Actual mentorship contact and cost
Calculate rates from distinct delivered, attended and reviewed minutes without turning tuition into an investment-return forecast.
Keep payment and contact definitions separate
A package cost and a scheduled-hour count describe different records. First identify which charges and refunds enter the package scope. Then identify what the session rows mean: listed time, completed delivery, attendance and individual review. The authored tool assumes entered fees belong to the full listed package, even when some slots remain planned. A current rate can therefore change as more contact is observed. It is not a fixed invoice allocation per session. Preserve the fee population and scope before comparing two offers. The fictional XCU example is not a programme quote, a bank receipt or an estimate of typical coaching prices.
Reconcile the time fields
A completed row records whole-minute delivered time, attended time and review time. Review is a subset of attendance, and attendance a subset of delivery in this exercise. Planned and cancelled rows require null actual observations. They remain in the listed schedule totals without adding contact. Completed delivery may exceed scheduled time; an overtime observation is not capped to make the plan appear exact. The exercise does not authenticate the minutes or determine whether the terms allowed the overtime. The useful record keeps plan and observation separate, including zero delivered minutes on a completed row when that is the supplied observation rather than inventing a positive duration.
Use the right denominator
The base teaching package has 350 units of entered charges, ninety delivered minutes, sixty attended minutes and thirty review minutes. That gives 233.333333 units per delivered hour, 350 per attended hour and 700 per review hour. The rates share a numerator but not a denominator. None measures learning quality or future returns. A comparison that selects the largest time population to produce the lowest advertised rate hides the service definition. The ledger reports all three. Zero time produces a null rate rather than infinity or a fabricated free-service claim. Keep those nulls when exporting or discussing a no-attendance case.
Treat refunds as entered status evidence
The tool subtracts entered refunds from entered charges and rejects a refund total above charges within its bounded exercise. That is a modelling limitation, not a universal payment rule. A real credit needs its own applicable agreement and settlement evidence. The full-refund variant simply changes the authored fee population so net cost becomes zero. It does not claim a provider offers that refund or decide eligibility. A zero cost with zero contact still has undefined hourly rates. Another product with pending credits, mixed currencies or unrelated compensation needs a different record rather than silently forcing its payments into this package definition.
End with service arithmetic, not ROI
The resulting rates can make a package comparison easier to inspect, but they cannot tell a buyer what to spend, whether the offer is suitable or whether mentoring will pay for itself. Opportunity cost, taxes, currency conversion and unentered obligations remain outside the tool. The cost of a useful critique also cannot be judged from duration alone. Keep a sample-task review and the contact ledger as separate evidence layers. A detailed learning trail may show an improved explanation while leaving trading validation unknown. No pass probability, investment return, learner score or causal improvement estimate is added to make the numbers look more persuasive.
An exact denominator check
Use the fictional package's ninety delivered minutes as one denominator and thirty review minutes as another. Converting both to hours gives 1.5 and 0.5. Dividing 350 by them yields different rates without changing any fee. The difference is not evidence that one format teaches better. A source record should say which minutes were observed and whether they represent the learner's work. Keep the cancelled and planned rows with the reconciliation so listed access is not silently treated as actual review.