Most firms build a rate card to express how they think about seniority. Procurement does not read it that way. It turns the card into something it can compare, and the comparison method is chosen before your document arrives.
Knowing which conversion is coming is most of the defence.
The evaluation machinery
Procurement starts by deciding what question it is answering. "Is this rate high?" and "can you actually deliver at this rate?" are different tests with different evidence.
For price reasonableness, the United States Federal Acquisition Regulation lists the playbook openly. Buyers may compare your proposal against competing bids, historical prices paid, published price lists and market prices, an independent cost estimate, market research, or supporting cost data. The first two, live competitive prices and historical prices, are the preferred techniques.
Broad bands against narrow bands
This is where procurement theory and procurement behaviour diverge, and the public bid protest record is unusually informative.
Give procurement many narrow bands and it does not necessarily weigh every row. It reduces the card. In the WJM Professional Services protest, dated 28 May 2024, the Coast Guard's pricing schedule contained 1,020 labour categories, of which 850 had zero estimated hours. The agency built a total evaluated price from the categories it expected to actually use. Granularity did not buy attention.
The opposite move is also survivable. In the CDO Technologies record, a supplier collapsed a more granular government hierarchy into two broad blended levels, and procurement did not reject the structure. It normalised the bands, understood the staffing implications and evaluated on that basis.
Taken together, those records kill both confident theses. Procurement does not demand granular cards, and more granularity does not force it to appreciate more nuance. There is no credible evidence that either shape systematically produces better commercial outcomes.
What benchmark data can tell you
There is a sourcing problem worth stating plainly. We found no current, public, primary-source dataset of commercial management-consulting bill rates broken out simultaneously by grade and geography. Commercial benchmark providers claim to hold this data, but the underlying observations are proprietary.
We would not convert a vendor's marketing claim into a headline market rate, and neither should a buyer. The best public comparator we found is Bureau of Labor Statistics labour cost data, which the General Services Administration itself instructs acquisition teams to use as a should-cost input.
The practical implication cuts both ways. If a buyer tells you your rates are above market, it is fair to ask which market and measured how. Frequently the answer is a proprietary dataset nobody can inspect.
Structures that survive scrutiny
A rate card alone does not resist commoditisation. The procurement record points at three stronger defences.
- Make the labour category falsifiable. "Principal" means nothing unless a buyer can see the experience, responsibilities, skills and type of work that justify the premium. The D&G record is the warning: a broad category was not inherently unacceptable, but the supplier could not demonstrate the crosswalk to the specific work.
- Show the economics at engagement level. A defensible total for a defined outcome resists line-item attack better than a defensible rate for an hour.
- State the assumptions the rate depends on. A rate premised on remote delivery, or on the client supplying data, is a different rate once those change. Saying so in the card prevents the argument happening later at your expense.
Seniority mix and named people
Procurement does not stop at the rate. It asks whether the person implied by the rate is the person you will actually deploy.
The FAR provision on professional employee compensation instructs evaluators to consider whether compensation is realistic and consistent with the skills required, the complexity of the discipline and the difficulty of the work. Unrealistically low compensation can be treated as evidence that the bidder does not understand the requirement, and a failure to provide an adequate compensation plan can justify rejection.
That is worth internalising, because it means undercutting on rate is not automatically safe. A rate low enough to imply you cannot staff the work as described is a scoring risk, not just a margin one.
Three questions to settle before you publish a card
Most rate cards are assembled from what the firm charged last year. These three questions produce a better one, and they take an afternoon.
Which comparison will this be subjected to? A card entering a competitive tender will be normalised against other bidders' categories. A card entering a sole-source negotiation will be compared against historical prices paid, quite possibly your own from a previous engagement. The two want different things: the first needs to survive a crosswalk, the second needs to explain any movement since last time.
What justifies each premium? For every band above your lowest, write the sentence you would say out loud if asked why that role costs more. If the sentence is about tenure rather than the work, procurement will treat the band as a title rather than a capability, and titles compress under pressure.
What is the smallest number of bands that survives? Given that granularity buys neither attention nor protection, the burden falls on each band to earn its place. A band you cannot staff distinctly, or explain distinctly, is a row someone will delete during normalisation anyway.
What would make this wrong
The uncomfortable evidence is the CDO Technologies record, and it is worth stating because it constrains what this article can claim.
A supplier collapsed a more granular government hierarchy into two broad blended levels. Procurement did not reject the structure, did not treat it as evasive, and did not penalise the bid. It normalised the bands, understood the staffing implications and evaluated the proposal on that basis. That kills any version of this article whose thesis is that procurement demands granular cards. It plainly does not.
The WJM and General Dynamics records kill the opposite thesis just as firmly. More granularity does not force a buyer to appreciate more nuance, and 850 categories with zero estimated hours got exactly the attention that number deserves.
What survives both records is not a shape. It is whether the buyer can map what you wrote onto the work they need done.
That is why the defences in this article are about falsifiability, engagement-level economics and stated assumptions rather than about band counts. Those hold regardless of which conversion procurement chooses, which is the only property worth optimising for when you cannot know in advance.
Common questions
Should we publish our rate card at all?
If you are selling into a process that will benchmark you regardless, a clear card that controls the crosswalk is better than one reconstructed from your invoices. If you are selling outcomes at engagement level, the card is a supporting document rather than the offer.
How do we handle a blended rate request?
Provide it with the assumed seniority mix attached. A blended rate without the mix is a number the buyer will re-derive under their own assumptions, which will not be yours.
Is it worth naming individuals?
Naming people strengthens the credibility of the rate and creates a commitment you have to honour. Name them where you can staff them and describe the profile where you cannot, rather than naming someone you intend to substitute.
What if procurement simply demands a discount?
Ask which comparator produced the target. Where the answer is a proprietary benchmark or a competitor bid for a different scope, the conversation can move back to what the work actually involves. Where it is a genuine like-for-like price, the card was never the problem.