A rate card does not fail at the moment you build it. It fails somewhere in month nine. Two new practices exist, half the roles have been retitled, the senior engineers you hired against have all left, and every open deal is priced against a number nobody has looked at since March.
The card still opens. The finance team still signs off each quarter. The sellers still quote from it. And the margin on closed deals keeps drifting away from the target nobody thinks to re-test.
A services rate card is a negotiation instrument, not a spreadsheet. It only survives if it is built to be re-used by sellers in the room, not filed by finance in a drive.
Why a services rate card fails in month nine, not month one
Most teams build the first rate card carefully. Someone pulls loaded costs per role, applies a target margin, rounds the numbers, circulates the sheet, and everyone agrees it looks right. Finance stamps it. The sellers paste it into a proposal template.
Then nothing changes it for eleven months.
During those eleven months, three things happen. The business hires differently from what the card anticipated, usually more seniors and fewer juniors than plan. A new practice shows up that the card has no roles for. Sellers pick the closest role and hope, or freelance a number into the quote. And the cost inputs drift. Salary bands, benefits, and on-cost percentages all move a few points each.
None of this is a disaster individually. Together, the card no longer represents how the business delivers work.
The sellers notice first. They start asking finance for "a one-off" rate on the new practice. The practice leads notice next, when delivered margin stops matching what finance is reporting. By the time anyone says "we should rebuild the card", a year of deals has been priced against a number that stopped being accurate in March.
Why one margin number hides every bad deal
The second failure is subtler. A single margin target, say 50%, applied uniformly across every role, is the default shape of most rate cards. It is also the reason the card cannot tell you which deals are quietly unprofitable.
A 50% blended margin across a deal does not mean every role is delivering 50%. It usually means a senior architect priced at 60% is carrying three juniors priced at 40%, and the weighted average lands on target. Then the deal staffs differently to plan. A harder piece of work needs two architects and one junior. The blended margin collapses, and nobody sees it until delivery reports come in.
The fix is not to raise the margin target. The fix is to stop using one number.
Role granularity is the first decision, and you will get it wrong
Before you price anything, decide what a "role" means on this card. This is the decision that matters most, and the one teams underthink.
Three common mistakes:
- Too granular. Separate lines for Senior Developer, Senior Backend Developer, Senior Node.js Developer. You cannot staff at that resolution, your juniors do not have the title, and the card becomes unusable within six months as titles evolve.
- Too coarse. One "Developer" line covering everything from a first-year junior to a fifteen-year platform architect. Every deal has to manually decide where in the range the specific resource sits. The card tells sellers nothing.
- Matched to the org chart. The roles on the card mirror the HR system. Looks tidy, estimates badly. HR tracks people. The card needs to track the shapes of work the business sells.
A rate card should have one role per distinguishable delivery shape. Senior vs junior matters. Frontend vs backend usually does not, unless your deals staff those roles differently. Architect, lead, senior, mid, junior is often enough across a practice. If you have ten practices, you may end up with forty to fifty roles. That is fine. What is not fine is two hundred roles nobody can remember.
A reasonable test: can a presales lead quote a new deal using only the card, without asking finance a single question? If not, either the card is missing roles or the roles are named badly.
Margin bands, not a single margin
The single most useful change you can make to a rate card is to stop quoting one margin and start quoting three. A low, a normal, and a high, per role, per card.
- Low. The floor. A deal priced below this on any role requires explicit approval, named in the record.
- Normal. The default the card calculates against. The number sellers see first.
- High. The ceiling. A deal priced above this is either ambitious pricing or a rate card that has fallen behind market. Both are worth flagging.
The three numbers do several things at once. They make approval thresholds mechanical instead of political. Any deal that dips below the low on any role needs a second signature. Any deal that clears the high is a commercial conversation, not a quiet win. They give sellers a negotiating range without a call to finance every time a buyer pushes back. And in the aggregate, they expose which roles are systematically pricing at the floor. That is usually the first symptom of either a pricing problem or a cost problem.
Set the floor so the business still delivers at acceptable margin when the deal gets ugly, not at the minimum anyone is willing to accept on a good day. The floor is a safety rail, not an aspiration. If you are using margin floors as a discount gate, the related move is to trade concessions for something back rather than cut the rate, covered in never discount without a trade.
Regions, practices, and rush: when one card stops being enough
A single rate card works until it does not. The usual triggers:
- Multiple currencies or regions. A developer in London and a developer in Bucharest deliver the same work, but the cost base is different by a factor of two or three. One card with a blended cost number misprices both markets.
