Inventory planning for a small brand: a forecast checklist for 20 SKUs or fewer
Most inventory forecasting advice is written for companies that have a demand planner, an ERP, and a warehouse team. You have a spreadsheet, a co-packer, and a to-do list with forty other things on it.
Here is the good news. If you sell 20 SKUs or fewer, you do not need any of that. A small range is not a handicap, it is an advantage. You can look at every single SKU by hand, understand it, and plan it well. Big brands cannot. They have to trust the model. You get to use judgment.
This is a practical inventory planning checklist for small brands. You can run it in an afternoon and repeat it monthly. I will give you the four numbers you need, the formulas that matter, and a few mistakes that cost real money. I know the mistakes because I made them at my own brand, and one of them is still sitting in a warehouse racing its best-before date.
First, the mistake that pays for this whole article
At Chilli Chan’s, my food brand, we bought close to two years of stock of one product ahead of a big retail launch. The forecast was built on the launch estimate, not on what we were actually selling. The launch came in far smaller than planned. That stock is now sitting in a warehouse, and we are working hard to move it before it expires.
Nothing about that mistake required a fancy tool to avoid. It required one habit: forecast on demand you can see, not on demand you are hoping for. Every item in the checklist below exists to protect you from a version of that mistake.
Before you forecast: get four numbers per SKU
You cannot forecast anything until you have four numbers for each product. Get these first. Everything else is arithmetic.
| Number | What it is | Where to find it | What trips people up |
|---|---|---|---|
| Sales velocity | Units sold per day or per week, split by channel | Shopify analytics, your wholesale orders, retail sell-through | Blending channels into one average hides the truth |
| Lead time | Days from placing the order to stock you can actually sell | Supplier date plus freight plus inbound and QC | Forgetting freight and the time it sits before it is sellable |
| On-hand plus on-order | Stock you hold now, plus stock already in transit | Your inventory sheet and open purchase orders | Counting in-transit stock twice, so you think you have more than you do |
| Shelf life or obsolescence risk | Best-before date, or how fast the product goes out of style | Product spec sheet, or your own judgment on trends | Leaving it out until it becomes a clearance problem |
That last watch-out on the in-transit count is not hypothetical. We once counted an incoming container as on-hand while it was still on the water. It made a month look far healthier than it was. In-transit stock is on-order, not on-hand. Keep the two columns separate.
The checklist
Work through this in order, SKU by SKU. It looks long written out. In practice, once you have done it once, each monthly pass takes minutes per product.
- 1Rank your SKUs by the money they make
Even with 20 SKUs, three or four of them almost certainly earn most of your profit. Sort your list by margin contribution, not by units sold. A cheap high-volume line can earn you less than a slow premium one. Spend your forecasting attention where the money is, and keep the long tail simple.
- 2Match the forecast method to the SKU
Different demand patterns need different maths, and the right one is usually simple. The table below maps each pattern to the method that fits it.
Table 2. Which forecast method to use, by SKU type Demand pattern Example from our range Method to use Why it fits Steady seller Crispy chilli oil, our hero 8 to 12 week moving average Demand is stable, so a plain average is honest and hard to game Trending up A new flavour gaining momentum Weighted moving average, recent weeks count more Catches growth a flat average would lag behind Seasonal Hot honey around the Q4 gifting season Same period last year, adjusted for growth The spike is predictable if you look at last year, not last month Brand new, no history A newly launched SKU Use a similar existing SKU as a proxy, forecast low, review weekly You have no data yet, so stay humble and correct fast Long lead-time import Our knife-cut noodles from Taiwan Forecast further out, order to demand, never to fill a container The lead time forces you to commit months ahead, so accuracy matters most here - 3Work out demand over the lead time
This is the number that decides when you reorder. It is simply how much you expect to sell while you wait for the new stock to arrive.
Lead-time demand = average daily sales x lead time in days
If you sell 12 units a day and your lead time is 21 days, you will sell about 252 units before a new order lands. That is the minimum you need on hand at the moment you reorder.
- 4Set safety stock on purpose, not by feel
Most small brands set safety stock by feel, which means too much on some SKUs and too little on the ones that matter. A simple, honest method:
Safety stock = average daily sales x buffer days
You choose the buffer days by how risky the SKU is. If you want to be more precise, use the peak-versus-average method:
Safety stock = (peak daily sales x longest lead time) - (average daily sales x average lead time)
This sizes the buffer to your actual variability. A steady local product needs a small buffer. A long lead-time import from another continent, paid in a foreign currency, needs a bigger one, because both the demand and the delivery date can move.
- 5Set the reorder point
Now combine the two. The reorder point is the on-hand level that should trigger a new order.
