Rules
Part of Model innovation: what to keep and what to drop
Model innovation checklist: sorting answers into three piles
A model innovation checklist ends in answers. Sort each into a document, a number or a capability, so every pile gets its own clock, owner and finish date.
| Pile | Contents | Working deadline | What completion requires |
|---|---|---|---|
| Documents | Constraint, change rule, exemptions, evidence and stop condition | This week | Written statements tied to a measure |
| Numbers | Realized price, effective rate and the arithmetic the new model requires | Before the next meeting | Calculations using actual records |
| Capabilities | Work the company cannot yet do with its people or systems | Set by a plan | A named owner and a date for each item |
Begin with the smallest useful decision record
Four entries are enough for a small company to begin: its constraint and measure, what the current model really earns, the arithmetic the proposed model requires, and the evidence that would lead to stopping. Other questions can wait until more teams or systems need to coordinate.
- State the constraint and how it will be measured.
- Calculate what the current model actually earns.
- Work out the arithmetic on which the proposed model relies.
- Define the evidence and the stop condition.
Put the written decisions on a short clock
Describe the constraint in a single sentence and include the measure that would demonstrate it. That measure separates a real limit from an ambition or a confident impression.
Possible constraints already in the decision include a calendar, a sales channel, a cost that will not decline, or the buyer's approval process.
Before You Decide: Four Checks
- Constraint in one sentence, with measure
- Has anything cheaper been tried?
- What does current model actually earn?
- Which business type are you running?
For every measure, record what evidence supports it, the direction it points, and the period it covers. Include confounding factors already known, then decide what result would bring the change to a halt. Set that rule before launch, so commitment to the decision does not outlast belief in its case.
Compute the model's dependencies from records
Use actual invoices and other records to calculate realized price against list price, or effective rate against quoted rate, before redesign work begins. The business description and the billing description can differ; business model types offers classification context, while business model axes help distinguish descriptions that may otherwise be blurred.
Next, calculate the sum the new model needs to achieve rather than treating it as a forecast. A design relying on recurring revenue depends on a loss rate; one relying on volume depends on the cost of discounts. These are thresholds to clear. The revenue constraints can help frame the arithmetic that rules options out.
Arithmetic Behind the Destination
- 3%monthly loss rate leaves 69% of cohort after a year
- 33 monthsimplied average customer life
- 40%gross margin at which 10% off needs 33% more units
If the calculation shows that the current arrangement performs better than expected, a replacement may not be needed. Finding that out with an afternoon of work is a useful outcome in itself. If a less disruptive adjustment could address the issue, test that before changing the model: a price increase, discount threshold or packaging change may be enough. The layers sit within pricing architecture.
Give capability gaps an owner and a date
Put every missing skill or system on a plan, with one responsible person and a deadline beside it.
This is often where the outcome is decided: the change may depend on a year of unglamorous work that no one has actually scheduled. A decision without provision for that work is not yet an executable plan.
Name the teams expected to change their behaviour, including sales, support, finance and delivery where relevant. Their work will be affected by the new arrangement, so include them in the plan before launch rather than informing them once it is already in motion. A model change requires people to do things differently.
Keep transition funding separate from the expected return of the new arrangement. The transition may involve two structures operating together and a sales team offering both. Funding the change with returns it has not generated yet creates a timing problem; MIT's entrepreneurial finance course explains why timing can matter more than the total amount.
Scale the evidence to how hard reversal is
Gather more evidence for a change that is difficult to undo than for one that can be reset. Some arrangements can be changed the following week.
A customer who has received a lower quote may not accept the previous structure, and invoices already issued on a new billing basis cannot simply be reversed.
Reversibility Checks
- If wrong, what does undoing cost?
- Can it run on a bounded group first?
- What are you deliberately holding constant?
If possible, test on a bounded group first, such as one segment, one region or new business only, and write down when the test ends. If the change cannot be narrowed, treat that as a reason to gather stronger evidence before committing rather than as a detail to resolve later.
Choose one element to hold constant: product, segment or delivery. Keeping one fixed makes it possible to attribute the result instead of explaining it afterward with a story. Strategic management teaching treats this choice as central; MIT's strategic management course discusses why the distinction matters.
