The Coordination Tax: What It Costs to Run Your Own Growth Engine

TL;DR

Our own website says that the traditional route to a working global growth engine costs $200,000–$300,000 or more a year, and that our $24,000 package is roughly 90% less. That is a marketing claim. This report tests it — using US federal wage data, published vendor rates, and arithmetic we show in full.

• Rebuilt from Bureau of Labor Statistics medians, a realistic in-house configuration costs $220,888 a year. That is inside our published range, in its lower half 12.
• The number in most founders' heads is missing about a third. BLS puts benefits at 30.1% of private-industry compensation, so an employer's real cost is the salary divided by 0.699 — a 1.431× multiplier on every figure you have been quoting yourself 2.
• Hiring one person for each of the seven roles would cost $869,127 a year fully loaded 12. Nobody does that, which is precisely why the in-house route usually fails quietly rather than loudly.
• Twelve press releases cost more to distribute than the whole package costs. At published national-distribution rates a 1,000-word release runs about $2,155; twelve plus membership is $26,055 3.
• And here is where the comparison is weakest, stated in section 7 rather than buried: most companies never spend $220,888. They assign the work in fractions to people who already have jobs, and get partial output. $24,000 does not buy 1.5 full-time equivalents of attention — it buys a productised system, and the deep regulatory work stays with you.

If you take one thing from this piece, make it the worksheet in section 9 rather than our number. Run the arithmetic on your own roles and your own market, and you will have something better than any vendor's comparison slide — including ours.

---

1. Why we are auditing our own marketing claim

Takeaway: a comparison figure you cannot check is a comparison figure you should not believe. Ours has been checkable in principle and unchecked in practice. This fixes that.

Every outsourced-growth vendor, us included, puts a version of the same slide in front of a prospect: here is what doing this yourself costs, here is what we cost, look at the gap. The slide is almost never sourced. The costs are usually salaries pulled from a job board, benefits are usually forgotten, and the vendor's own number is the only one on the page with any precision behind it.

We publish a version of that claim. Our website states that the comparable traditional route runs $200,000–$300,000+ per year, and that our Done-for-You subscription at $24,000 per year is roughly 90% less 4.

There are two honest things to do with a claim like that. One is to keep saying it. The other is to rebuild it from data anyone can check, publish the arithmetic, and say clearly where it holds and where it does not. This report does the second.

The standard we are applying to ourselves is the one a buyer should apply to every vendor in this category, and the practical value of the piece is not our conclusion. It is the method — which is in section 9, and which will produce a different and more useful number for your company than ours.

---

2. The number in your head is missing a third

Takeaway: benefits are 30.1% of private-industry compensation. Every salary figure in your build-versus-buy model is roughly 43% too low as an employer cost.

Start with the most common modelling error, because it distorts everything downstream.

When a founder estimates what a marketing hire costs, the number that comes to mind is the salary. The number that lands on the company's books is total compensation, and the gap is not small.

The Bureau of Labor Statistics measures this directly. In its Employer Costs for Employee Compensation release for March 2026, employer compensation costs for private industry workers averaged $46.60 per hour worked: wages and salaries were $32.60 (69.9%) and total benefits $14.01 (30.1%) 2.

The operational consequence is a single multiplier. Divide the salary by 0.699 — or multiply by 1.431 — and you have the employer's cost. A $100,000 salary is a $143,100 line item.

Two notes before we use it. Benefits here include paid leave, insurance, and legally required contributions such as Social Security, Medicare and unemployment insurance. And for civilian workers overall the benefit share is slightly higher at 31.6%; we use the private-industry figure throughout because that is what a commercial life-sciences company is 2.

---

3. What the seven roles cost

Takeaway: the six services in a growth engine map to seven distinct occupations, and BLS publishes what each of them is paid.

A global growth engine is not one job. Building and running a website, keeping it operating, managing a LinkedIn presence, holding a domain portfolio, producing a hundred-plus source-grounded articles a year, and putting out a monthly press release require different skills, and the labour market prices them separately.

Here is what the market pays, using BLS median annual wages for May 2024 1:

Every one of these roles pays above the all-occupations median. That is worth noticing on its own: there is no cheap corner of this function.