- Separately managed practices. A data practice with dedicated seniors and a shared development practice with fluid staffing have different cost structures. Forcing them onto the same card means one subsidises the other invisibly.
- Rush work. Compressed timelines cost more to deliver. A separate "rush" card, used only when the deal commits to a shortened delivery window, makes the uplift visible and defensible instead of hidden in a padded estimate.
- Managed services vs project work. Recurring run work and one-off project work price differently. The run rate includes availability, handover, and lower utilisation assumptions. Conflating them on one card means the project side looks too expensive and the run side looks too cheap. If you are still picking between those shapes, how to pick the right pricing model is the prior conversation.
The rule of thumb: if you are regularly manually adjusting the card during pricing, that adjustment probably wants to be its own card.
Rounding is not cosmetics
Rounding matters more than teams expect, for two reasons.
The first is signalling. A rate of £1,247 per day reads as calculated, not deliberate. A rate of £1,250 reads as a price. Buyers subconsciously assume calculated numbers are soft. Rounded numbers feel set.
The second is negotiation. If every role on the card is rounded to the nearest £25 or £50, a concession of "take £100 off the senior rate" is a defined move, not a slippery slope. Unrounded rates invite the buyer to chip away in fifties and hundreds of pounds until the card's internal logic collapses.
Pick a rounding policy: nearest £10 for juniors, nearest £25 for mid, nearest £50 for senior, nearest £100 for principal. Apply it consistently. Document it once, with the reasoning, so nobody rebuilds the card next year without it.
Review triggers, not review cadence
Annual rate card reviews are a coping mechanism for cards that nobody trusts to reflect reality in the meantime. They are cheaper than a trigger-based rebuild, and they are how most teams end up with a card that reflects last March's business.
A better model: name two or three triggers that force a partial rebuild, and accept that the card is otherwise living.
Useful triggers:
- Hiring plan shift. If the next-quarter hiring plan moves the role mix by more than 15%, the card rebuilds for the affected roles. New practice, new roles. Shifted seniority mix, shifted costs.
- Pipeline shape change. If the pipeline starts showing a class of deal the card was not built to price, like larger fixed-price deals, longer-running retainers, or rush work, the card gets a new card or a new role band for that shape.
- Margin drift. If delivered margin on closed deals drifts more than three points from the card's target for two quarters in a row, something is wrong in either cost inputs or the margin band. Rebuild the inputs, not the target. The pattern of not seeing margin until after the deal is signed is its own problem, worked through in stop guessing margin at the point of sale.
None of these are calendar events. All of them are things RevOps should already be watching. Wiring the rate card review into the same cadence as hiring and pipeline review means the card keeps up with the business that uses it, not the business that existed when the card was approved.
In practice
A rate card that lives next to the tools that price deals survives longer than one that lives in a drive. In Estii, every deal picks its rate card at creation, and each phase can switch cards when the shape of the work changes. The low, normal, and high margin bands enforce themselves in-line as sellers negotiate. A role priced below the floor highlights in the deal. A deal-level target margin recalculates the whole card. Rate rounding applies automatically so numbers stay deliberate, not calculated. Multiple cards cover region, practice, and rush without a manual adjustment per deal. Each deal snapshots the card version it was priced against, so rebuilding the card tomorrow does not disturb deals already in play. The card can keep moving without the open pipeline paying for the change.
A minimum policy you can adopt this quarter
- Name a rate card owner. Not a committee, a person. RevOps is the usual home.
- Split any role that is pricing from one number into a senior and junior line. At least one. Most cards need three.
- Add low, normal, and high margin bands per role. Make the low an approval gate, not a suggestion.
- Pick a rounding policy per seniority band and apply it once. Document the reasoning.
- Name the two or three triggers that rebuild the card. Calendar review is not one of them.
None of this is expensive. The expensive version is doing nothing and discovering, at the end of the year, that a third of the closed deals were quietly priced against a card that no longer represents how you deliver work.
Wrapping up
The test of a rate card is not whether it looks right when it is approved. It is whether a seller can quote a real deal from it at month nine without quietly asking finance for an exception. Most cards fail that test. The ones that survive are built on role granularity that matches how deals staff, margin bands instead of a single number, enough cards to cover the real shapes of work, and a review rhythm driven by hiring and pipeline rather than the calendar. Build it that way once, and the card becomes the thing the business negotiates with, not the thing the business works around.
