Engineering rate cards: how to read them
Engineering rate cards: how to read them
An engineering rate card is a pricing menu published by an agency, consultancy, or talent platform. It lists what each role costs by seniority, region, and engagement model (hourly, daily, or weekly). To read one, look at three axes: role + seniority on the left, region or geography across the top, and the billing unit underneath. The headline rate is almost never the true rate; markup, minimums, and overhead reshape it.
If you have ever asked a vendor "what does a senior backend engineer cost?" and gotten back a 3-page PDF with $95 to $385 per hour spread across six geographies, you have read a rate card. They look intimidating. They are not. Once you understand the structure, you can compare a Big Four consultancy quote against a boutique agency quote against a marketplace like Toptal or Cadence in about 90 seconds.
This post breaks down the anatomy of a rate card, the math behind blended rates, the difference between hourly, daily, and weekly billing, and how to spot when a number is inflated. We will end with a side-by-side of three real pricing models so you can see what the same engineer actually costs from different vendors.
What a rate card actually is
A rate card is a vendor's published price list for billable labor. It exists for the same reason a restaurant publishes a menu: it speeds up sales conversations, sets anchor expectations, and protects the vendor from undercharging.
Most rate cards live in one of three places:
- A PDF emailed during the sales cycle (consultancies)
- A pricing page on the vendor's site (talent platforms, agencies)
- An MSA appendix locked in once you sign (Big Four, large SIs)
Public rate cards tend to be lower than private ones. The version on a vendor's website is the marketing rate. The version in your MSA is the actual rate, often 15 to 35 percent higher once you add overhead, expenses, and minimum hours per engagement.
A rate card is not a contract. It is an anchor. The number you actually pay depends on scope, duration, payment terms, and how hard you negotiate.
The three axes every rate card uses
Every rate card you will ever read has the same skeleton. Get familiar with these three axes and the rest is detail.
Axis 1: Role and seniority
Roles are usually one of: software engineer, mobile engineer, data engineer, ML engineer, DevOps engineer, designer, product manager, QA. Seniority bands are typically: junior (0 to 2 years), mid (2 to 5 years), senior (5 to 10 years), principal or staff (10+ years), and sometimes architect or fractional CTO at the top.
A clean rate card lists every role across every seniority band. A messy one lumps everything into "developer" with no seniority breakdown, which is a sign the vendor wants pricing flexibility (read: they will assign whoever is on the bench).
Axis 2: Region or geography
The same role costs very different amounts depending on where the engineer sits. A senior software engineer in the US runs $150 to $250 per hour at most agencies. The same engineer in Poland or Mexico runs $60 to $110. In India or Vietnam, $30 to $65. Eastern European mid rates and US junior rates often overlap, which is why a lot of US agencies quietly run a Krakow or Lisbon office.
If a rate card does not break out region, the vendor is either fully local (small US boutique) or hiding a nearshore arbitrage they would rather not advertise.
Axis 3: Engagement model and billing unit
This is where the comparison gets tricky. Vendors bill in different units:
- Hourly is most common for staff augmentation, marketplaces (Toptal, Upwork Top Rated), and short engagements under 100 hours.
- Daily is the Big Four and large SI default. A day usually means 8 billable hours, but check the contract.
- Weekly is the talent-marketplace model (Cadence, some Andela engagements). One flat rate, full week of focused capacity.
- Monthly retainer is the offshore dedicated-team model (Toptal teams, Andela, EPAM long-engagement).
- Project-fixed is most common at boutique agencies for defined scope.
The same human can show up at $150 per hour, $1,100 per day, or $5,500 per week depending on which sales motion the vendor runs. Sometimes those numbers reconcile cleanly. Often they do not.
How to read a sample rate card
Here is a realistic mid-market agency rate card. We will use it as a worked example.
| Role | Junior (US) | Mid (US) | Senior (US) | Mid (LATAM) | Senior (LATAM) |
|---|---|---|---|---|---|
| Software Engineer | $95/hr | $145/hr | $215/hr | $75/hr | $115/hr |
| Mobile Engineer | $110/hr | $160/hr | $225/hr | $80/hr | $120/hr |
| Data Engineer | $115/hr | $175/hr | $245/hr | $85/hr | $130/hr |
| ML Engineer | $135/hr | $195/hr | $275/hr | $95/hr | $145/hr |
| DevOps Engineer | $110/hr | $165/hr | $235/hr | $80/hr | $125/hr |
| Designer | $95/hr | $135/hr | $185/hr | $70/hr | $105/hr |
| Product Manager | $125/hr | $175/hr | $225/hr | $90/hr | $135/hr |
Reading this card top-to-bottom:
A US senior software engineer is $215 per hour. Assuming a normal 35 billable-hour week (the other 5 hours go to standups, code review, and Slack), that engineer costs you $7,525 per week, or roughly $30,000 per month before any expenses or PM overhead.
