What data engineering actually costs, and how I price it

Here is the whole rate card, in USD:

Two audits a month; one full-time engagement at a time.

That is the part most consulting sites make you ask for. The rest of this post is the reasoning behind those numbers — why builds are a range rather than one price, and why none of it is a day rate.

Why there's no day rate

You are not buying hours. You are buying an outcome — a platform that works, a bill that drops, a pipeline you can trust — without the overhead of hiring for it.

A day rate anchors the conversation on the wrong number and creates a genuinely misaligned incentive: the faster and more experienced I am, the less an hourly model earns. That rewards padding and punishes efficiency. Neither of us wants that.

So builds are quoted fixed or milestone-based. You get a number you can budget and approve up front, and I am rewarded for delivering rather than for dragging it out. It requires doing the scoping properly first, which is exactly what the audit is for.

Why the build number is a range

$15–40k is a wide band, and the honesty is in the width. What moves it:

• How many sources, and how clean. Two tidy Postgres databases is a different project from a dozen third-party APIs that each change their schema on their own schedule.
• Batch or real-time. Batch is natural and cheap. Streaming ingestion from many writers is a different architecture with a different price.
• How much modelling and testing. "Get the data somewhere" is a fraction of "get the data somewhere and have people trust the numbers".
• Governance. Roles, row-level security and audit trails are real work, and most early-stage teams do not need them yet.

The audit is what turns that range into your number. A week of senior work on your actual stack produces a ranked fix-list, a cost model comparing your spend to a lean rebuild, and a real estimate. You keep it whether or not we work together, and the $3,000 credits in full toward a follow-on build. I wrote separately about why build costs range so much.

How to think about what it's worth

Instead of "what's the rate", ask two questions.

What would the alternative cost? For a build, the real comparison is the fully-loaded cost of the senior data engineer you would otherwise hire — around $150–200k a year, plus months of recruiting before they write a line of code (see consultant vs. full-time hire). For ongoing work, it is a fraction of that same salary. The other alternative is the cloud-warehouse or SaaS-stack bill a leaner build would cut, which you can estimate for yourself.

What does getting it right — or wrong — cost the business? A platform that is late, brittle or quietly wrong is not just an infrastructure line item; it is decisions made on bad data. The value of doing it well is usually a multiple of the build price.

Priced against those, this stops looking like an hourly expense and starts looking like what it is: buying an outcome at a fraction of the cost of the team you would assemble to get it.

The sequence

1. A free 30-minute scoping call. No pitch. I only recommend the audit after seeing your cost export.
2. The Data Platform Audit — $3,000, one week, a written plan and a real number. If day one shows no credible opportunity, you pay nothing.
3. A scoped build, quoted from what the audit found, with the audit fee credited in full.

You can stop after any step and keep what you have. The full mechanics — how I work, what I need from you, timezones, contracts — are written down on how I work.

If you want your number rather than a range, the scoping call below is free.