Descriptive analytics reports what happened. Diagnostic analytics explains why. A descriptive output says win rate fell last quarter. A diagnostic output says the drop sits almost entirely in mid-market, where average selling price fell at the same time and losses to one competitor rose.
Most sales reporting stops at the first sentence, which is why most reporting generates a follow-up meeting instead of a decision.
The four analytics layers in order
| Layer | Question | Typical output |
|---|---|---|
| Descriptive | What happened? | Win rate, bookings, coverage, attainment |
| Diagnostic | Why did it happen? | Segment isolation, driver decomposition |
| Predictive | What happens next? | Forecast with a confidence range |
| Prescriptive | What should we do? | Ranked actions with expected effect |
Naming the mechanism, not the symptom
Diagnostic work is finished when you can name a mechanism. ORM describes the most common reason forecasts fail as something changing in the business or the market while the forecast still runs on old assumptions, and gives concrete mechanisms rather than symptoms.
A new competitor enters and creates pricing pressure, so average deal size falls. Interest rates rise, private equity firms slow capital deployment, portfolio companies cut costs, and fewer companies buy, so win rates fall. Market uncertainty means fewer decisions, so deals take longer from qualified to closed. Territories get redrawn and sellers get distracted, so execution suffers while the pipeline still looks fine on paper.
Each of those is testable. "Win rate declined" is a symptom and cannot be acted on. "Win rate declined because we now lose on price in one segment" points at pricing, packaging, or qualification.
The diagnostic loop
Start with the descriptive movement. Split it along one dimension at a time until the movement concentrates in a slice rather than spreading evenly. Then check whether driver metrics inside that slice moved in a direction that explains the result, and look for a plausible alternative explanation before accepting the first one.
Applied to a win rate decline, that means cutting by segment, source, deal size, and rep tenure, then testing whether cycle length, discount depth, or competitive presence moved in the same slice. Applied to deal slippage, it means separating deals that slipped once from deals that have slipped repeatedly, since those two groups behave differently.
Why the split matters for forecasting
Forecast accuracy depends on assumptions that match current conditions. Descriptive reporting tells you the assumption was wrong after the period closes. Diagnostic work tells you which assumption broke and why, which is the only version of the finding you can correct before the next period starts.Frequently Asked Questions
What is the difference between descriptive and diagnostic analytics?
Descriptive analytics states what happened, such as win rate declining last quarter. Diagnostic analytics isolates why, such as the decline sitting almost entirely in one segment where average selling price fell at the same time. Descriptive produces the number, diagnostic produces the cause.
Where do descriptive and diagnostic sit relative to predictive analytics?
They come first. Descriptive establishes what the data says, diagnostic establishes the mechanism, and predictive models what happens next. A predictive model built without the diagnostic step often extends a pattern whose underlying cause has already changed.
What techniques does diagnostic analytics use?
Segmentation, cohort comparison, and driver decomposition. The pattern is to split the aggregate along dimensions such as segment, rep, source, and deal size until the movement isolates in one slice, then check whether the driver metrics in that slice moved in a way that explains the result.
Why do most sales reports stop at descriptive?
Because descriptive output is easy to standardize and diagnostic work needs the underlying records, a defined set of dimensions to cut by, and someone willing to reject the first plausible explanation. Reports that only describe get read and then generate a request for the analysis that should have accompanied them.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like descriptive vs diagnostic analytics into prescriptive action for your team.
Schedule a Demo