the cost, sometimes using the same therapist network. We prioritized function over appearance because of budget constraints.

This pattern is not unusual. Across the conversations Thoughtful's team has had with benefits leaders in the past year, it shows up nearly universally. The vendor is fine. The therapists in the network are good. The interface might be dated, or might not be. None of it explains the utilization gap. Newer entrants charging eight times more solve the interface problem without moving the engagement number.

Another conversation, this one with a benefits leader at a mid-market technology company: only 3 percent of registered users engage with their digital mental health vendor. The team has noticed an uptick in mental health leaves. They are questioning whether they should keep paying for a benefit that touches 3 percent of the workforce while the other 97 percent surface their distress elsewhere — most often, on a mental health leave the company pays for in a different ledger.

Why the category moved to Fee For Service

By the mid-2010s, the limitations of PEPM had become visible enough that buyers started pushing back. Employers paying flat fees for benefits with single-digit utilization began asking the obvious question: what are we actually paying for?

The largest digital mental health vendors in the category responded by shifting their pricing structures. Fee For Service became the new default — and quickly, the new consensus. Employers would now pay per session, per visit, or per unit of clinical care actually delivered. The model spread fast. Once one major player moved, the others followed within a few years.

On its own terms, the FFS shift was a meaningful improvement. Vendors were no longer rewarded for low engagement. Their revenue scaled with usage. Buyers could finally see a direct line between what they paid for and what employees received. The category had, by its own account, fixed its alignment problem.

Except it had only fixed half of it.

The new problem the new model created

Fee For Service pricing moves engagement risk from the vendor's books to the employer's. Under PEPM, a high-engagement quarter compressed the vendor's margin. Under FFS, that same high-engagement quarter produces an unexpected bill the employer has to absorb.

Consultants we spoke with for this report have a phrase for what this can produce: the Dry January problem. They described a client experience in which a substance use vendor on a usage-based contract saw a sharp engagement spike during Dry January — employees enrolled at higher rates, used the product more, and the vendor's revenue followed engagement up. The employer received an invoice substantially larger than they had budgeted for. The reaction was not gratitude that more employees engaged with the benefit. It was sticker shock.

Clients sometimes struggle with usage-based models because of unexpected costs from high utilization periods. The trend is shifting toward utilization-based pricing — but clients want to know the cost is bounded.

This isn't a one-time anomaly. It's the structural property of FFS. Any pricing model in which the employer's per-user cost is unbounded creates an incentive for the employer to quietly discourage utilization. The discouragement doesn't have to be explicit. It can be as subtle as not aggressively promoting the benefit. As soft as delaying renewals to absorb a budget shortfall. As indirect as choosing vendors with smaller session caps over vendors with larger ones.

The result is that FFS has produced a category where vendor incentives and buyer incentives are now misaligned in the opposite direction from PEPM. The vendor wants engagement up. The buyer, in practice, quietly wants engagement bounded. The benefit sits in the middle of that tension.

Consultants we spoke with describe FFS-era utilization as generally higher than PEPM-era utilization. But it's still meaningfully below the engagement rate buyers say they want — and there's a structural reason for that. The buyer cannot fully champion engagement when each unit of engagement adds variable cost. The model improved one dimension and broke another.

The structural economics of PEPM and FFS each produce a different misalignment. Walking through both makes clear why the trade-off the category has accepted is not inherent — and why a third model can resolve it.

PEPM: vendor wins when usage stays low

Under Per Employee Per Month pricing, the employer pays a fixed fee per enrolled employee. A 5,000-employee company at $3 PEPM pays $15,000 monthly. That number does not change if 50 employees engage or 500.

The math on the vendor side runs as follows. Fixed platform costs spread across all enrolled employees regardless of usage. Variable costs — provider time, support, crisis escalation, clinical oversight — only get incurred when an employee engages. At 5 percent engagement, the vendor collects fees from 100 percent and pays variable costs on 5 percent. Margin is structurally high. At 50 percent engagement, the vendor collects the same fees but pays variable costs on ten times as many users. Margin compresses significantly.

The vendor with the most profitable economics is the one whose product is used least. Not because of bad faith — because the pricing structure mathematically rewards that outcome. A rational vendor on PEPM under-invests in engagement, prices expansion modules conservatively, resists engagement-linked outcome commitments, and favours broad enrolment over deep usage. None of these decisions are malicious. Each makes sense given the unit economics. Together, they produce the industry-wide pattern of high enrolment and durable single-digit utilization.

