The 5 numbers every founder needs to know about their inventory

Most founders can tell you their revenue, their margin, and their month-over-month growth without hesitating.

Ask them their fill rate and you get a pause.

Ask them their forecast accuracy and you get a longer one.

Ask them their reorder point and some of them will ask you what that means.

This isn't a criticism. These numbers don't show up on a P&L. They don't come up in investor meetings. Nobody asks about them until something goes wrong — and by then, the cost of not knowing them is already sitting in a warehouse somewhere, or missing from a retailer's shelf.

Here are the five numbers that tell you whether your inventory is actually working. What each one means, how to calculate it, and what it costs you not to know it.

1. Safety stock

What it is: The buffer inventory you hold specifically to absorb uncertainty — demand that comes in higher than forecast, a supplier delivery that arrives late, or both at once.

How to think about it: Safety stock is not just "a bit of extra inventory." It's a calculated number based on how variable your demand is and how long your lead time is. A product with stable, predictable demand and a short lead time needs a small buffer. A product with volatile demand and a long lead time needs a large one. Treating all your SKUs the same way — holding two weeks of stock across the board regardless of variability — is a common mistake that leaves you over-buffered on some products and dangerously thin on others.

A simple starting point: Multiply your average daily demand by your lead time in days, then add a buffer for variability. If your demand swings by 20 to 30 percent month to month, your safety stock should reflect that. If it's consistent within 10 percent, you can hold less.

What it costs you not to know it: Two things. If your safety stock is too low, you stock out when demand comes in above forecast or a shipment runs late. If it's too high, you're holding cash in inventory that isn't earning anything. Either way, you're paying for not having calculated the right number.

2. Reorder point

What it is: The inventory level at which you need to place a new order — calculated so that the new stock arrives before you run out.

The formula:

Reorder point = (average daily demand × lead time in days) + safety stock

An example: You sell 50 units a day on average. Your lead time is 30 days. Your safety stock is 300 units.

Reorder point = (50 × 30) + 300 = 1,800 units.

When your inventory hits 1,800 units, place the order. By the time it arrives 30 days later, you'll have used roughly 1,500 units and be sitting just above your safety stock level.

Why most brands get this wrong: They use production lead time instead of full supply chain lead time. Production might be 2 weeks. But add transit to your warehouse, transit to the retailer's DC, and time from the DC to the shelf, and your real lead time might be 7 or 8 weeks. A reorder point calculated on 2 weeks of lead time when your real lead time is 8 weeks will leave you out of stock every single time.

What it costs you not to know it: Every unplanned stockout. Every emergency air freight order. Every conversation with a retailer explaining why your fill rate was short. Almost all of these trace back to either not having a reorder point or having one calculated on the wrong lead time.

3. Weeks of stock

What it is: How many weeks of demand you currently have on hand. A simple ratio that tells you at a glance whether your inventory position is healthy, dangerously low, or holding more than you need.

The formula:

Weeks of stock = on-hand inventory ÷ average weekly demand

An example: You have 2,000 units in the warehouse. You sell 400 units a week on average. Weeks of stock = 2,000 ÷ 400 = 5 weeks.

What the number tells you: Whether you have enough — and whether you have too much.

Too low means you're exposed. If your lead time is 8 weeks and you have 5 weeks of stock on hand, you're already past the point where you should have placed an order.

Too high means your cash is tied up in inventory that's sitting. A product with 40 weeks of stock and a 6-week lead time is overstocked by a significant margin — money that could be elsewhere is sitting in a warehouse.

The right number varies by product. A fast-moving core SKU with a long overseas lead time might warrant 12 to 14 weeks. A domestic product with a 2-week lead time might only need 6. The target isn't universal — it's specific to the lead time and demand variability of each SKU.

What it costs you not to know it: You make inventory decisions based on how the warehouse feels rather than what the numbers say. "We have plenty" and "we're running low" are impressions. Weeks of stock is a fact. Without it, overstocking and understocking both happen on instinct rather than on data.

4. Forecast accuracy (MAPE)

What it is: A measurement of how close your demand forecast is to what actually sold. MAPE stands for Mean Absolute Percentage Error — which sounds technical, but the concept is simple: on average, by what percentage was your forecast wrong?

The formula:

MAPE = (absolute difference between forecast and actual ÷ actual) × 100

An example: You forecast 1,000 units. You sold 800. The error is 200 units, which is 25 percent of actual. Your MAPE for that SKU that month is 25 percent.

Do this across all your SKUs for several months and average the results. That's your overall forecast accuracy.

What good looks like:

• Brand stage: Early stage, limited history | Acceptable MAPE: 35–45%
• Brand stage: Growing brand, $5M–$15M | Acceptable MAPE: 25–35%
• Brand stage: Established brand, $15M+ | Acceptable MAPE: 15–25%
• Brand stage: Enterprise / world class | Acceptable MAPE: 10–15%

If you've never measured this before, your first number will probably be higher than you expect. That's fine. The value isn't in the first number — it's in watching it improve as your process gets better.

The other metric to track alongside it: forecast bias. MAPE tells you how far off you are. Bias tells you which direction you're consistently wrong in. A brand with decent MAPE but strong upward bias is systematically over-forecasting — building more inventory than it needs, month after month. A brand with downward bias is systematically under-forecasting — consistently running short. Bias is a process problem. MAPE is an accuracy problem. You need both to understand what's actually happening.

What it costs you not to know it: You have no way of knowing whether your planning is getting better or worse over time. Every inventory surprise feels like a one-off. Some of them are. But if your MAPE is 40 percent and has been for a year, the surprises aren't random — they're structural, and they're fixable.

5. Fill rate

What it is: The percentage of demand you were able to fulfill completely and on time. If a customer or retailer ordered 100 units and you shipped 95, your fill rate for that order is 95 percent.

Why it matters more than it sounds: Fill rate is the number your retail partners are tracking whether you are or not. Most major retailers have minimum fill rate requirements — typically 95 percent or above — and will issue chargebacks for orders that fall short. A sustained fill rate below their threshold can put your account at risk.

Beyond retail compliance, fill rate is the clearest external signal of whether your demand planning is working. A high fill rate means your forecast was accurate enough, your inventory was positioned correctly, and your supply chain executed. A low fill rate means at least one of those things broke down.

The version most brands should track first: Line item fill rate — for each SKU on each order, what percentage of the requested quantity did you ship? This is more useful than order-level fill rate because it shows you which specific products are creating the shortfall, which points directly to where the planning problem is.

**What it costs you not to know it: **Chargebacks you didn't see coming. Retailer relationships that quietly deteriorate. The moment a buyer decides your brand is operationally unreliable is rarely announced — it just shows up in a smaller reorder, or no reorder at all.

How to start tracking these

You don't need new software to know these five numbers. A spreadsheet that you update monthly is enough to start.

For each SKU, track:

• Current on-hand inventory (gives you weeks of stock)
• Lead time in days (gives you your reorder point)
• Forecasted demand vs. actual demand each month (gives you MAPE)
• Units ordered vs. units shipped on each order (gives you fill rate)
• Safety stock level (calculated once, reviewed quarterly)

Set aside an hour once a month to update them. The first time you do it will take longer because you'll be calculating baselines. By month three it will be routine.

The goal isn't to hit a perfect number on any of these. It's to know where you are — so that when something moves, you see it coming rather than discovering it after the fact.

If you'd like help building a tracking system for these metrics — or if the numbers you're running right now are raising questions you're not sure how to answer — we're happy to take a look.