What is bottom-up forecasting?
Bottom-up forecasting is a way to build a forecast from the smallest parts you can count, such as single deals, reps, products, or customers, then add them up to a total. Top-down forecasting runs the other way and starts from a target or market size. In sales, the parts are open deals and the reps who sell them. Updated October 2026: added the finance meaning, a model table, a worked example and Pete Furseth's three revenue sources.The principle is simple: estimate what each unit will produce, then sum. Because every dollar in the total traces back to a deal, a rep, or a documented rate, a bottom-up forecast is the easiest kind to defend and the easiest to act on in the current quarter.
What is a bottoms-up model?
A bottoms-up model is the same idea used for any projection: build revenue or cost from the drivers that make it, then roll them up. Finance teams write it as volume times price, split by product, region, or channel. Sales teams write it as deals times the odds of closing, plus rep capacity. Supply chain teams forecast each item on its own and add them.Pick the unit and the formula follows:
| Model type | Unit | Formula | Best for |
|---|---|---|---|
| Pipeline forecast | Open deal | Sum of (deal amount x probability of closing this period) | Current and next quarter calls |
| Full sales forecast | Revenue source | Carry-over deals + in-quarter deals + pull-forward deals | Quarterly forecast that covers deals not yet in CRM |
| Capacity forecast | Rep | Sum of (reps x productive quota per period) | Hiring plans, annual plans |
| Unit-economics model | Product or plan | Sum of (units x net price) | Budgets, operating models, investor models |
| Customer model | Account | Starting ARR + new + expansion, minus churn and contraction | ARR plans, retention planning |
| Demand forecast | Item per location | Sum of item-level forecasts | Inventory and supply planning |
How is a bottom-up sales forecast built?
Score each deal, weight it by your own historical conversion, sum it, then add the revenue that is not in the CRM yet. The deal-level work follows three steps: Step 1: Deal-level assessment. For each deal in pipeline, evaluate: - Current stage and how often deals at that stage close - Deal amount, set against what deals like it have closed for - Expected close date, and whether it has already moved - Risk factors (champion status, executive engagement, competitive position) Step 2: Probability weighting. Apply historical stage conversion rates to each deal. A deal in evaluation with a 30% historical conversion rate and a $100K value contributes $30K to the weighted pipeline. Step 3: Aggregation. Sum the weighted values by rep, by segment, and in total. Many teams then group deals into commit, best case, and upside categories so leaders can see how much of the number rests on high-confidence deals.That covers deals you can see. Pete Furseth, ORM's COO, argues that most teams stop there and miss the revenue they cannot see yet. A complete bottom-up forecast splits the quarter into three sources:
1. Carry-over deals: already in pipeline on day one and expected to close this quarter. 2. In-quarter deals: not created yet, but will be created, qualified, and closed inside the quarter. 3. Pull-forward deals: deals from future periods that close early, often with a discount and a cost to next quarter.
Illustrative worked example. Target: $4.0M. Day-one pipeline with close dates this quarter: $10.0M, which is 2.5x coverage. - Carry-over: Pete says roughly 20% of the pipeline value dated to close in a quarter on day one usually closes in it (it varies by customer). $10.0M x 20% = $2.0M. - In-quarter: the same quarter in prior years averaged $1.2M created and closed inside the quarter. Add $1.2M. - Pull-forward: $0.3M of next quarter's deals that could close early after a discount. Add $0.3M, and note the hole it leaves next quarter. - Forecast: $2.0M + $1.2M + $0.3M = $3.5M, or 87.5% of target. The gap is $0.5M, and it is visible on day one, not in week 13.The step-by-step build guide covers each layer and the capacity check in detail.
How does bottom-up compare with top-down forecasting?
Top-down divides a target down; bottom-up sums the parts up. The short version:| Bottom-up | Top-down | |
|---|---|---|
| Starts from | Deals, reps, products | Target, growth rate, market size |
| Best horizon | Current and next quarter | Annual plan, long range, new markets |
| Main risk | Inherits pipeline bias and misses deals not yet created | Ignores what the pipeline actually shows |
| Output | A number you can trace deal by deal | A number you can set fast |
Why bottom-up forecasting matters for revenue teams
Bottom-up forecasting exposes the gap between target and reality at the deal level. If your top-down target is $5M and your bottom-up forecast shows $3.8M (illustrative), you know the gap is $1.2M and can name which deals need to close, which need to move faster, and how much new pipeline is required.That only helps if you see it early. "Getting the forecast right in the last week of the quarter does not help anyone," Pete says. "By then, the quarter has already happened." The value is knowing the likely shape of the quarter on day one.
It is also what earns a board's trust. Pete's test: the forecast rolls up consistently from region to total, you can drill from any number down to the deals behind it, and it shows a range (a low case, an on-target case, and an upper bound) rather than one point. A bottom-up build gives you all three. Forecast accuracy improves because the forecast is grounded in deals you can see and influence. There is room to improve: Gartner puts median sales forecast accuracy at 70% to 79%, with only 7% of sales teams at 90% or better.
