Sales forecasts and demand plans are built for different jobs. A pipeline forecast rolls up deal-stage probabilities to track progress toward a revenue target. A demand plan sizes the production run, the purchase order and the stock position. Both are useful, and the trouble starts only when one number is asked to do both jobs at once, which happens more often than anyone plans for. Across many businesses, the pipeline figure arrives first, carries the sales director’s signature, and nobody builds a second number. Operations then plans a month of production on a figure built to motivate a sales team. Telling the two apart, and checking whether the sales input helps, protects both teams.
What Does Sales Forecasting Software Measure?
A pipeline forecast rolls up deal-stage probabilities and rep-reported close estimates into a revenue number. The number exists to track progress against a target and to steer sales activity. Its purpose gives it a structural lean toward optimism, and nobody has to be dishonest for that lean to appear.
Three properties matter for operations.
First, it carries revenue, and a plant builds units. A pipeline worth $18 million in revenue for a detergent range says nothing about how many 250 g, 500 g, and 1 kg packs that means. Converting revenue to units needs a mix assumption, usually last year’s. The mix is exactly what moves.
Second, it updates when deals move stage in the CRM. A rep marks a deal “verbal commit” on the 20th, and the forecast jumps. No customer placed an order that day. Demand stayed put.
Third, it covers the accounts in the pipeline. The base business underneath, such as repeat orders from 400 small distributors who never appear as deals, sits outside it. When that base is half the volume, the pipeline forecast describes the other half, and the plant guesses the rest from history someone pulls by hand at month end.
What Does Demand Planning Require?
Demand planning combines a statistical baseline with sales input, promotional plans and cleaned demand history into a consensus forecast sized for operations. The design goal is accuracy at the level and horizon each decision needs. Revenue tracking is a separate job.
In practice, that requires three things. Units by item and location, since plants make and ship units. A horizon covering the cumulative lead time of the longest item, so purchasing has a forecast for the weeks it is buying. And a history cleaned of supply effects, so stock-out weeks and month-end loading do not pass for demand.
The sales forecast is one input to this process, often a valuable one, so at this point the two numbers stop competing. Salespeople hear early. A new listing or a lost account reaches them weeks before the history shows it. The question is how much weight that input deserves. Direction matters too. Our guide to demand planning covers all five inputs and how they combine.
What Does the Research Say About Sales Adjustments?
The published research is more precise than the general claim that sales numbers run high, and more useful for an operations team. Read closely, it shows which adjustments to trust and which to question.
The largest field study looked at more than 60,000 forecasts across four supply chain companies. Robert Fildes, Paul Goodwin, Michael Lawrence and Konstantinos Nikolopoulos published it in the International Journal of Forecasting in 2009. Across their data, judgmental adjustments to statistical forecasts behaved very differently by direction and size. Upward adjustments improved accuracy far less often than downward ones. Large adjustments tended to help, while small ones tended to damage the number.
An earlier field study of judgmental sales forecasting in 13 manufacturing organisations reached a related conclusion. Michael Lawrence, Marcus O’Connor and Bob Edmundson published that study in the European Journal of Operational Research in 2000. The company forecasts they studied were not uniformly more accurate than a simple naive forecast. Much of the error came from bias and from errors that repeated period after period. Accuracy, the authors noted, may not have been the forecast’s only objective.
The point about objectives explains the pattern, because a number that also serves as a target, a motivator or a negotiating position will drift from the most likely outcome.
The practical reading for operations is narrow. The useful question is which direction of adjustment to question hardest. Large downward changes from sales deserve attention. Small upward nudges deserve the most scrutiny.
Where Do Sales Forecasts and Demand Plans Diverge?
Five dimensions pull the two numbers apart. Each one changes what operations should do with the sales figure. The table gives the short version, and the prose under it carries the operational detail.
| Dimension | Sales pipeline forecast | Demand plan |
|---|---|---|
| What it is built for | Tracking progress toward a revenue target | Sizing production, purchasing and stock |
| Unit of measure | Revenue by account | Units by item and location |
| Typical direction of error | Optimistic, following deal confidence | Corrected across several inputs |
| Update trigger | CRM stage changes | Actual demand and the planning cycle |
| Coverage | Accounts in the pipeline | Total demand, base business included |
What it is built for
A revenue forecast succeeds when the sales team hits the target. Success for a demand plan means the right stock in the right place at the lowest cost. The two goals overlap less than they seem. A director who forecasts 5% high to stretch the team is doing that job well, yet any plant building to that number carries the extra 5% as stock.
