If you’ve searched for inventory forecasting advice, you’ve probably seen the same formula repeated everywhere. It’s simple, easy to understand and easy to calculate. In many businesses, it’s also wrong.
Reorder point = average daily sales × lead time + safety stock
Let’s look at why — and at what to use instead.
Problem 01Average sales hide what’s actually happening
The biggest flaw is right there in the name. Suppose your last six months looked like this:
| Month | Daily sales |
|---|---|
| January | 18 |
| February | 22 |
| March | 30 |
| April | 42 |
| May | 58 |
| June | 76 |
- 41/dayThe average
- 76/dayToday
If you reorder based on the average, you’re planning for a business that no longer exists. Growing brands should forecast where demand is heading — not where it used to be.
A better question is: “What am I likely to sell during my supplier’s lead time?” That’s very different from asking what you’ve sold historically.
Alternative 01The forward-looking sales rate
Instead of averaging months together, weight recent performance more heavily.
Forecast sales = (last 7 days × 60%) + (last 30 days × 40%)
- 60/dayLast 7 days
- 50/dayLast 30 days
- 56/dayForecast
It better reflects businesses that are growing — and it reacts much faster when sales begin slowing.
Problem 02Lead time isn’t one number
Most spreadsheets ask for one value. Lead time: 45 days. Done. Except — what happens when your supplier delivers in:
- 38 days
- 43 days
- 47 days
- 58 days
- 63 days
Which one is your lead time? They’re all correct.
Your supply chain doesn’t have one lead time. It has a range.
Alternative 02Plan around the worst reasonable case
Instead of using your average lead time, use the lead time you can actually depend on.
Average
43 days What you’d normally plan aroundLongest observed
61 days What actually happenedInstead of forecasting for 43, forecast for 55–60. Yes, you’ll carry slightly more inventory — but you’ll dramatically reduce stockout risk.
Problem 03Safety stock is usually guesswork
Ask ten sellers how they calculate safety stock. Many will answer: “I keep about two weeks.”
Why? “Because it feels right.”
That’s not forecasting. That’s intuition.
Alternative 03Safety stock based on lead-time variability
Instead of choosing an arbitrary number of days, calculate how inconsistent your suppliers actually are. Say your last eight deliveries took:
- 42
- 44
- 43
- 41
- 58
- 46
- 61
- 42
- 47Average
- 61Maximum
- 14Days of difference
- 700Safety stock at 50/day
Now your safety stock reflects real-world supplier performance — not a guess.
Problem 04The formula doesn’t know about growth
Imagine you’re launching PPC. Running promotions. Improving reviews. Increasing conversion rate. Sales are expected to grow 20%.
The classic formula assumes none of that happens. It plans for a static business — and Amazon businesses are rarely static.
Alternative 04Demand scenarios
Instead of one forecast, create three.
- 40/dayConservative
- 50/dayExpected
- 65/dayAggressive
Calculate inventory requirements for all three. Instead of one answer you have a range — much closer to how experienced operators think.
Problem 05It ignores cash flow
Imagine the formula recommends ordering 7,000 units. Perfect. Except you only have enough cash for 4,500.
Now what? The spreadsheet doesn’t know.
Inventory planning without cash planning is only solving half the problem.
Alternative 05Cash-constrained forecasting
Instead of asking “how much inventory should I buy?”, ask:
What’s the maximum inventory I can safely finance without putting the business under pressure?
Sometimes ordering slightly less — and ordering more frequently — is the better decision. Especially for businesses growing quickly.
Problem 06It treats every product the same
Should you manage a $9 phone case the same way you manage a $120 standing desk? Of course not. High-value inventory carries a much higher opportunity cost.
Alternative 06Inventory value coverage
Instead of only measuring units, measure money. How many dollars are currently sitting in inventory?
Product A
$6,000 1,000 units × $6 landed costProduct B
$42,000 1,000 units × $42 landed costBoth have identical inventory levels. Only one is tying up serious capital. Sometimes inventory dollars matter more than inventory units.
Problem 07It assumes forecasting is a calculation
This is probably the biggest misconception. Forecasting isn’t a formula — it’s a process. The formula only gives you one number.
Good operators continuously ask questions:
- What changed this week?
- Did PPC improve?
- Did competitors go out of stock?
- Did my supplier delay production?
- Has demand accelerated?
The quality of your questions matters more than the complexity of your spreadsheet.
A better mental model
Instead of using one equation, think of forecasting as balancing four moving pieces:
- Demand
- Lead time
- Cash
- Risk
Every decision changes one of them. Forecasting isn’t about finding a perfect answer — it’s about understanding the trade-offs.
The formula I’d rather use
Instead of this:
Reorder point = average daily sales × lead time + safety stock
I prefer thinking like this:
Inventory required = expected demand during lead time + protection against uncertainty − inventory already on the way
There isn’t a single “average”. There isn’t one fixed lead time. And there isn’t an arbitrary safety stock. Everything is tied to how your business is performing today.
Final thoughts
The classic reorder point formula has survived for decades because it’s simple. And simplicity has value. But simplicity can also create a false sense of precision.
Amazon businesses operate in an environment where demand changes quickly, suppliers miss deadlines, promotions distort sales, and cash is almost always constrained.
The best inventory planners don’t rely on a single equation. They combine mathematics with judgment. They question assumptions instead of accepting averages.
Forecasting isn’t about producing a perfect number — it’s about consistently making better decisions than competitors who are still planning for yesterday’s business.