Why the Cheapest Offshore Rate Usually Costs the Most

An $18/hr engineer and a $55/hr engineer are not the same product at different prices. They are different products, and the cheaper one transfers work to you rather than removing it. This page shows where that transfer happens and how to price it, because "we found someone cheaper" is a reasonable thing to say right up until you have run the numbers.

We are not a cheap supplier. Our bands run $22–85/hr. The argument below is why that is the rational choice rather than the expensive one — and it includes the cases where the cheap rate genuinely is correct.

Rate is a price. Cost is rate divided by delivered work

The only number that matters is the cost of a completed, reviewed, deployed unit of work. Four things sit between the rate and that number.

1. Rework

Work that has to be done twice costs twice, plus the review that caught it, plus the delay. Rework on infrastructure and security work is particularly expensive because it is often discovered late — during an audit, an incident, or a cloud bill.

2. Review load transferred to you

A less experienced engineer needs more supervision. That time is yours, it is senior, and it never appears on an invoice. An engineer who needs 8 hours of senior review a week is consuming roughly $720 of your capacity at a $90/hr internal cost — on top of their own rate.

3. Decisions you now have to make

Experience is mostly the ability to decide without escalating. A cheaper engineer escalates more: which library, which pattern, is this edge case real. Each escalation costs a context switch on your side, and context switches are the most expensive small thing in engineering.

4. What gets built that should not have been

The most expensive outcome is not slow work. It is work that was completed competently and should never have been started — the service that did not need to exist, the abstraction nobody needed, the migration that solved a problem you did not have. Seniority is largely the judgement to say "we do not need this", and that judgement is exactly what the cheapest rate does not include.

The arithmetic

One engineer, one month, 160 hours, with your internal senior review costed at $90/hr.

	Budget offshore	Techtweek senior
Rate	$18/hr	$55/hr
Invoice	$2,880	$8,800
Effective delivery rate	~55%	~90%
Delivered engineering hours	88	144
Senior review required	8 hrs/week = 32 hrs	1.5 hrs/week = 6 hrs
Your review cost at $90/hr	$2,880	$540
Rework allowance	~20% of delivered work	~5%
Total cost	$5,760	$9,340
Cost per delivered hour	$65	$65

Two things are worth noticing.

First, at the point of comparison they are roughly level. A cheap engineer is not automatically a bad deal — this is the honest version, and any vendor showing you a table where their option wins by 3x has chosen the inputs to make that happen.

Second, the budget column consumed 32 hours of your senior engineer's month. That is most of a working week of your most constrained resource, spent reviewing rather than building. If that senior has nothing better to do, the cheap engineer is genuinely fine. If they are your architect, you have just bought $2,880 of review by spending your scarcest capacity.

Where the cheap rate is genuinely right

• Well-specified, repetitive work — data entry into a system, bulk content migration, test writing against an existing specification, straightforward CRUD against a settled schema.
• When you have surplus senior review capacity and would rather spend it than money.
• Short, bounded tasks where a mistake is cheap and visible immediately.
• When you are genuinely experimenting and the code is expected to be thrown away.

Anything where a mistake surfaces months later — infrastructure, security, data modelling, auth, billing — is where the cheap rate stops being cheap. Those are also the four places we see the most expensive remediation work, which is a sampling bias worth declaring, but the pattern is consistent.

What you are paying for between $18 and $55

• Judgement about what not to build — the largest single saving, and the hardest to see on an invoice
• Decisions made without escalation, which is your context, protected
• Production experience of the failure modes — the specific ways a Terraform state gets corrupted, an EKS upgrade goes wrong, a migration locks a table
• Work that arrives reviewable, with tests and a rollback path, rather than work that arrives needing to be understood first
• Evidence the work will survive an audit — for us specifically, engineers built inside a practice that runs PCI DSS, ISO 27001 and SOC 2 engagements

How to tell which one you are being offered

Ask for the effective delivery rate, not the hourly rate. Then ask these:

1. How many hours of our senior time will this engineer need per week?
2. What does this engineer decide without asking us?
3. Show me work they shipped that is still running two years later.
4. Which of the last three engagements needed a replacement, and why?
5. What will they tell us not to build?

A supplier who answers question 5 with a straight face is selling judgement. A supplier who does not understand the question is selling hours.

The rate is the wrong negotiation

If the budget is fixed, negotiate the shape, not the rate: fewer hours of a better engineer, part-time instead of full-time, a narrower scope done properly. Twenty hours a week of someone who decides well beats forty hours of someone who escalates — and it costs less.

That is a conversation we will have honestly, including telling you when the answer is that you do not need us this quarter.

Related

• The hidden costs of offshore engineers nobody quotes you
• Hire dedicated developers — our bands and what is in them
• Staff augmentation vs managed services vs outsourcing