The Hidden Costs of Outsourcing Nobody Puts in the Proposal

The proposal says thirty five dollars an hour. You do the arithmetic in your head, compare it against what the same role costs you locally, and the decision feels obvious. It usually is obvious. That is not the problem.

The problem is that the number on the proposal is roughly sixty percent of what the arrangement will actually cost you in year one. Not because the vendor is lying. Because the vendor is quoting the only thing they can quote, which is their own labour, and the rest of the cost lands on your side of the line where nobody itemises it.

I am not here to talk you out of outsourcing. I am here to make sure you budget for the real number, because the companies that get burned are almost never the ones who understood the full cost and went ahead anyway. They are the ones who budgeted for sixty percent and then spent the next six months quietly wondering why nothing felt like a win.

The biggest hidden cost is your own calendar

Start here, because it is the one everyone underestimates by the widest margin and it is the one you cannot invoice to anybody.

Somebody on your team has to manage this. Not "check in occasionally." Manage. Write briefs, answer questions, review output, chase clarifications, sit in a call at an awkward hour, catch the thing that went sideways, re-explain the thing you already explained once.

Rough shape of what this costs, in hours per week, from what I have watched happen repeatedly:

• Weeks 1 to 4: eight to fifteen hours a week. This is onboarding. It is intense and it is unavoidable.
• Weeks 5 to 12: five to eight hours a week. Volume drops, the questions get better, you are still correcting a lot.
• Steady state, month four onward: two to five hours a week per outsourced function, assuming it is going well.

Now price that. If the person doing the managing is a founder or a senior operator, their loaded cost is not their salary divided by two thousand. It is what the business loses when they are not doing the thing only they can do. Most founders, if they are honest, put that number somewhere between one hundred and three hundred dollars an hour and then immediately feel embarrassed about it.

Five hours a week at even a conservative one hundred and fifty is thirty thousand dollars a year of management overhead on a contract you budgeted at seventy thousand. Nobody puts that in the proposal. Nobody could.

Here is the annoying part. This cost does not go away. It gets smaller and it changes shape, but delegation never becomes free. Outsourcing does not remove the management job. It converts a doing job into a managing job, and if you preferred the doing job you are going to be miserable regardless of how good the vendor is.

Documentation, which you have to build before anyone can start

You cannot hand over a process that exists only as a habit in one person's head. So before the vendor's first billable hour, somebody has to write it down.

Realistically that is fifteen to forty hours of internal work per meaningful function. Screen recordings, step lists, decision rules, edge cases, the four exceptions that everyone in the office knows about and nobody has ever articulated. This is genuinely valuable work and you should do it anyway. But it is work you are doing because you decided to outsource, at the exact moment you were hoping to have more time, and it happens before you see a single unit of output.

The related trap: teams skip this, hand over a half-explained process, and then pay for it later in rework at a much worse exchange rate. Forty hours of documentation up front is cheaper than four months of correcting misunderstandings. It is not close. We covered the mechanics of this in the daily operations playbook, and it remains the single highest-leverage thing you can do before a handover.

The productivity dip that everyone forgets to plan around

For the first six to ten weeks, total output across your team goes down. Not up.

Your existing people are answering questions instead of working. The new team is producing at maybe forty to sixty percent of eventual throughput, and a chunk of what they produce needs reworking. You have added cost and subtracted capacity simultaneously.

This is normal. It is not a sign you picked badly. But if you promised your board a cost saving in Q1 and you signed the contract in January, you have promised something the calendar will not give you. Plan for the curve to cross over somewhere in month three or four, and treat anything earlier as a pleasant surprise.

The dip is worse when you outsource something urgent. Outsourcing under time pressure is like moving house during a fire. Technically possible. Rarely a good idea.

Knowledge transfer and rework

Every misunderstanding costs three times: the wrong work, the conversation about the wrong work, and the right work.

In the early months, expect somewhere between fifteen and thirty percent of output to need meaningful revision. That percentage is a direct function of how good your documentation was and how specific your feedback is in weeks one and two. Vague feedback is expensive feedback. "Looks good" in week one is a bill that arrives in month five.

There is also the reverse flow that nobody mentions. Your outsourced team accumulates context. When they leave, that context leaves. Which brings us to the cost people plan for least.

Attrition and re-ramping

People leave. Vendors rotate staff. The engineer you interviewed in March is on a different account by September, and the replacement is "equally senior," which is a phrase that means nothing.

Every rotation costs you the ramp-up again. Not the full onboarding, because the documentation exists now, but four to eight weeks of reduced output and renewed management load. If you are running a vendor with twenty five percent annual churn on your account, you are paying a re-ramping tax roughly every eighteen months per seat, and the tax does not appear anywhere in your rate card.

This is the practical reason to care about commercial models that align the vendor with retention rather than churn. A staffing vendor billing a monthly markup is economically indifferent to who fills the seat, as long as the seat is filled. A recruiter charging a one-time placement fee with a replacement guarantee attached has skin in the game for exactly as long as the guarantee runs. The placement fee model Zemenay Tech runs out of Ethiopia, for instance, prices at roughly eight to twelve percent of first year salary with a six month replacement window and no recurring cut, which means their revenue does not survive a bad match the way a monthly billing model does. Whether that model suits you depends on whether you want an employee or a rented seat. But the incentive difference is real and it shows up in your re-ramping costs.