Fee For Service: buyer absorbs engagement risk

Fee For Service flips the misalignment to the other side of the table. The employer pays per session, per visit, or per unit of clinical care delivered. The vendor's revenue scales directly with engagement. The PEPM problem is resolved.

The numbers bear this out. The largest digital mental health vendors in the category publicly report 20–25%+ utilization in their own marketing materials, versus the 2–5% range that traditional EAPs have sustained for over a decade. Some of these vendors cite utilization increases of up to 600% for customers migrating from legacy EAP contracts. By the engagement metric alone, the FFS shift has been a genuine improvement.

But the model produces a new problem for the buyer. The annual cost of the program is no longer known at contract signing — it's a function of how much the workforce uses the benefit. A high-engagement quarter produces a higher bill. An awareness campaign that successfully drives engagement produces an invoice the CFO hadn't planned for. Layoff events, restructuring, and seasonal stress periods all become unbudgeted expense risks.

The behavioural consequences are subtle but real. A benefits leader on FFS has a structural reason to be quietly cautious about engagement. They want the program used, but only so much. They want awareness up, but only so much. The benefit sits in a strange place: nominally promoted, but practically rate-limited by the budget anxiety the pricing model creates.

Higher engagement is good. Unpredictable cost is hard to defend in a budget review. The category traded one misalignment for another.

The trade-off the category has accepted

PEPM trades engagement alignment for budget predictability. FFS trades budget predictability for engagement alignment. Each addresses one half of what the buyer needs and ignores the other. Benefits leaders have come to treat this trade-off as inherent. It isn't. It's an artefact of the two specific models the category has so far tried.

The sequence playing out in mental health benefits is not unique. The most directly parallel case is in healthcare itself, which walked through nearly the same three-step arc over the past century — and the resolution it landed on tells us where this category is going next.

For much of the twentieth century, payers and providers used capitation models — a fixed per-member-per-month fee regardless of services delivered. Capitation produced the same problem PEPM produces in benefits: providers were paid the same whether patients received care or not. The financial incentive ran toward under-treatment.

Fee-for-service replaced capitation as the consensus model and held through most of the twentieth century. It fixed the under-treatment incentive: providers were now paid for each service delivered. But it produced its own problem. The provider's revenue scaled with volume, not with health outcomes. Payers found themselves spending more for more services without proportional improvements in population health. Engagement risk lived on the payer's books.

By the 2010s, the misalignment in fee-for-service had become impossible to ignore. The response was value-based care. Beginning with Medicare and spreading to commercial insurers, payers moved toward contracts that paid providers based on outcomes — bundled payments, shared-savings models, capitated arrangements with outcome accountability. The new model resolved both prior problems: providers were aligned with patient outcomes rather than visit volume, and payers had cost predictability through bundled and capped structures.

The mental health benefits category is currently in step two of this sequence. Starting in the mid-2010s, most major vendors moved from PEPM into FFS. The limitations of FFS are now becoming visible — engagement risk on the buyer's books, the budget anxiety it produces, the structural ceiling it places on how much engagement buyers can champion. The next step in the sequence is one that combines vendor alignment with cost predictability. In healthcare it took roughly a decade to mainstream. In mental health benefits it has already begun.

The model that resolves both prior trade-offs is Per Participant Per Month, with unlimited usage included in the participant fee. The structure works like this: the employer pays only for employees who actually participate in the benefit — typically defined as activating their account or completing a baseline level of engagement — within a given month. Once an employee triggers the participant fee for that month, all usage that month is included. There is no per-session, per-visit, or per-hour additional charge.

This single structural choice does several things at once.

It fixes vendor alignment. The vendor's revenue scales with engaged users, not with nominal enrolment. A vendor who can't engage employees doesn't get paid. The PEPM problem is gone.

It fixes buyer cost-unpredictability. The employer's exposure per engaged user is bounded by the participant fee. A user who has one session and a user who has thirty sessions cost the employer the same in a PPPM month. The Dry January problem is gone.

It introduces a useful new metric. The employer's cost per engaged employee is the participant fee directly. No engagement modelling required. The number you see is the number you pay.

The trend is shifting toward utilization-based pricing — but clients want to know the cost is bounded. PPPM is particularly compelling for mid-market employers, where the math of flat fee structures has always been most punishing.

Two of the major US benefits consultancies have told us, in private conversations, that they expect PPPM-style structures to become more common in this category over the next 18 to 36 months, with particular momentum in the mid-market segment. The reasoning on mid-market matters.

