A weekly figure for how much to order, with its own margin of error
Demand Forecasting Setup
Back to homeThe promise
Over five weeks, we configure a forecasting step for a defined range of your products, using your own sales history alongside the calendar effects that shape demand in Japan — long holiday periods, seasonal changeovers, and local event patterns. Each week you receive a figure with a stated margin of error, delivered to your existing spreadsheet, so ordering decisions rest on something more specific than habit alone.
The problem
Ordering by feel works reasonably well until a holiday shifts a fortnight earlier than expected, or a new competitor opens nearby and last year's numbers stop applying cleanly. Most buyers already sense when their usual method is straining — more urgent reorders, more markdowns on stock that didn't move. It's rarely one big miscalculation; it's a slow accumulation of small ones.
The solution
We configure a forecasting model against your transaction history, tuned specifically to calendar effects relevant to Japanese retail. Rather than a single projected number, you receive a forecast paired with its confidence range, so you can see at a glance how much weight to give it for any particular week or line.
The experience
We begin by reviewing the sales history and current range together, then configure the model over the first week or so. From there, output arrives weekly in your existing spreadsheet — no new software to learn. At the end of the first full month, we sit down to review how the forecasts performed against actual sales and adjust where needed.
The investment
¥29,000
Five-week scope, for retailers with at least two years of transaction history across twenty or more lines.
- · Model configuration for your defined range
- · Weekly output delivered to your existing spreadsheet
- · A review session after the first full month
The proof
The method relies on established forecasting techniques rather than anything unproven, applied to your own data instead of a generic industry template. Progress is measured by comparing forecast against actual sales each week, which is part of why the month-one review exists — to check the numbers against reality before drawing conclusions.
The guarantee
If, after the first two weeks of output, the forecasts clearly don't fit your range, we'll discuss adjusting the scope or stopping the engagement rather than continuing regardless. An initial conversation about your data carries no obligation to proceed.
Next steps
Send a short note about your product range and how long you've been tracking sales. We'll let you know whether this service is a good fit, and if so, what we'd need from you to begin.