Optimized Sales Optimized Marketing Target Accounts For CROs For CFOs For CMOs Blog News Glossary Compare Tools About Schedule a Demo
Sales Forecasting

Hybrid Forecast Model

ORM Technologies
Home/ Glossary/ Hybrid Forecast Model
Definition A hybrid forecast model combines a bottom-up roll-up of individual deals with a top-down statistical projection, then reconciles the difference between them into one committed number.
A hybrid forecast model runs two independent forecasts and reconciles them. The bottom-up half rolls up individual deals with their close probabilities. The top-down half projects from historical conversion rates, seasonality, and capacity without looking at any single deal. The committed number comes out of explaining the gap between them.

Each half has a known failure mode, and they fail in opposite directions. Deal roll-ups inherit rep optimism and stale close dates, so they run hot. Statistical projections extend the recent past, so they miss a shift in either direction until it shows up in the numbers. Running one method alone means owning its bias with nothing to reveal it.

What each half contributes

The bottom-up half supplies specificity. It names the deals, which means every forecast number has an owner and an action attached to it. It is the only half that can answer what to do this week.

The top-down half supplies memory. It knows that pipeline entering the quarter has historically converted at a given rate and that certain months close harder than others. ORM's data shows the calendar effect clearly, with Q2 and Q4 stronger than Q1 and Q3 and the third month of a quarter closing more than the first two. A roll-up built in week two has no way to price that in. The statistical half does it automatically.

Reconciling the gap

The gap is the product of the model, so give it a written explanation every cycle rather than splitting the difference.

Start with what the bottom-up half structurally misses. ORM's decomposition of a quarter names three revenue sources: deals already in pipeline on day one that close this quarter, deals created and closed entirely inside the quarter, and deals pulled forward from future periods. A deal roll-up on day one can only see the first source. Teams that forecast from the roll-up alone systematically under-model in-quarter creation and understate what pulling future deals forward costs the following quarter.

Then check the direction of the gap. When the roll-up runs above the statistical projection, the usual cause is deals valued higher in the CRM than they close for. ORM's point is that deals routinely close for less than the value carried in the CRM. Take a pipeline with an $80,000 average deal size that produces $40,000 average closed-won deals, which is enough to invalidate a roll-up on its own. When the roll-up runs below, in-quarter creation is doing work the historical model already knows about.

Keeping the model honest

Two habits keep a hybrid model from decaying into a rubber stamp.

Log both raw numbers before reconciliation, every quarter, without editing. Once you have a dozen quarters, you know which half deserves more weight at week two versus week ten, and the weighting stops being an opinion.

Second, resist a coverage ratio as the tiebreaker. ORM reports 3x to 5x coverage as the standard range with most customers near 3.5x, and treats the ratio as an input rather than a conclusion. A team can hold 4x and still miss badly when the pipeline is aged, concentrated, or sitting in the wrong segment. Reconcile on composition instead. Compare the two forecasts against pipeline coverage and a running forecast accuracy record, and the hybrid number becomes defensible rather than negotiated.

Frequently Asked Questions

How do you weight the top-down and bottom-up numbers?

Weight by measured accuracy rather than by preference. Track both methods against actuals for several quarters, then weight each by how close it has historically landed. Expect the statistical projection to carry more weight early in the quarter and the deal roll-up to carry more in the final weeks, then confirm that against your own logged results before fixing the weights.

What do you do when the two numbers disagree?

Treat the gap as the finding rather than as a problem to average away. A roll-up above the statistical projection usually means rep optimism or stale close dates. A roll-up below it usually means unmodeled in-quarter creation. Each explanation points at a different action, and averaging destroys the signal that told you which one applies.

Is a hybrid model the same as consensus forecasting?

No. Consensus forecasting reconciles the judgments of different people, such as reps, managers and finance. A hybrid model reconciles different methods, such as a deal roll-up and a statistical projection. A mature process runs both, since agreement between people means little if every person is reading the same optimistic pipeline.

Does a hybrid model reduce forecast bias?

It reduces one-directional bias because the two methods fail differently. Rep roll-ups lean optimistic, and statistical projections lean toward the recent past. Holding both visible makes a persistent lean in either one measurable and correctable, which a single method cannot do because it has nothing to be measured against.

Put these metrics to work

ORM builds custom revenue forecast models that turn concepts like hybrid forecast model into prescriptive action for your team.

Schedule a Demo