Demand forecasting
History makes a better forecast.
Weekly and monthly baselines across a 52-week horizon, built from what you sold, not last year plus a percentage.
You have been sold magic before.
Every vendor promises accuracy. You have traded a season where the number broke and nobody could say why. So: what this does, and nothing beyond it.
Demand shows up as a baseline.
The forecast view carries 6 measures for the merchandise scenario in front of you.
- Horizon
- How far forward it goes, out to 52 weeks.
- Average weekly demand
- The weekly rate across the horizon.
- Peak demand
- The highest week in the period.
- Total forecast units
- The full unit count across the horizon.
- Weekly baseline
- Week by week, where intake decisions get made.
- Monthly baseline
- The same, rolled up by month.
FORGE · DEMAND FORECASTING
The forecast meets the week.
BlueIllusion · FY27 · All categories · Illustrative data
Weekly sales · forecast vs actuals
Sales value · AUD · W1–W8
Where the variance sits
W6 · Latest complete week| Category | Forecast | Actuals | Variance |
|---|---|---|---|
| $480k | $525k | +9.4% | |
| $390k | $405k | +3.8% | |
| $290k | $285k | -1.7% | |
| $180k | $195k | +8.3% |
View weekly source values
| Week | Forecast | Actuals |
|---|---|---|
| W1 | $1,120,000 | $1,090,000 |
| W2 | $1,180,000 | $1,210,000 |
| W3 | $1,240,000 | $1,200,000 |
| W4 | $1,190,000 | $1,260,000 |
| W5 | $1,280,000 | $1,320,000 |
| W6 | $1,340,000 | $1,410,000 |
| W7 | $1,390,000 | Not yet available |
| W8 | $1,450,000 | Not yet available |
Reforecasting is normal here.
Trade will disagree with the forecast. Mid-season, that is the expected case.
- HistoryWhat you sold, weekly and monthly.
- ForecastThe baseline, out across the horizon.
- OverrideYour change, where you know something the data does not.
- ReforecastAdjust, compare, restore an earlier version, commit.
Active overrides stay listed, and the commit history keeps what changed.
A forecast you cannot argue with is a forecast that gets ignored.
Simulation comes before the action.
When trade moves away from the forecast, yukti raises a risk with the affected products and the financial exposure attached. Simulate the action, see the estimated effect, then act or dismiss it.
Forecasts feed the WSSI (MFP).
Sales history feeds the forecast; the forecast feeds WSSI, open to buy and continuity, and size curves turn the units into a size mix. All of it is part of core, not a separate license.
Questions, answered.
How accurate is it?
We do not publish an accuracy figure, and you should distrust anybody who does without your data in front of them. Judge it on your own history in the POC.
What is the forecast built from?
Your sales history, weekly and monthly, across 52 weeks.
Can I override a line?
Yes. Active overrides stay listed and the commit history keeps what changed.
Can I reforecast mid-season?
Yes. Adjust, compare versions, restore an earlier one, commit.
Does it forecast by store?
Forecast views are built for merchandise scenarios. Store-level forecasting is UNCONFIRMED.
Your history decides it.
1 hour with sales and product, then a free 6-week POC on your own data.