The problem
Executives had defined a new pricing strategy, but the rollout failed. The rules existed on paper; applying them consistently across every service, customer segment, and division in the price books that CPQ and Salesforce consume was manual, slow, and error-prone.
Pricing errors reached customers, and fixing them meant paying contractors to rework price books by hand each time a rule changed. The business needed pricing that was correct by construction and could change as fast as the executives changed their minds.
What I did
-
Treated pricing rules as code. Sat with the executives who owned the strategy, translated each rule into an explicit, testable definition, and kept those definitions in one place so a rule change is a code change with a history.
-
Designed a modular architecture, 48 modules in total: data ingestion from the source systems, rule definitions, the pricing engine that composes them, validation against expected outcomes, output generation in the formats CPQ and CRM load, and an audit trail of what was priced and why.
-
Built validation into the pipeline so no price book publishes until every rule passes and the results reconcile against the previous version. Differences are surfaced for review rather than silently shipped.
-
Calibrated the rules on evidence rather than opinion: outlier screening on the sales history that feeds the engine, and a price-elasticity study on the order-line panel so the executives could see how much volume a given increase was likely to cost before approving it.
-
Wrote the program to be run and maintained by one person, with documentation, tests, and structured outputs, so it does not depend on a contractor engagement to stay current.
- Ingest Sales history and service data from the warehouse and source systems.
- Screen Outlier screening and filtering on the history the engine learns from.
- Classify Facilities and price-book changes sorted into the classes the rules address.
- Price Reference and realized prices, adjustments, aggregation: the rules as code.
- Validate & reconcile Every rule passes and the book reconciles against the last version, or nothing publishes. gate
- Export The formats CPQ and the CRM load.
- Audit What was priced, by which rule, and why.
Stage names describe the program at the pattern level. The rules themselves are confidential.
Outcome
$15.7M of revenue was priced through the program in its first year.
Pricing errors were eliminated after cutover, and contractor spend on price-book maintenance was reduced because rule changes no longer require manual rework.
Executive pricing decisions now have a direct path into the systems reps quote from, which closes the loop between strategy and execution.
In the stack
Where it sits
The highlighted nodes are the systems and people this project touched.