How to improve bottom-up forecasting
- Calibrate probabilities to your data. Do not use generic stage probabilities. Work out your own past win rates by stage, segment, and deal size, and update them each quarter. Your rate for a stage can sit well below the benchmark someone quoted you. - Use different weights for different motions. Pete names this as the second reason stage weights lie: one set of weights applied to new business, expansion, and renewal, or to enterprise and SMB, when each converts differently. Stage weights also fail when stage entry and exit criteria are loose and reps decide the stage. More in why stage-weighted forecasting misses. - Layer in deal-level signals. Stage alone is a crude signal. Adjust probabilities up or down based on champion activity, multi-threading status, and time-in-stage against the median. Watch what changes as well as where a deal sits. A deal that moved into Proposal yesterday tells you more than one that has sat there for months, and Pete calls a changed close date the best signal that a deal will slip. - Validate with top-down forecasting. When your bottom-up forecast drifts far from past booking trends, find out why. Either the bottom-up number carries rep optimism, or the top-down trend has missed a real shift in the market. - Forecast by rep as well as by deal. Some reps call their numbers well and some run high or low every quarter. Measure each rep's history and correct for rep-level bias. - Add seasonality by week. ORM sees the third month of a quarter as the largest and the first as the smallest. One example 13-week profile from Pete weights week 13 at 2.27 against 0.70 for week 1. Without it you cannot tell whether you are ahead or behind. See how to add seasonality to a sales forecast. - Do not wait for clean data. "Garbage in does not have to equal garbage out," Pete says. If your CRM is imperfect in a consistent way, a model can learn the pattern and still forecast accurately.
Common mistakes with bottom-up forecasting
Trusting deal amounts without validation. Most deals close for less than the amount in the CRM. Pete's example: a pipeline with an $80,000 average deal size where closed-won deals average $40,000. Weight amounts by what similar deals actually closed for, not by what reps entered. More on why deals close for less than forecast. Treating the visible pipeline as the whole forecast. A bottom-up forecast built only from today's CRM misses the in-quarter deals that will be created and closed before quarter end, and it treats coverage as if it were the forecast. For future quarters even more of the bookings will come from deals not yet created, so blend bottom-up with pipeline generation forecasts. Building on last year's assumptions. The most common reason a forecast misses, in Pete's view, is that the business or market changed and the model did not. A new competitor lowers deal size, buyers delay decisions and cycles stretch, a territory change distracts sellers. In 2026, ORM has seen longer cycles as buyers delay decisions while they wait to see what AI will change. Refresh conversion rates, deal sizes, and cycle lengths as they move, not once a year.Frequently Asked Questions
What is a bottoms up model?
A bottoms up model (also written bottom-up model) projects revenue or costs from operating drivers and sums them to a total. In finance that means units times net price, or customers times revenue per customer, by product or region. In sales it means open deals weighted by their chance of closing, plus the deals still to be created, plus rep capacity. The same idea: build the total from the parts.
What is the difference between top-down and bottom-up forecasting?
Top-down forecasting starts from a big number, such as a revenue target, last year's growth rate, or a market size, and allocates it down to segments and reps. Bottom-up forecasting starts from individual deals, reps, or products and adds them up. Top-down is fast and works without pipeline. Bottom-up is slower but traceable to specific deals. Most revenue teams run both and study the gap.
What is a bottom-up analysis?
A bottom-up analysis looks at a business by starting with its parts, such as deals, accounts, products, or line items, and working up to the total. A top-down analysis starts with the total and breaks it apart. A bottom-up forecast is one kind of bottom-up analysis. It estimates what each part will add to future revenue, then sums the estimates.
How accurate is bottom-up forecasting?
It depends on the inputs. Gartner reports that the median sales forecast accuracy is 70% to 79%, and only 7% of sales teams reach 90% or better. Pete Furseth, ORM's COO, puts the usual level on new and expansion business near 90%, reached with heavy manual effort. ORM targets 95% accuracy with a model that updates from day 1 to day 90 of the quarter.
What data does a bottom-up forecast require?
Deal-level data: stage, amount, close date, and the dates those fields last changed. Historical data: stage conversion rates, closed-won amounts versus pipeline amounts, cycle length by segment, and how much revenue is created and closed inside a quarter. Rep-level data: attainment history and ramp status. The data does not need to be perfect. It needs to be consistent.
When should you use bottom-up vs. top-down forecasting?
Use bottom-up for the current and next quarter, where you have deal-level visibility and can act on specific deals. Use top-down for annual plans, board targets, and new markets with no pipeline yet. For anything in between, run both: the gap between the bottom-up number and the top-down plan tells you how much pipeline or capacity you still need.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like bottom-up forecasting into prescriptive action for your team.
Schedule a Demo