Why mid-market is where PPPM lands first

Mid-market employers — companies in the 500 to 5,000 employee range — face a sharper version of the PEPM problem than either enterprise or small-business buyers. At enterprise scale, a low single-digit utilization rate still produces a meaningful absolute number of engaged employees that the program team can point to. At small-business scale, the per-employee fee is low enough that the absolute spend doesn't draw scrutiny. Mid-market sits in the middle: enough employees that the program is a real budget line, not so many that low utilization produces a defensible number of engaged users.

Mid-market HR teams are also more open to pricing innovation than enterprise procurement. Mid-market buyers don't typically have an established RFP template that excludes non-PEPM structures. They make decisions faster. They feel the engagement-or-cost trade-off more acutely because their budgets have less room for either side of it to go wrong.

The result is that mid-market is where PPPM is going to get adopted first, and where the proof-points for the model will be visible earliest. Enterprise will follow once the mid-market data shows that PPPM produces better cost-per-engaged-employee numbers than either PEPM or FFS.

How PPPM resolves the remaining objections

Two concerns about engagement-aligned pricing have historically slowed adoption. Each is addressable with model design.

Budget predictability is solved through forecasting and budget envelopes. PPPM contracts can include expected participation rates and contracted budget ranges. A buyer planning for 25 percent participation in year one can model that as their expected spend. Variance from that number is visible and reviewable; it isn't a surprise. Some vendors offer participation floors and ceilings as part of the contract structure, giving the buyer a known maximum exposure. The CFO conversation becomes more honest, not harder, because the budget reflects expected engagement rather than enrolment fiction.

The engagement-risk problem is solved by including unlimited usage in the participant fee — which is what distinguishes PPPM from straight FFS. Once an employee engages and triggers the participant fee for the month, additional engagement from that same employee costs the employer nothing additional. A user who has one session and a user who has thirty sessions cost the employer the same. Engagement spikes don't produce sticker shock because per-user cost is capped at the participant fee.

What changes for the buyer

On a PPPM contract, the buyer's primary metric stops being utilization in the abstract and becomes cost per engaged employee. This is a more honest number, and it's the number that maps to the outcome conversations CFOs actually care about.

A traditional benefit at low single-digit utilization is not actually cheap when you do the math per engaged employee. If utilization sits at 5 percent, the real cost per engaged employee is roughly twenty times the headline PEPM rate — the same math runs in any direction the utilization rate moves. A PPPM benefit priced at five times the PEPM headline rate, if it reaches 25 percent engagement, has a meaningfully lower cost per engaged employee than the PEPM benefit at 5 percent utilization. This is illustrative math, not a benchmark — actual numbers depend on the specific vendors and contract structures. But the pattern is robust. The intuitive read of pricing in this category often reverses once you express cost per engaged employee.

What changes for the vendor

On a PPPM contract with unlimited usage included, the vendor's commercial interest aligns precisely with what buyers want: more engaged employees, with no incentive to ration usage per engaged employee. Every additional engaged employee is revenue. Every disengaged employee is missed revenue. Every user who increases their usage costs the vendor more to serve but produces zero additional revenue — which means the vendor is structurally motivated to make engagement productive, not just frequent. The vendor's incentives now run toward driving meaningful engagement broadly, while ensuring per-user economics work through better product design rather than usage limits.

Most benefits leaders reading this report are not on PEPM anymore. They're on FFS, or on a hybrid structure their largest vendor has been moving them toward over the past few years. The practical recommendations therefore look different depending on where you sit. Some guidance for the next twelve months.

If you're still on PEPM

In your next renewal conversation, ask your current vendor a question they likely haven't been asked before: "At what utilization rate would your unit economics become unsustainable for you?" Most vendors haven't framed the question this way internally — and the conversation that follows tells you a lot about how their pricing model is structured, and which model will serve you better over the next contract cycle.

If you're on Fee For Service

The question to ask your vendor is different: "What's our cost trajectory if engagement doubles?" Most FFS contracts will produce a number that gives the CFO pause. Then ask the harder follow-up: "How would you price this if our per-user cost were capped at a flat monthly rate, regardless of usage?" Vendors who can answer this question have already started thinking in PPPM terms. The conversation reveals how much engagement-risk exposure is in your current contract — something many buyers haven't fully modelled.

Run the cost-per-engaged-employee math, on every model

When evaluating any mental health vendor, calculate cost per engaged employee — not the headline PEPM, not the per-session FFS rate. Take the total expected contract value, divide by realistic engagement projections, and look at the resulting number. This is the number the CFO will eventually be asked. Better to know it than to discover it during a budget review.

