The math on startup health insurance flips around 50 employees and most founders don’t see it coming. The PEO that was the obvious choice at 15 employees — large-group rates, simple administration, one vendor — quietly becomes the most expensive option in the stack. By the time the renewal notice arrives showing a 22% premium increase, the founder is staring at a six-figure annualized cost and asking whether the alternative would have been cheaper.
Yes. Usually. By a lot.
There are three real structures for startup health insurance: a Professional Employer Organization (PEO), a direct group plan through a broker, or individual stipends (typically through an ICHRA). Each one is the right answer at a specific stage and the wrong answer outside it. This post is the operator’s comparison — what each structure actually does, what it actually costs per employee, and the signals that say you’ve outgrown one and need to migrate to the next.
The three structures at a glance
| Structure | Right for | Per-employee cost (employer side) | Setup time |
|---|---|---|---|
| PEO | 5-50 employees, low-state-count | ~$150-200/employee/month plus benefits cost | 2-4 weeks |
| Direct group plan via broker | 25+ employees, any footprint | Plan cost only; no platform markup | 4-8 weeks |
| ICHRA / individual stipends | Remote teams that want to opt out of group plans entirely | $400-700/employee/month stipend | 2-4 weeks |
The structural choice frames every other benefits decision — plan design, dependent contribution policy, broker relationships, payroll integration. Get this right and the rest of the benefits stack falls into place. Get it wrong and you’ll be unwinding the choice 18 months later.
Option 1: PEO (Professional Employer Organization)
A PEO co-employs your team. Legally, your employees are co-employed by the PEO for tax and benefits purposes; the PEO becomes the employer of record for benefits enrollment, which lets them pool your headcount with their other clients’ headcount and access large-group benefits rates.
The practical effect: a 20-person startup gets benefits priced as if it were part of a 50,000-employee pool. That’s a meaningful improvement over what a 20-person startup could negotiate on its own.
The two PEOs we see most often are Justworks (modern, founder-friendly, transparent pricing) and TriNet (older, broader, more enterprise-leaning).
How the cost stacks up. PEO fees are charged on top of the actual benefits cost. The fee structure is typically either a percentage of payroll (2-4% is common) or a flat per-employee-per-month rate ($150-200 PEPM). For a 30-person startup with average salary of $150K, that’s roughly $135K-180K per year in PEO fees, on top of the underlying benefits cost.
The benefits cost itself sits on top of the PEO fee. Medical/dental/vision through a PEO typically runs $500-800 per employee per month employer-paid, comparable to what you’d pay through a direct group plan.
When PEOs work. 5-50 employees. Single state or low state count. Founders or ops leads who want to outsource HR compliance and benefits administration without thinking about it. Companies that don’t yet have an in-house HR function and don’t want to build one immediately.
When PEOs stop working. Above ~50 employees, when the PEO fees start meaningfully exceeding what you’d pay for direct administration. Above 75 employees, almost always. Companies with multi-state complexity that exceeds what their PEO is set up to handle. Companies that want to integrate cleanly with a modern payroll stack — PEOs are not usually compatible with running Rippling or Gusto in parallel, because the PEO is the payroll provider.
The migration trigger. When the all-in PEO cost (fees + benefits) starts approaching $30K+ per employee per year, and your headcount is over 40, run the numbers on a direct group plan. Most companies discover they can save $1,500-3,000 per employee per year by migrating, which on a 50-person team is $75K-150K in annual savings — enough to fund the in-house HR hire that the migration requires.
Option 2: Direct group plan via a broker
You contract directly with insurance carriers — typically through a benefits broker that aggregates carrier options, negotiates renewals, and supports open enrollment. Benefits run through your own payroll stack (Rippling, Gusto), and you administer the day-to-day directly.
The two brokers we recommend most often are Newfront (modern, tech-forward, strong for venture-backed companies) and Sequoia (tier-one alternative, particularly strong for later-stage and pre-IPO transitions). Regional brokers can be the right answer for geography-specific or industry-specific needs; ask for references and for their actual client roster at your size before committing.
How the cost stacks up. No platform fee. You pay only the underlying benefits cost. Medical/dental/vision typically runs $500-900 per employee per month for the employer portion, depending on plan generosity, geography, and demographics. Dependent coverage is on top of that — usually a defined employer contribution to the dependent premium (50-75% is common), with the employee covering the rest.
The broker compensation question. Brokers are typically paid commission by the insurance carrier, not by you. This is fine — it’s the standard structure — but you should understand it. The commission is usually 3-7% of the premium, baked into the rate. A broker recommending a more expensive plan structure could in theory be earning more commission on it. The mitigation: ask your broker how their commission works on each plan they’re proposing, and ask whether they’d recommend the same plan if they were paid a flat fee.
When direct group plans work. 25+ employees. Multi-state footprint (modern brokers handle this fluently). You have a Rippling or Gusto stack that handles benefit deductions cleanly. You want full control over plan design — which plans to offer, what to contribute, how to structure dependent coverage.
When direct group plans don’t work. Below ~15 employees, when most carriers won’t quote you competitive rates because you’re too small. Below ~25 employees, often, when the PEO is still cheaper after accounting for administrative time. When you genuinely don’t have or want an HR function — the direct plan requires somebody to own the broker relationship, manage open enrollment, and handle benefits administration through payroll.
Setup expectations. A direct group plan migration takes 6-10 weeks from broker selection to first benefits-effective payroll. You’ll go through a discovery process (current plan, demographics, geography, budget), the broker shops the market and presents quotes, you select plans, the broker handles enrollment, and you configure deductions in your payroll stack. Open enrollment for the first plan year runs alongside this.
