The 14% revenue opportunity hiding inside one rate card
One rate card can conceal several markets with very different responses to price.
The pricing decision was where to spend margin to change buying behavior, and where not to.
Executive pricing conversations often collapse into one choice: raise, hold or discount.
That is usually the wrong level of decision for a bid-driven business. A company does not buy growth by lowering an average price. It gives up margin on specific opportunities in exchange for a better chance of winning them. The commercial question is whether that exchange creates value, and in which parts of the book.
At one specialty contractor, roughly three years of bids showed why the distinction matters. The company-wide numbers looked coherent. Once the bids were separated by job scale and operating conditions, they revealed several markets with very different responses to price.
A historical replay applied bounded price moves only where lower prices had been associated with meaningfully better conversion. It projected 14% more probability-weighted revenue and 9% more gross profit.
Those are modeled figures, not realized performance. The five-percentage-point gap between them is still instructive. It is the price of buying additional conversion: projected revenue rose faster than gross profit because selected bids surrendered some margin.
The opportunity was therefore more precise than ‘charge less.’ Most bids did not need a new price. A small part of the portfolio needed a better commercial rule.
The historical replay projected 14% more probability-weighted revenue and 9% more gross profit. Projected, not realized.
Portfolio averages cannot price the next job
Company averages answer an important question: how did the portfolio perform? They do not answer the next one: what should this bid cost?
No average customer accepts an average quote for an average job. A particular quote reflects a particular scope, customer type, travel burden and physical complexity. The buyer also has a particular set of alternatives. Two bids that sit beside each other in a monthly report can belong to different economic markets.
The useful unit of pricing policy is therefore a decision segment. A good segment has four properties. The business can identify it before the quote goes out. Operators recognize it. It contains enough repeatable work to measure. Its economics or buying behavior differ enough to justify another rule.
That standard matters. A variable discovered only after close cannot govern a quote. A statistically tidy group that sales cannot recognize will not survive in the field. A tiny leaf with an extreme result is evidence to investigate, not a market to price.
When the contractor’s bids were separated this way, the company-wide win rate broke into sharply different local patterns. Lower prices were associated with much better conversion in some groups and almost no improvement in others. The same rate could be restrictive in one pocket, sensible in another and unnecessarily generous in a third.
Before changing price, leadership needs to decide what the business is actually pricing.
Nearly two-thirds of the modeled gross-profit opportunity came from just two job groups.
Treat margin concessions like investment capital
The modeled opportunity was highly concentrated. Two job groups accounted for 63% of the projected gross-profit improvement. One contributed nothing.
That concentration changes the executive decision. A discount is not a sales gesture. It is an investment of margin, made now, in the hope of buying additional conversion later. Like any scarce capital, it should go where the expected return is highest.
A blanket discount would have deployed that capital across the entire portfolio, including bids that were already competitive and groups where price movement appeared to buy little. A blanket increase would have protected margin while missing the few markets where a lower quote could create more value than it surrendered.
Pricing teams also have limited testing capacity. Every additional rule requires training, controls, measurement and enough volume to learn from. Concentrating first on the two strongest pockets improves both the economics and the quality of the evidence.
The 14% headline did not require making the company cheaper. It required directing commercial flexibility toward the small number of situations where it had a plausible return.
That is a more useful approval request for leadership: not a company-wide rate change, but a governed pricing policy for defined segments.
Revenue growth has an acquisition cost
A price reduction has an immediate, certain cost. The company gives up margin on every job it would have won at the old price.
The benefit is conditional. A lower quote may convert work that would otherwise have been lost, or it may simply make an existing win less profitable.
This is why win rate is a dangerous executive target on its own. It allows the commercial organization to buy growth with margin while the dashboard reports success. The useful starting measure is expected gross profit per opportunity:
chance of winning × gross profit if won
A discount earns its way back only when the increase in win probability covers the contribution surrendered. If operating capacity is constrained, the denominator should become tighter still: expected gross profit per crew day, machine hour or other scarce unit. Winning more low-quality work can crowd out better work and make a positive quote-level result unattractive at portfolio level.
The gap between the projected 14% revenue gain and 9% gross-profit gain makes this visible. Selected bids traded margin for conversion. The modeled trade remained positive, but the growth was not free.
Leadership should therefore ask three questions of any pricing move: How much additional conversion are we buying? What contribution remains after we buy it? Which scarce resources will the new work consume?
‘Win rate improved’ answers none of them.
Illustrative curves: only a genuinely price-responsive segment earns the discount back.
A rate card and a model are not a pricing policy
The decision trees did not discover a magic price. They separated the bid history into combinations of conditions with different outcomes.
That is useful pattern detection. It is not commercial governance.
A deployable pricing policy needs more than a prediction. It needs an eligibility rule that can be applied before quoting, an allowed price band, a margin floor, an owner for exceptions, a record of overrides and a review date. Without that policy layer, a model either becomes unused advice or creates discretion the business cannot audit.
Human review matters for another reason. Historical quote prices are partly a record of seller judgment. Teams often discount the opportunities they already believe will be difficult to win. A tree can rediscover that judgment and mistake it for a customer response. Operators must decide whether a pattern reflects the market, the estimating process or a change in the work itself.
The strongest rules in this case described recognizable work, had plausible buying logic and could be bounded. That made them suitable for a test.
The model’s output was a map. The executive output should be a controlled decision: where the rule applies, how far price may move, who can override it, what success means and when the business will stop.
A projection should end in a governed test
Historical bids come with a dangerous advantage: we already know which ones were won.
That creates selection risk. Lower-priced jobs may also have involved different customers, sellers, competitors, timing or operating conditions. Price itself may have been lowered because the seller expected a difficult contest. Segmentation can reduce this confusion. It cannot prove that the price caused the conversion.
The 14% revenue and 9% gross-profit figures came from replaying historical bids under different rules. They identify an opportunity worth testing. They do not establish a realized result.
A credible rollout should begin with one or two high-potential segments and a concurrent comparison group that keeps the current rule. Leadership should approve the exposure in advance: the permitted price band, maximum margin at risk, test duration, minimum evidence and stop conditions.
The measurement should separate leading and lagging indicators. Quoted price, win rate and expected gross profit appear quickly. Realized gross profit, cost-to-serve, rework and capacity consumption arrive later. A policy that wins attractive work on paper can still disappoint when completed-job economics close the loop.
If the new rule improves conversion without giving away the projected economics, expand it carefully. If it does not, stop. Either result is more valuable than another quarter of debating the company average.
What the executive team should ask for
The next pricing review should fit on one page per decision segment.
Show quote volume, median price, win rate, expected gross profit per opportunity and realized gross profit for completed work. Add the relevant capacity measure if labor, equipment or working capital constrains growth. Display enough history to distinguish a persistent pattern from a noisy month.
Then ask five questions:
Where is there credible evidence that price changes buying behavior?
How much margin must be invested to capture the additional conversion?
Does the work remain attractive after capacity and cost-to-serve are considered?
What guardrails limit the downside while the company learns?
What result, by what date, will cause us to expand, revise or stop the rule?
This changes the meeting. Sales no longer asks for broad discretion. Finance no longer defends one margin threshold across unlike jobs. Operations can identify whether the proposed growth consumes scarce capacity. Leadership approves a measurable commercial policy instead of a general view on whether prices feel high.
A single rate card may remain useful to administer the business. It should not persuade executives that the business operates in a single market.
Comments