top of page

Bottom-Up vs Top-Down Forecasting: What AI Changes About the Debate

The old debate is over. Evidence-based forecasting won.

_______________________________________________

The Two Forecasting Philosophies

Top-down forecasting starts with a target and works backward. The board says $50M. The CRO divides that across regions. Regional VPs divide it across reps. Everyone works backward from the number they've been given, and the forecast becomes an exercise in gap math.


Bottom-up forecasting starts with deals and works forward. Each opportunity gets assessed independently. Deal-level scores aggregate into a pipeline-level prediction. The forecast comes from the data, not from the target.


In theory, both have value. In practice, most organizations run top-down and call it forecasting when it's really target allocation with a pipeline overlay.


Why Top-Down Forecasting Produces Bad Surprises

It's Target-Driven, Not Evidence-Driven

When the starting point is a revenue target, every downstream conversation is about coverage ratios and commit totals — not about whether individual deals are real. The question becomes "do we have enough pipeline?" instead of "is this pipeline going to convert?"

Reps Commit to Numbers, Not to Reality

In a top-down system, reps are pressured to hit a commit number. So they include deals they're hopeful about. Manager pressure amplifies the optimism. By the time the rolled-up forecast reaches the CRO, it reflects what everyone wants to be true, not what the deal evidence supports.

Risk Hides in the Aggregate

A top-down forecast might show $12M in commit with $18M in pipeline. That looks healthy at a 3x coverage ratio. But if 40% of that pipeline has stalled engagement and 30% lacks a confirmed economic buyer, the actual likely outcome is far lower. The aggregate hides deal-level risk.


What Bottom-Up Forecasting Requires

Pure bottom-up forecasting is more accurate, but it's also harder to execute manually. It requires:

  • Consistent deal-level assessment: Every opportunity needs to be scored on the same criteria. MEDDPICC is the most common framework for this.

  • Evidence-backed scoring: Scores need to reflect buyer-confirmed reality, not rep opinion. This means evidence from calls, emails, and documented commitments.

  • Frequent updates: Deal health changes weekly. A monthly forecast review using stale data defeats the purpose.

  • Aggregation without manipulation: The roll-up from deal scores to pipeline prediction needs to be mechanical, not subject to manager judgment calls that re-introduce bias.


Before AI, these requirements made pure bottom-up forecasting impractical for large teams. The data collection burden was too high.


How AI Makes Bottom-Up Forecasting Work at Scale

AI solves the three problems that made bottom-up forecasting impractical:

  1. Automated data capture. AI listens to every call, reads every email, and tracks every CRM update. Deal-level evidence is captured without rep effort. Spotlight.ai does this across all conversation channels and maps signals into structured qualification data.

  2. Continuous scoring. AI reassesses deal health with every new interaction. The forecast updates in real time, not at the next pipeline review.

  3. Objective aggregation. AI rolls deal-level scores into pipeline predictions using historical outcome patterns. The math is transparent and consistent. No manager can inflate the number by reclassifying a deal without corresponding evidence.

📊 Only 7% of sales organizations achieve forecast accuracy of 90% or higher. AI-powered bottom-up forecasting is the clearest path to joining that group. — Gartner, 2025

The Hybrid Approach That Works

The best-performing revenue organizations use AI-powered bottom-up as the primary forecast and top-down as a planning tool. The bottom-up model tells you what the pipeline will actually produce. The top-down target tells you where the gap is.

Spotlight.ai enables this hybrid approach by generating evidence-based, bottom-up forecasts directly in Salesforce while giving RevOps teams the pipeline health data to model scenarios against targets.


Bottom-Up vs Top-Down Forecasting

_______________________________________________


FAQs About Bottom-Up vs Top-Down Forecasting


Is bottom-up forecasting always more accurate?

When built on evidence-backed deal scores, yes. The accuracy advantage disappears if bottom-up scores are still based on rep self-assessment rather than objective data.


How do I transition from top-down to bottom-up?

Start by running both in parallel for one to two quarters. Compare predictions against actual outcomes to build confidence in the bottom-up model.


Does bottom-up forecasting work without AI?

In theory, yes. In practice, the manual data collection required makes it unsustainable for teams with more than 20-30 active opportunities. AI makes it practical at any scale.


What role does MEDDIC play in bottom-up forecasting?

MEDDIC provides the scoring framework. Each qualifier contributes to a deal health score. Those deal-level scores aggregate into the bottom-up forecast. Without a structured qualification model, there's nothing to score against.

_______________________________________________

Comments

bottom of page