Option 3: ICHRA and individual stipends
You give employees a fixed monthly amount (typically $400-700) to buy their own coverage on the individual marketplace. An ICHRA — Individual Coverage Health Reimbursement Arrangement — is the tax-advantaged structure for doing this. Stipends paid outside an ICHRA are generally taxable to the employee.
How the cost stacks up. Employer cost is the stipend amount, plus minimal administration fees from an ICHRA platform if you use one. Total cost is typically $5,000-8,500 per employee per year — meaningfully lower than group plans on paper.
When ICHRAs work. Fully-remote teams across many states where group plans are operationally complicated. Companies with very small headcount (under 10) where group plan rates are non-competitive. Companies whose employees skew young, healthy, and able to find better individual-market coverage than the group rate provides.
When ICHRAs don’t work. Most startups, honestly. Two problems:
The first is recruiting. “We don’t offer health insurance; we give you money to buy your own” is harder to recruit against than “we offer Blue Cross PPO with $50K employee premium covered.” Even when the math works out the same, candidates read the latter as a real benefit and the former as an abdication.
The second is employee experience. Group plans have negotiating power and a unified network. Individual marketplace plans have higher deductibles, narrower networks in some states, and the administrative load of selection and enrollment falls on the employee. Employees with families and ongoing health needs often end up worse off under ICHRA, even with a generous stipend.
We recommend ICHRA only when the specific use case justifies it — typically a fully-remote company with 10-20 employees scattered across many states where group plan logistics are genuinely intractable, and where the team has agreed in advance that this is the structure.
The migration: when and how
The most common migration is PEO to direct group plan. The trigger is typically headcount (40+) plus the renewal notice that shows the PEO economics no longer working.
Timeline. Budget 4-6 months from decision to live. You’ll need:
- A broker selection (2-4 weeks)
- A new payroll provider if migrating off the PEO’s bundled payroll (Rippling or Gusto, see the payroll administration post for the migration mechanics)
- A new HRIS, ATS, and 401(k) provider if the PEO was bundling those — Guideline for 401(k), Rippling or a standalone ATS for hiring
- Benefits selection and enrollment for the first non-PEO plan year
- Coordination with employment counsel on the legal change in employment structure
Cost. Net cost should be negative — the PEO fee savings will exceed the migration cost. Migration cost itself typically runs $15K-50K in one-time expenses depending on complexity (broker setup, payroll migration, counsel time, internal time).
Sequence. Stand up the new stack in parallel with the PEO. Run the first full payroll cycle through the new stack while the PEO is still active. Cut over at a clean month-end, with the new benefits effective the following month. Don’t try to migrate off the PEO before benefits are confirmed live in the new stack — the gap is what creates exposure.
What to ask your broker before you sign
The broker selection is more consequential than founders realize. Five questions that separate the brokers worth working with from the ones to avoid:
Who specifically will be our day-to-day contact, and what’s their book of business? You want a named individual, not a “team.” You want to know if you’re going to be a small account on a 50-account book or a meaningful account on a 12-account book.
How many of your current clients are in our stage and our industry? Specifically: how many venture-backed Series A/B companies of 25-75 employees do you currently serve? A broker that primarily serves established companies will not know how to navigate the early-stage benefits decisions.
Walk me through your renewal process. A good broker starts the renewal conversation 90 days before the policy anniversary, with market data, plan alternatives, and a clear recommendation. A bad broker emails you the renewal notice 30 days before, with the carrier’s price increase, and asks if you want to accept.
How do you handle multi-state issues? Specifically: which carriers do you have strong relationships with in California, New York, Texas, and the other states where we have or will have employees? Multi-state benefits coordination is real work and a broker who can’t articulate their approach to it is not the right broker.
Can you give me three references from clients at our stage? Talk to those references unfiltered.
Plan design — the variables that actually matter
Once you’ve picked the structure, the plan design conversation centers on three variables.
Plan tier. PPO (more expensive, lower deductible, broader network) vs. HDHP-HSA (lower premium, higher deductible, paired with a Health Savings Account). Offering both gives employees agency. Offering only HDHP-HSA saves money but reads as cost-conscious to candidates.
Employer contribution percentage. The benchmark for Series A/B startups is 80-100% of the employee premium and 50-75% of dependent premiums. Below 80% on the employee side starts to become a recruiting issue. 100% is generous and increasingly common in competitive markets.
Dependent contribution policy. This is the biggest cost variable hidden inside the benefits decision. A single employee might cost the company $700/month for medical. The same employee with a spouse and two kids costs $2,200-2,500/month. The dependent contribution policy you set has a much larger impact on total benefits cost than the plan tier you pick. We see most Series A/B startups land at 50-75% employer contribution to dependent premiums; below that becomes a recruiting issue with parents.
When to bring in operator support
You probably don’t need outside help if you’re under 25 employees on a PEO with no immediate scaling pressure. The decision is simple.
You likely want operator input if:
- You’re 25-50 employees, on a PEO, and trying to decide whether to migrate to direct group
- You’re evaluating brokers and want a second opinion on the proposals you’re seeing
- You’re a fully-remote company considering ICHRA and want to make sure the recruiting tradeoff is understood
- You’re preparing for a Series B raise and want benefits to read as competitive to senior hires
- You’re inheriting a benefits stack and aren’t sure what’s broken
Pegacorn Group works with venture-backed Series A and B startups on benefits structure and migration. If you’re trying to make this decision and want a clear-eyed read on what’s right for your stage, let’s talk.
This post pairs with: The Series A/B benefits stack, How to administer payroll and 401(k) plans at a startup, Equity compensation 101, and Multi-state employment compliance.