Tooling, seats and access

Small, boring, and it adds up faster than people expect.

Every person you add needs licences. Project management seat, chat seat, design tool, code repository seat, CRM seat, password manager, VPN, whatever your stack is. Fifty to two hundred dollars per person per month is a normal range once you count everything, and for a team of six that is a line item somewhere between three and fourteen thousand dollars a year that appeared on your credit card without anyone deciding it should.

Then there is the security work. Access reviews, offboarding checklists, device policy, the awkward conversation about whether a contractor should have production access. If you are in a regulated industry, add compliance documentation and possibly an audit conversation. None of this is optional and none of it is in the vendor's number.

Quality assurance and review overhead

If the outsourced work touches a customer or a ledger, someone reviews it. That reviewer is a cost.

For content, code and finance work, a sensible planning assumption is that review consumes ten to twenty percent of the hours the production work consumed. It drops over time as trust builds and as you move from checking everything to sampling. It never goes to zero, and the moment it does you will discover why it should not have.

The version of this that stings is when you hire cheap and the review burden turns out to be forty percent. At that point the cheap rate is not cheap. It is a subsidy you are paying with senior staff time, which is your most expensive currency.

Context switching and time zone friction

If your team and their team share two hours a day, every question that needs a human answer costs you a day of latency.

That is not a disaster for well-scoped, independent work. It is a serious tax on anything tightly coupled, exploratory or ambiguous, where the work advances in a series of small clarifications. Twelve hours of overlap gap means a five-message thread takes most of a week.

The mitigation is either structural, meaning you pick a region with more overlap, or procedural, meaning you write better briefs so fewer questions are needed. Both are real options and we go through the tradeoffs in the offshore, nearshore and onshore decision. What you should not do is assume goodwill will absorb it. It will not, and asking people to permanently work nights produces exactly the attrition you were trying to avoid.

The compounding cost of a monthly markup

Here is the one that is invisible in year one and enormous by year three.

Most outsourcing is sold as a rate with a margin baked in. That margin is fine. It is how the vendor stays in business. But it is charged on every hour, forever, and it does not decay as the relationship gets easier.

Take a role where the underlying market salary is fifty thousand dollars. A vendor bills you at a rate that works out to eighty thousand a year. The thirty thousand difference is margin, admin and risk premium, and it is a perfectly reasonable price for flexibility in month one.

Over thirty six months you have paid ninety thousand dollars in markup. On a role you now know you need permanently, for a person you would happily hire directly, whose replacement would cost a one-time fee somewhere in the single-digit thousands.

Nobody selling you the monthly model is going to run that comparison for you. Run it yourself. Sometimes flexibility genuinely is worth ninety thousand dollars, particularly if the need is seasonal or the project has a defined end. Often it is not, and the honest test is simple: do you expect this seat to exist in two years? If yes, the markup model is the expensive option and you should at least know that before you sign.

The worked example

Illustrative numbers, not a benchmark. The point is the shape of the gap, not the digits.

A mid-sized company outsources a three-person operations pod. The quoted rate is twenty five dollars an hour per person, full time.

Line item	Year one cost	Where it appears
Quoted vendor cost, 3 people	$156,000	The proposal
Internal management time (6 hrs/week avg at $150)	$46,800	Nowhere
Documentation and SOP build (80 hrs at $100)	$8,000	Nowhere
Productivity dip, weeks 1 to 10	$18,000	Nowhere
Rework and correction	$12,000	Nowhere
Tooling and licence seats	$5,400	A credit card statement
QA and review overhead	$14,000	Nowhere
One mid-year replacement and re-ramp	$9,000	Nowhere
Loaded year one total	$269,200	Your actual P&L

The quoted cost is fifty eight percent of the real cost. That is the number to plan around.

And before anyone concludes this kills the case: the in-house equivalent of that pod, at Western salaries plus employer taxes plus benefits plus recruiting fees plus desk cost, is comfortably north of three hundred thousand and carries most of the same management overhead. Outsourcing still wins here. It just wins by a smaller margin than the proposal implies, and the smaller margin is the one you should be defending in a budget meeting.

Which of these can you actually avoid

Not all of these costs are equal. Some are structural and some are self-inflicted.

Price of admission, do not fight it:

• Your management time. It shrinks, it never disappears.
• The first six to ten week dip.
• Tooling seats.
• Some level of review overhead.

Avoidable, and usually cheaper to fix than to absorb:

• Documentation debt. Pay it before handover, not after.
• Most rework, which is a briefing quality problem wearing a competence costume.
• Time zone latency, which is a region selection problem you make once.
• Churn costs, which are largely a function of the commercial model and how the vendor's incentives are set.
• Markup compounding, which is a contract structure decision you can revisit at renewal.

The pattern is that the avoidable costs are all decided in the first eight weeks, and the unavoidable ones are the ones people spend their energy complaining about. Reverse that. Spend the energy on scope, documentation and commercial structure, which is where a scope of work that actually holds up earns its keep, and accept the rest as the cost of not doing the work yourself.

Outsourcing is still one of the best trades available to a small company. Just stop calling thirty five dollars an hour the price. It is the deposit.