The same role in LATAM is $115 per hour, or about $4,025 per week. That is a 47 percent discount for what is often the same skill in a more convenient timezone than offshore India.
Now look at the structural pattern. Mid-LATAM ($75/hr) sits roughly equal to Junior-US ($95/hr). That overlap is the single most important number on the card. It tells you the vendor's "cheap" line, the cheapest billable human they will put on your project. Below that, you are getting an intern or a bait-and-switch.
Blended rates: the number you actually pay
Real engagements rarely use a single rate card line. You get a team: one senior architect, two mid engineers, one junior, and a fractional PM. The vendor blends those rates into one number for budgeting.
A typical 4-person blended rate calculation:
- 1 senior at $215/hr × 10 hours = $2,150
- 2 mids at $145/hr × 35 hours each = $10,150
- 1 junior at $95/hr × 35 hours = $3,325
- 1 PM at $175/hr × 10 hours = $1,750
Total billable cost per week: $17,375 across 125 hours, for a blended rate of $139/hr.
The blended rate is what you put in the budget spreadsheet. The line-item rates are what you use during scope negotiations ("Can we drop the senior architect to 5 hours and reallocate to a mid?"). If a vendor will only quote you a blended number and refuses to show line items, the markup is probably hiding in there.
Engineering team economics matter at every team size. We track this more deeply in our breakdown of engineering team cost by country, which shows how the same 4-person team prices out across the US, Western Europe, Eastern Europe, and South Asia.
Day rate vs weekly rate vs hourly: which is honest?
Each billing unit hides different costs. Knowing which is which saves you from comparing apples to oranges.
Hourly billing rewards speed against the vendor and is dishonest by default. The engineer who takes 8 hours to ship a feature earns more than the one who takes 4. Marketplaces (Upwork, some Toptal engagements) reinforce this by tracking hours via screenshot software, which produces presence theater rather than output. Hourly is fine for short, scoped work (a 12-hour migration). It is a bad model for ongoing capacity.
Day rate billing is what consultancies prefer because it smooths the bill. A day is sold as 8 hours but is usually 5 to 6 hours of focused work plus 2 to 3 hours of meetings, status updates, and context-switching. Big Four day rates of $1,800 to $3,500 per senior consultant work out to $225 to $440 per "real" hour once you strip the ceremony.
Weekly billing flattens the unit entirely. You pay a single number for one engineer for one week of focused work. No screenshot tracking, no day-rate math, no surprise overruns. Cadence uses this model: junior at $500 per week, mid at $1,000, senior at $1,500, lead at $2,000. The trade-off is that you lose granular accounting (you cannot bill 3 hours to project X), but you gain forecasting clarity and remove the perverse incentive to drag work out.
Monthly retainer is the most opaque. You pay $X for a "dedicated engineer" or "dedicated team" with vague output expectations. This model works when you have stable, predictable workload and a clear point of contact. It fails when the vendor swaps people silently or counts an engineer on three accounts as "100% dedicated" to each.
The honest hierarchy, in our experience: weekly > project-fixed > daily > monthly > hourly.
When consultancies inflate
Big Four and tier-1 consultancies (Accenture, Deloitte Digital, McKinsey Digital, BCG X) publish rate cards with $400 to $700 per hour for senior engineers and $800 to $1,400 per hour for partners. These numbers are real. They are also wildly inflated for delivery work.
The math: a Big Four senior consultant earns $200,000 to $280,000 in base + bonus. At $500 per hour billed across 1,600 billable hours per year, that engineer generates $800,000 in billings. Gross margin runs 60 to 70 percent. The "rate card" funds the partner pyramid, sales motion, and the bench of people not currently on a project.
You are paying for the brand and the insurance. If something goes wrong, you can sue a Big Four. If your boutique vendor disappears, you cannot. That insurance is real and worth real money for risk-averse enterprise buyers. It is not worth it for a 6-week sprint to ship a product feature.
Spot the inflation: rates above $400 per hour for delivery (non-strategy) work, a $200k or 6-month minimum, a long "discovery phase" billed before any shipping begins, multi-tier teams where the partner attends one call per week and bills 4 hours, and "knowledge transfer" billed at full rate. None of this is fraud; it is the consulting model working as designed. Built for governance and political cover, not for shipping fast.
The same applies (in milder form) at top-tier strategy consultancies. We have written about why engineers leave high-rate firms, and a lot of it is that the high billing rate funds overhead the engineer never sees in compensation.