Now apply the multiplier from section 2.

That $869,127 is not a strawman we are about to knock down — it is the honest upper bound, and it is useful for one reason. The gap between it and the salary sum of $607,520 is the benefits load most models forget. The gap between it and what companies actually spend is the subject of the next section, and it is where the real story lives.

---

4. What a real configuration looks like

Takeaway: nobody hires seven people. They assign fractions — and a properly staffed fractional model still costs $190,833 a year in labour.

No life-sciences company with one cleared product hires a marketing manager, a web developer, a writer, an editor, a PR specialist, a designer and a marketing specialist to run its digital presence. What actually happens is that each role is filled fractionally: some hours from a person who already has another job, some from a contractor, some from the founder at eleven at night.

So the honest model is fractional. Here is a configuration that would genuinely deliver the six functions — not brilliantly, but competently:

Two of those lines deserve defending, because they are where a sceptical reader will push.

The writer at 0.5 FTE. Our own content engine produces 110–180 source-grounded articles a year 4. Even at the low end that is roughly two a week, every week, each one researched against primary sources and written for a regulated-market audience. Half a full-time writer is not a generous allocation for that; it is a tight one.

The marketing manager at 0.25 FTE. This is the coordination line, and it is the one most models omit entirely. Somebody has to brief the writer, approve the design, chase the developer, schedule the release, and make sure the LinkedIn post and the website page and the press release describe the same product the same way. That person is not producing anything. They are preventing the six pieces from drifting apart — which is the coordination tax this report is named for, and section 8 is about what happens when nobody pays it.

---

5. The bills that are not salaries

Takeaway: distributing twelve press releases costs more than the entire annual package this report is comparing against.

Labour is most of the cost but not all of it. Two non-salary lines matter enough to model.

Wire distribution. Published rates for PR Newswire national distribution run about $805 for the first 400 words, plus a per-100-word charge that 2026 rate guides state inconsistently — $225 to $275 — on top of an annual membership of $195–$249. Multimedia items add about $325 each and a company logo about $495 3.

A 1,000-word national press release is therefore $2,155 to $2,455 before you add a single image. Twelve of those, plus membership, is $26,055 to $29,655 a year. We carry the lower figure through the rest of this report, because it is the conservative choice: it makes the in-house total smaller and our own comparison claim harder to support, not easier.

Sit with that number for a moment, because it is the most quotable line in this report: the wire distribution alone, for one release a month, costs more than the entire $24,000 annual package that includes those releases plus five other services — and that holds at either end of the published rate range.

Infrastructure. Domains across a global brand strategy, hosting, uptime monitoring, backups and security tooling. We have modelled $4,000 a year. Unlike every other figure in this report, that one is an estimate with no primary source behind it, and section 7 says so again.

---

6. The result, against our own claim

Takeaway: $220,888. The claim on our website survives the audit — at the lower half of the range it states.

At the upper end of the published wire rate the total is $224,488 rather than $220,888. Nothing in the conclusion changes.

Our website says $200,000–$300,000+ and "roughly 90% less" 4. The rebuild lands at $220,888 and 89.1%.

The claim holds. It is also worth saying what kind of holding this is: our number sits in the lower half of the range we publish, which means the top of our own range — the $300,000+ end — describes a more generously staffed configuration than the one we modelled here, not the median case. If we were optimising this report for sales we would have modelled the generous configuration and reported a bigger gap. We modelled the tight one.

---

6b. Who this decision is actually in front of

Takeaway: the companies facing this choice are overwhelmingly single-product, first-time entrants with no existing growth function to extend — which is exactly the profile for which the in-house arithmetic works worst.

It is worth knowing who is making this decision, because the answer changes depending on what a company already has.

In our previous report we analysed the 2026 post-clearance cohort from the FDA record. Through 27 June 2026 there were 1,647 K-number decisions, including 869 rows from non-US applicants. After conservative name normalisation those resolve to 645 non-US applicant entities, of which 511 had a single 2026 clearance and 201 appear to be receiving their first FDA clearance ever 5.