A low PEPM with a low engagement rate is rarely cheaper than a higher PPPM with a higher engagement rate. A low FFS rate with high session caps may be cheaper than a flat PPPM. A higher PPPM with unlimited usage may be cheaper than an FFS structure where heavy users drive the bill. The intuitive read of pricing in this category often reverses under scrutiny. The cost-per-engaged-employee number is the equaliser.

If the historical pattern from healthcare holds, third-model pricing structures will become the consensus in mental health benefits over the next three to five years. The shift is starting in mid-market and will work upward. Buyers who move early will have better data and better unit economics than buyers who wait. The shift is not optional in the long term — only in the short term, and that option has a value.

This report has made a category-level argument. It would be incomplete without naming where Thoughtful sits in the picture, since Thoughtful published the report and operates on the model the report recommends.

Thoughtful's pricing structure is Per Participant Per Month, with unlimited usage included in the participant fee. Customers pay only for employees who engage with the product in a given month. Each participant fee includes unlimited AI sessions and clinically-overseen safety escalation. In the Thoughtful Complete tier, the participant fee also includes unlimited provider sessions and unlimited safety escalations. There are no per-session fees. There are no usage caps within a participant month. The model is the third model this report describes.

This is not a marketing choice. It's the choice that aligns Thoughtful's commercial incentives with the buyer's stated outcomes. When customers pay only for engaged users, Thoughtful's product priorities reorganize around engagement. When usage is unlimited within the participant fee, Thoughtful has to build a product that produces meaningful outcomes per session — because we can't margin our way out of low-value engagement by capping it. Onboarding becomes a critical investment because users who never activate never pay. Retention becomes the metric that defines the business.

Thoughtful did not invent the structural argument in this report. Pricing innovation has happened in other categories before, and the path the mental health benefits category is going to walk has been walked elsewhere. Thoughtful's contribution is to be one of the first vendors in mental health benefits to operate on the third model, and to operate it in a way that takes advantage of the alignment the model creates.

The argument of this report would stand without Thoughtful in it. The category needs a different pricing model. The buyers who shift early will benefit from it. Thoughtful is one example of what the next model looks like in practice. There will be more.

This report combines secondary research on benefits industry utilization data, qualitative primary insight from Thoughtful's conversations with benefits leaders and consultants over the past year, and historical analog research on the pricing-model shift in healthcare from capitation through fee-for-service to value-based care.

Primary sources

Conversations with benefits leaders at multiple US-based and global employers, ranging from 500 to 50,000+ employees, conducted between January and May 2026. Insights are aggregated and anonymised; specific buyer quotes are paraphrased with permission and details adjusted to protect identifying context.

Conversations with senior benefits consultants at two of the major US benefits consultancies, conducted in dedicated deeper-dive sessions in early 2026. Consultant quotations are anonymised by firm and individual at consultant request.

Secondary sources

Industry utilization data: SHRM Employee Benefits Survey, Mercer National Survey of Employer-Sponsored Health Plans, National Business Group on Health (NBGH) Large Employers' Health Care Strategy and Plan Design Survey, Willis Towers Watson Best Practices in Health Care Survey, Lyra Health State of Workforce Mental Health Report.

FFS-era utilization figures: Publicly available marketing materials and case studies published by the largest digital mental health vendors in the category. Numbers are reported as the vendors themselves describe them; this report does not independently verify them but treats them as the vendors' own published benchmarks for the FFS pricing era.

Healthcare pricing-shift analog: CMS Innovation Center reports on Accountable Care Organizations, Health Affairs research on value-based care contracts, NEJM Catalyst case studies on shared-savings programs and capitated arrangements.

What this report does not claim

This report does not claim that any specific vendor in the mental health benefits category is operating in bad faith. The structural argument applies to incentive alignment under particular pricing models, not to vendor intent. Vendors operating on PEPM or FFS structures are operating rationally within the incentive structures their pricing creates. The argument is about those structures, not about individual companies.

This report does not name specific competitors by company. The category dynamics described — the PEPM era, the shift to FFS led by the largest digital mental health vendors in the mid-2010s — are public knowledge to anyone in the industry. We've kept the framing structural rather than naming specific players to keep the argument focused on the economics rather than on inter-vendor disputes.

This report does not claim that PPPM is universally superior to PEPM or FFS in every benefit category. The argument is specific to engagement-driven outcome categories, of which mental health is one. Other categories may have different optimal pricing structures.

This report does not claim that Thoughtful's pricing model is the only viable third model. Other engagement-aligned structures with cost predictability built in are possible. Thoughtful's contribution is to be early to the shift in this specific category.