Unit of measure
Unit of measure causes the most damage in practice. Revenue hides the mix. Take a quarter. The pipeline shows $18 million in revenue, the same as last year. Planning splits it by last year’s mix: 40% in 1 kg packs, 45% in 500 g and 15% in 250 g. Meanwhile a large retail account has moved to 250 g packs, so the real mix runs closer to 30%, 40% and 30%. Actual revenue landed within a percent of the forecast, so on the sales review the quarter looked perfect. The plant built a third more 1 kg packs than it needed, and ran out of 250 g by week six.
Typical direction of error
Pipeline forecasts follow deal confidence, and confidence runs ahead of signed orders. A demand plan corrects that by setting the sales view against the statistical baseline and the history. The aim is a number with no consistent lean in either direction over several months.
Update trigger
CRM stage changes move the sales forecast. Mark a batch of deals “likely” in the quarter’s last week and it can jump 15% overnight. The demand plan moves on actual orders, shipments and the planning calendar, which is the rhythm production and purchasing need. Neither can follow a number that jumps on pipeline hygiene.
Coverage
Coverage is the second row that does real damage. A pipeline covers deals someone is working. Repeat orders, small accounts and spare parts often never appear as deals. With base business at half the volume and the plan running on the pipeline alone, operations is guessing at half its demand. A demand plan starts from total history and layers the pipeline on top for the accounts it covers.
How Do You Test Whether the Sales Input Helps?
A narrow forecast value added test answers it. Measure forecast error at the statistical baseline and again after the sales input, over the same items and periods, then compare the two.
Work one example. Over three months, a family of 30 items shows a baseline error of 24%. After sales adjustments, error is 27%. The sales step cost three points of accuracy.
A three-point loss sounds small. On an item family moving a few hundred pallets a month, that can mean an extra week of stock on half the range and shortages on another quarter of it. Operations has usually been complaining about exactly that. Split the adjustments by direction and the picture sharpens. In this example the downward adjustments improved the number and the small upward ones made it worse, which tells planners which conversations to have with sales.
Our guide to forecast value added covers running the same test across every stage of the process.
Where Does Oritiq Fit Between Sales Numbers and the Plan?
Oritiq is the decision layer above the ERP. Its demand forecasting engine takes stock-out periods out of the demand history and estimates the sales lost while the shelf was empty. It then builds a statistical baseline for each SKU from that SKU’s own demand pattern. Promotions are held as separate events, so a trade uplift never leaks into the baseline.
The sales view arrives as one input to consensus building, and every override is logged with a reason category and the name of the person who made it. Forecast value added by stage shows whether the sales step helped, and planners whose overrides keep degrading accuracy get flagged. S&OP reconciliation then carries the agreed number into the supply plan.
FAQs on Sales Forecasting Software vs Demand Planning
1. What is the difference between sales forecasting and demand planning?
They answer different questions. A sales forecast estimates revenue from deals in progress so the team can track a target. Plants need the other number, units by item and location, to size production and stock. A pipeline worth $18 million says little about how many 250 g packs a plant should run.
2. Can a CRM pipeline forecast replace demand planning?
Rarely. A pipeline covers accounts being actively sold to, in revenue, updated when deals change stage. Plants need units for every item, including base business that never enters the CRM. When repeat orders make up half the volume, a pipeline-only plan leaves half of demand unplanned.
3. Why do sales forecasts lean optimistic?
They serve more than one purpose, and a forecast that also works as a target or a motivator drifts above the likely outcome. A field study of over 60,000 forecasts found upward adjustments improved accuracy far less often than downward ones, which fits that pattern.
4. What inputs go into a demand plan besides the sales forecast?
A statistical baseline built from cleaned history, promotional plans, known customer changes such as a new listing, and the sales view. A consensus step then reconciles them into one number by item and location. With 2,000 active items, the baseline does most of the work.
5. How do you test whether the sales input is improving the forecast?
Compare error before and after the sales step. Use 3 to 6 months on one item family. Rising error, say from 24% to 27%, means the sales input is costing accuracy. Then split adjustments by direction and size to see which kind does the damage.
6. How often should a demand plan be reconciled against the sales forecast?
At each S&OP cycle, where sales and operations agree one number. Between cycles, only large changes from sales justify reopening the plan. A lost account or a new listing worth 10% of an item’s volume qualifies, and a hopeful note about next quarter does not.
7. Does demand planning replace sales forecasting software?
No. Sales teams still need pipeline tracking. Deals and targets depend on it. The demand plan uses the sales forecast as one input, weighted by how much it has helped in the past. Both tools stay, and the business stops using one number for two different jobs.
Which Test Should You Run First?
Run the three-month test on one item family. Pull the baseline forecast, the sales-adjusted forecast and actual demand for each month. Where the sales step lowered accuracy, you now know which direction of adjustment to question first. A result like that settles more of the argument than any feature list.
Bring three months of baseline and sales-adjusted forecasts for one item family to a conversation about where your plan’s error comes from.
Contact Oritiq about a scoped pilot.