Read those proportions against the build-versus-buy question.

Most of this cohort has one product, not a portfolio. A company with a single cleared device does not have enough recurring work to justify a full-time marketing manager, let alone seven roles. Every line in the section 4 configuration is a fraction — which is exactly the situation in which fractional hiring fails, because 0.10 FTE of an editor is not a job anyone takes.

Most of them are entering the US market for the first time. A first-time entrant has no existing English-language commercial surface to extend, no established agency relationships, and no internal precedent for what "done" looks like. The in-house route asks them to build a function and a product launch simultaneously.

And the work is front-loaded exactly when the company is busiest. Clearance is not the end of the regulatory workload; it is the start of the commercial one, arriving while the team is still handling post-clearance obligations. The 0.25 FTE coordination line in section 4 is, in practice, usually the founder — during the quarter they have least to give.

None of this makes a managed subscription automatically right. It does mean that for the modal company in this cohort, the honest in-house comparison is not $220,888 against $24,000. It is what actually gets done by a busy team of two against $24,000 — and that comparison is decided on output, not on cost.

---

6c. Three routes, honestly compared

Takeaway: cost is the least interesting axis. The routes differ most in who carries coordination, and what happens when a person leaves.

We sell the third column, and we have an obvious interest in how that table reads. So note what it says against us: if you already have a marketing team and one market to serve, extending what you have is usually the better answer, and we will say so on a call. The third column earns its place when there is no function to extend and more than one market to reach — which, per section 6b, describes most of the current cohort but not all of it.

---

7. Where this arithmetic is wrong

Takeaway: five things that could make this comparison mislead you. A number you cannot argue with is a number you should not trust.

Most companies never spend $220,888. This is the honest one, and it is the most important. The figure is what doing the job properly in-house costs. What actually happens at most small life-sciences companies is that the work is assigned in fractions to people who already have full jobs, nothing is anyone's primary responsibility, and the output is correspondingly partial — a website that was current at launch, a LinkedIn page that posts when someone remembers, a content plan that survived two quarters. The real comparison for many readers is not $220,888 against $24,000. It is partial output at low cash cost against complete output at $24,000, and that is a different argument than the one the arithmetic makes.

The agency route is not modelled. Appointing three or four specialist agencies sits between hiring and a managed subscription on cost, and we have no primary rate data we would stand behind. Its distinguishing feature is not price anyway — it is that the coordination tax in section 8 gets larger, not smaller, because now the calendars belong to other companies.

$24,000 does not buy 1.5 FTE of attention. It buys a productised system running on shared capacity. That is the honest description of how any subscription at this price works. Work that is genuinely deep and client-specific — regulatory review, clinical claims, medical-legal-regulatory sign-off — sits outside it and stays with you. Our own content follows the client's review path; it does not replace it 4.

The $4,000 infrastructure line is an estimate. Every other figure here traces to BLS or a published rate card. That one traces to judgement.

These are US wages. A company hiring in a lower-cost labour market has a completely different arithmetic and should redo it with local data. The method transfers; the numbers do not.

---

8. The part that is not a number

Takeaway: the coordination tax is not in any of these tables, and for most companies it is the part that actually decides the outcome.

Everything above prices capacity. None of it prices coherence, and coherence is what a growth engine actually produces.

Consider what has to line up for a single product announcement to work. The website product page has to describe the device using the language the regulatory decision actually supports. The press release has to say the same thing, in a form a wire service will accept, on a date that makes sense. The LinkedIn post has to point at the page, not at a PDF. The article published that month has to reference the announcement rather than contradict it. The domain the release links to has to resolve, hold a valid certificate, and be the one the brand strategy says is canonical.

Six pieces. In the in-house model, six different part-time attentions. In the multi-agency model, six different companies with six different calendars, six different account managers and six different definitions of "done".

The failure this produces is rarely dramatic. It looks like a press release pointing at a page that describes a previous product version. Like a LinkedIn page whose "About" text predates the clearance. Like three domains, two of which redirect somewhere stale. Individually, each is a small embarrassment. Together they are the reason a buyer who checks you finds four slightly different companies.

