The short answer
Because a single number that ignores scarcity and seniority is not one you can plan against. The same platform skill costs very different amounts depending on whether the engineer leads or executes, how few people have run the platform in production, and constraints such as clearance or residency. Conntal publishes the rate method instead.
A rate card implies that a RabbitMQ engineer has a price. In practice the price depends on three things a card cannot capture: whether the engineer leads the change or executes it, how many people have actually run that platform in production, and whether constraints such as clearance, US-person requirements or data residency narrow the pool.
Scarcity in particular moves the number. For messaging and Cloudera work its effect is substantial, which is exactly where a published average would mislead most.
So rather than a number, Conntal publishes the method — the factors that set a rate and how each one moves it — on the delivery models page. That lets you sanity-check any quote you receive, from Conntal or from anyone else. The location model changes the rate; it never changes the screen.