Guide · C99
Update marketing-scenario probabilities
How can probability be revised after positive or negative evidence without returning to intuition?
Direct answer
The update turns the prior probability into odds, applies each item of evidence’s diagnostic strength and converts the result back to probability.
01Keep a pilot running, strengthen the evidence, scale up or stop according to pre-set thresholds.
02The rule, the evidence and the next review date stay visible.
How to use this
Move from the question to the deliverable.
How can probability be revised after positive or negative evidence without returning to intuition?
Keep a pilot running, strengthen the evidence, scale up or stop according to pre-set thresholds.
Sequential probability calculation
Calculation · SQ-99
Revise the probability without erasing the starting point
Every item of evidence changes an initial probability according to its ability to tell success from failure. One negative result can cancel several weak positive signals.
Fictitious case: an initial probability of 40% that a new segment reaches its target economics, updated after three tests.
- 01Freeze
Document the initial probability, its source and its date before observing the test.
- 02Qualify
Estimate how much more likely the evidence is under success than under failure.
- 03Update
Multiply the previous odds by the likelihood ratio of the evidence.
- 04Decide
Compare the revised probability with the pilot, extension and stop thresholds fixed in advance.
Three items of evidence move the probability from 40% to 57%
Two positive signals lift the scenario to 77%, then a negative activation brings it back below the extension threshold.
Posterior odds = prior odds × likelihood ratio; probability = odds ÷ (1 + odds)| Step | Evidence | Likelihood ratio | Prior odds | Posterior odds | Probability | Decision |
|---|---|---|---|---|---|---|
| Starting point | Comparable history | — | — | 0.67 | 40% | Explore |
| Message test | Positive signal | 2.8 | 0.67 | 1.87 | 65% | Pilot |
| Price test | Positive signal | 1.8 | 1.87 | 3.37 | 77% | Prepare scale-up |
| Pilot activation | Negative signal | 0.4 | 3.37 | 1.35 | 57% | Correct the pilot |
01The extension threshold was set at 70%; the negative activation therefore forbids scaling.
02The probability stays above the stop threshold of 30%, which justifies a corrected pilot rather than an abandonment.
03The record keeps the prior of 40% and avoids presenting 57% as a direct or certain measurement.
Fictitious teaching case. Likelihood ratios must come from comparable data and be tested for sensitivity.
Coherence controls
Minimum conditions before validation.
The status, the evidence and the owner must be available for review.
The status, the evidence and the owner must be available for review.
The status, the evidence and the owner must be available for review.
The status, the evidence and the owner must be available for review.
Usable deliverable
Sequential probability calculation
Explore the decision further with an interactive control, or start from a structured template to produce your deliverable.