What Mid-Market Companies Get Wrong About HR Outsourcing
I have spent almost 25 years in payroll and HCM, and the last 15 building a boutique firm that has now implemented for 571 clients. Companies between 50 and 2,000 employees call us with the same opening question every time: should we outsource HR?
That question cannot be answered. The useful version is narrower. Which functions, at what cost per employee per month, and who owns the outcome when something breaks.
The three tiers, and what they actually cost
Pricing in this market is deliberately murky, so here are the real numbers.
Payroll processing only runs $5 to $10 per employee per month. You are buying tax filing and a paycheck. Nothing else.
Managed payroll services land between $30 and $50 PEPM. Someone else runs the payroll, chases the timecard exceptions, and handles the agency notices when they show up.
Full HR outsourcing on a real HCM platform sits at $50 to $75 PEPM. Payroll, benefits administration, onboarding, compliance support, and a named human who knows your business.
Plenty of companies are paying single-digit PEPM for something a sales rep called managed payroll. That is not managed payroll. That is processing with a support inbox, and the gap shows up the first time a tax notice arrives.
The question that separates real partners from vendors
Ask any provider this: who runs my implementation, and how many of these have they personally done?
Watch what happens. Most national providers will tell you about their implementation team, their methodology, their onboarding portal. They will not give you a name, because the name changes three times before you go live.
Implementation is where outsourcing succeeds or fails. The software is roughly the same across the serious platforms. What differs is whether the person configuring your earning codes has done it 200 times or is learning on your company.
We have inherited enough broken configurations to know how it goes. Complex pay rules get skipped during the original build, usually the union rates or the shift differentials nobody wanted to map. Somebody in accounting rebuilds them by hand in a spreadsheet. That spreadsheet quietly becomes the real system of record, and the platform you are paying for turns into an expensive place to store names.
What you are actually buying
Outsourcing HR is not a cost reduction play at this size. Companies that pitch it that way are selling you a spreadsheet exercise that stops being true in month seven.
You are buying capacity and risk transfer. Your HR generalist stops losing most of the week to payroll reconciliation and starts doing retention work. Your CFO stops absorbing penalty notices from the Department of Labor because somebody misclassified a group of field employees.
Price that against the fully loaded cost of the two HR hires you were about to make. That comparison is honest. The one where outsourcing pays for itself in six months usually is not.
How to run the evaluation
- Inventory what you do today. Every HR and payroll task, who owns it, how many hours. Most companies have never written this down, and the exercise alone surfaces two or three things nobody knew were happening.
- Separate what must stay in-house. Employee relations, culture, and hiring decisions do not outsource well. Transaction processing and compliance filing do.
- Get named references at your headcount and in your industry. Not logos on a slide. Phone numbers of people who went live in the last 18 months.
- Ask about year two. Implementation gets attention. Find out what your service model looks like once the go-live team moves on, and get it in writing.
Companies evaluating HR outsourcing solutions tend to spend 80% of their time comparing feature lists and 20% comparing the people who will run the thing. Invert that ratio and you will pick correctly.
One last thing
The smoothest go-lives I have been part of all share a trait. The company mapped its own processes before it talked to a single provider, so it walked into every demo knowing exactly what it needed. Those evaluations run weeks instead of quarters, and those companies go live on schedule.
The worst outcomes all started the same way. Somebody got a demo, liked the dashboard, and signed.