We measured a version of this in our previous report: across a 60-company audit of newly cleared manufacturers, identity fragmented between the regulatory record and the commercial surface far more often than anyone would predict. The cause was almost never incompetence. It was that no single person or contract owned the whole surface.

That is the coordination tax. It does not appear on an invoice. It appears as the gap between what you paid for and what a buyer actually finds.

Why it is structurally worse than it looks

Three properties make coordination cost behave badly, and none of them are obvious when you are drawing the org chart.

It scales with connections, not with pieces. Six functions do not create six coordination relationships; they create fifteen possible pairs that have to stay consistent. Adding a seventh function does not add one unit of coordination — it adds six. This is why the multi-agency route so often feels harder than the in-house one despite costing less: you did not add capacity, you added edges.

Nobody's job description contains it. The writer is measured on articles, the developer on uptime, the PR contractor on releases sent. Coherence between them is not any of their deliverables, so it is nobody's failure when it slips. It becomes visible only downstream, as the thing a buyer noticed and you did not.

It fails silently and on a delay. A missed coordination step does not throw an error. The stale page sits there being wrong for eight months, and the cost lands when a distributor, a partner or an AI answer reaches it. By the time you see the effect, the cause is a decision nobody remembers making.

The practical implication for the arithmetic in this report is that the 0.25 FTE marketing-manager line in section 4 is not overhead you can trim to make the in-house model competitive. It is the line that makes the other 1.25 FTE produce something coherent. Cut it and you have not saved $57,593 — you have bought $133,240 of uncoordinated output.

That is the real argument for a single accountable contract, and it has nothing to do with the price comparison. A managed subscription does not remove the coordination work. It moves the obligation to a party who has agreed to carry it, and gives you one number to call when the release points at the wrong page.

---

9. How to run this arithmetic for yourself

Takeaway: do not take our number. Take the method, and produce a better one for your own company.

Three tests to apply to the result.

Is any line below 0.15 FTE? If so, you are not going to hire it. You are going to assign it to someone as an extra duty, or leave it undone. Model it that way, or model buying it.

Who holds the coordination line? If the answer is "the founder", price the founder's time at their real opportunity cost, and then ask whether that is the highest-value use of it.

What happens in month seven? Most in-house growth plans survive two quarters. Whatever your model says, run it against the version of your company that is busy, behind, and dealing with something urgent — because that is the version that has to execute it.

If you want a second opinion on the result, we will look at your numbers and tell you honestly whether a managed subscription is the right answer for you. Sometimes it is not — a company with an existing marketing team and one market to serve is usually better off extending what it has.

---

Sources

1. US Bureau of Labor Statistics — Occupational Employment and Wage Statistics, May 2024 median annual wages, as published in the Occupational Outlook Handbook for marketing managers (11-2021), web developers (15-1254), market research analysts and marketing specialists (13-1161), editors (27-3041), writers and authors (27-3043), public relations specialists (27-3031) and graphic designers (27-1024). Accessed July 2026. https://www.bls.gov/ooh/
2. US Bureau of Labor Statistics — Employer Costs for Employee Compensation — March 2026, USDL-26-0827, released 12 June 2026. Private industry total compensation $46.60 per hour worked; wages and salaries $32.60 (69.9%); benefits $14.01 (30.1%). Accessed July 2026. https://www.bls.gov/news.release/ecec.htm
3. PR Newswire national press-release distribution rates as published in 2026 rate guides: approximately $805 for the first 400 words, approximately $275 per additional 100 words, annual membership $195–$249, multimedia approximately $325 per item, logo approximately $495. Accessed July 2026.
4. VayoMed — published offer, pricing and comparison claim: $24,000 per year, 12-month term, six core services, 110–180 source-grounded articles per year, one press release per month, and the stated comparison that the traditional route runs $200,000–$300,000+ per year. Accessed July 2026. https://vayomed.com/
5. VayoMed — After 510(k), Can Buyers Actually Find You?, analysis of the 2026 post-clearance cohort from the FDA 510(k) record through 27 June 2026. Accessed July 2026. https://vayomed.com/blog/after-510k-medtech-commercial-readiness-audit