Comparing three rate models head-to-head
Let us put a senior backend engineer in front of three different pricing models. Same human (5 years experience, ships React + Node, knows Postgres), three different bills.
| Vendor | Model | Headline rate | True weekly cost | Minimum engagement | Replacement timeline |
|---|---|---|---|---|---|
| Big Four (US) | Hourly | $425/hr | $14,875 (35 hrs) | 12 weeks | 4-6 weeks |
| Toptal | Hourly | $90-$200/hr | $4,725 (avg $135, 35 hrs) | None (often 20 hrs/wk min) | 1-2 weeks |
| Boutique agency (LATAM) | Hourly | $115/hr | $4,025 (35 hrs) | 3 months | 2-3 weeks |
| Andela | Monthly retainer | $9,000/mo | $2,250 | 6 months typical | 2-4 weeks |
| Cadence | Weekly flat | $1,500/wk | $1,500 | None (week-by-week) | 48 hours |
A few honest observations:
Big Four wins on governance, audit trail, and risk transfer. If you need a SOC 2 attestation lineage and a name a board recognizes, you pay $14k a week. That is the product.
Toptal wins on engineer quality and process maturity. Their top 3 percent claim is real for the senior tier, and the platform UX is polished. The hourly model and the (often hidden) 20-hour minimum nudge bills higher than the headline rate suggests.
Andela wins on stable, long-engagement Africa / LATAM teams. Their model assumes you want a "dedicated" person for 6 months or more. Great for enterprises building out an offshore arm, expensive for a 4-week sprint.
Cadence wins on speed and flexibility. Weekly billing, 48-hour trial, no minimum engagement, replace any week. Every engineer on Cadence is AI-native by default, vetted on Cursor, Claude Code, and Copilot fluency before they unlock bookings. The trade-off: no large-account governance layer, no PM-as-a-service, no enterprise procurement integration. If you need an MSA with a Fortune 500 procurement department, Cadence is not the right shape. If you are a founder or PM who wants to book a vetted engineer Tuesday and have them shipping by Thursday, the math is hard to beat. Our internal data shows a 27-hour median time to first commit across the 12,800-engineer pool.
For the full apples-to-apples math, our revenue per engineer benchmarks page walks through how each model affects the per-engineer ROI calculation.
What to do with a rate card you have been sent
Three steps, in order. Skip none.
1. Translate everything to one billing unit. Pick weekly. Convert hourly to weekly at 35 billable hours (not 40; meetings exist). Convert daily to weekly at 4.5 days (not 5; nobody bills Friday afternoon at full velocity). Now the comparisons are direct.
2. Add 25 percent for the things not on the card. Project management, account management, "ramp-up" hours, weekly status meetings, "knowledge transfer" between team rotations, the inevitable 1-week onboarding when their preferred engineer is on vacation. Some vendors fold this in. Most do not. Add it.
3. Stress-test the replacement story. What happens if the engineer they assigned does not work out? At a consultancy, this is a 4 to 6 week reassignment with no guarantee the next person is better. On Cadence, daily ratings drive auto-replacement and you can request a different engineer next week with no notice period. The replacement story is hidden value or hidden risk depending on the vendor.
If you want to skip the rate card translation entirely, run the numbers on a weekly booking instead. One rate, one unit, no math required.
Cadence quotes a single weekly rate per tier (junior $500, mid $1,000, senior $1,500, lead $2,000), with a 48-hour free trial. No procurement loop, no minimum engagement, replace any week. Book your first engineer and skip the rate card decoding entirely.
FAQ
What is the difference between a rate card and a quote?
A rate card is the published price per unit (per hour, per day, per week). A quote is the total for a specific scope, calculated by multiplying the rate card by the estimated time. Rate cards are anchors; quotes are commitments. A vendor that will give you a rate card but not a quote is hedging on scope; a vendor that gives a quote without a rate card is hedging on accountability.
Why do engineering rate cards vary so much by region?
Because cost of living, currency, and local labor competition vary by 5x across the geographies most vendors operate in. A senior engineer in San Francisco competes for offers at Google ($350k+ TC); the same engineer in Krakow competes at a local market median of $60k to $90k. Rate cards reflect that gap, plus a margin layer for the vendor.
Is an hourly rate card always more expensive than weekly?
Usually yes, but not always. Hourly billing rewards the vendor for taking longer, which inflates total spend on open-ended work. Weekly billing caps the bill but assumes you have enough scope to fill the week. For a 6-hour bug fix, hourly is cheaper. For 6 weeks of feature work, weekly almost always wins.
How do I tell if a rate card is inflated?
Three signals. First, the senior rate is more than 3x the junior rate (most healthy markets show 2x to 2.5x). Second, there is no published rate card at all and you have to "request a quote" (signals price discrimination). Third, the day rate divided by 8 hours produces an hourly rate higher than the published hourly card (signals an undocumented day-rate premium).
Can I negotiate a rate card?
Yes, on three levers: volume (longer engagement, more headcount), payment terms (net-15 vs net-60 is often worth 5 to 8 percent), and exclusivity (locking the vendor out of a competitor in your space). Smaller boutiques negotiate freely. Big Four firms negotiate at the partner level. Marketplaces with published flat rates (Cadence, parts of Toptal) generally do not negotiate per-engineer rates but will discount multi-engineer bookings.