Reorder point = lead-time demand + safety stock
Table 3. Reorder points for three very different SKUs (illustrative numbers) SKU Avg daily sales Lead time Lead-time demand Safety buffer Reorder point Chilli oil (local co-packer) 12 21 days 252 10 days = 120 372 Knife-cut noodles (imported, sea freight) 8 75 days 600 21 days = 168 768 Hot honey (steady, plus a Q4 spike) 5 30 days 150 14 days = 70 220 baseline Look at the noodles row. Same small brand, same size of business, but the reorder point is more than double the chilli oil because the lead time is more than triple. This is the single most important thing long lead-time products teach you: you are committing to a demand forecast for three months from now, not next week. Get that one wrong and you get the warehouse problem I described at the top.
- 6Decide the order quantity against three limits
The reorder point tells you when. The order quantity tells you how much. Aim to order enough to cover demand until your next planned order, then check it against three real-world limits: your supplier minimum order quantity, your cash, and your shelf life. The right order is the demand-based number, adjusted only as far as those limits force you.
This is where discipline pays. A supplier offering a discount for a bigger order is not offering you a deal if the extra units expire on your shelf. A minimum order quantity that gives you two years of cover on a product with an 18 month shelf life is not a minimum, it is a write-off with a delay.
- 7Forecast per channel, then add them up
DTC, wholesale, and retail do not behave the same way. DTC is steady and driven by your marketing. Wholesale comes in lumps whenever a stockist reorders. Retail depends on rotation, which is how fast each store sells through per week. A blended average hides all of this.
Forecast each channel on its own terms, then sum them for the total you need to buy. When a retail buyer gives you an opening order, treat it as a one-off event in its own row, not as a new baseline. Ours felt like a new baseline. It was not.
- 8Plan launches and promotions separately
A launch order, a promo, or a seasonal gift push is a spike, not your normal demand. If you bake a spike into your baseline forecast, you will over-order for months afterward. Keep one-off events in a separate line, order for them specifically, and go back to your baseline once they are done.
- 9Run the cash and expiry gate before you commit
This is the step that would have saved us. Before you send any purchase order, run it past three questions.
Table 4. The pre-order gate Check The question to ask Stop or shrink the order if Cash Can we pay the deposit and the balance without starving marketing and payroll? The order ties up cash you need to actually sell the stock Coverage How many months of cover does this create at current velocity? It is far beyond your normal reorder cycle for no clear reason Expiry Will this sell through before its best-before date at realistic velocity? Projected sell-through is lower than the stock you are about to hold If an order fails any of these, you do not need a better spreadsheet. You need a smaller order.
- 10Set a review cadence and keep it
Forecasting is not a once-a-quarter event. Give it a light weekly glance to catch anything moving fast, and a proper monthly reforecast where you update velocities and reorder points. Ten SKUs, twenty minutes. The brands that get burned are not the ones with the wrong model. They are the ones who set a forecast in January and never looked again.
The mistakes that actually hurt small brands
Most inventory advice warns you about stockouts. Stockouts sting, but for a small brand the more dangerous mistake is the opposite one, because it is quiet and it locks up your cash.
| Mistake | What it looks like | The fix |
|---|---|---|
| Forecasting on hope | Ordering to a launch or pitch estimate instead of real sales | Order to demand you can already see, treat the launch as upside |
| Ordering to the discount | Buying more to unlock a lower unit price | Only if it clears the cash and expiry gate |
| One method for all SKUs | Same buffer and cycle for a local product and an overseas import | Match method and buffer to lead time and variability |
| Ignoring shelf life | Stock expiry is a surprise, not a plan | Put shelf life in the gate before every order |
| Forecasting the total, not the channels | One blended average across DTC, wholesale, retail | Forecast each channel, then sum |
If you avoid only one of these, make it the first one. It is the most tempting, it feels like ambition, and it is the one that put two years of stock in our warehouse.
Where the spreadsheet starts to break
Everything above works fine in a spreadsheet when you have five SKUs and one channel. It gets genuinely painful at fifteen or twenty SKUs across DTC, wholesale, and retail, each with different lead times and different seasonality. The maths does not get harder. The upkeep does. You end up rebuilding the same reorder-point calculations every month and hoping you did not fat-finger a cell.
That gap is why I built OrderBee. It connects to your Shopify data, tracks velocity per SKU, and calculates reorder points and timing for you, so the monthly reforecast is a review instead of a rebuild. To be clear about what it is and is not: OrderBee is forecasting and reorder timing for small Shopify brands. It is not a warehouse system, it is not full inventory management, and it does not do your bookkeeping. It does one job, which is telling you when and how much to reorder, and it tries to do that job honestly.
You do not need it to use this checklist. The checklist stands on its own. But if you are doing this by hand every month and it is eating your evenings, that is exactly the point where a tool earns its keep.
The checklist, in one place
Copy this and run it monthly.
- Pull the four numbers per SKU: velocity by channel, lead time, on-hand plus on-order, shelf life
- Rank SKUs by margin contribution and focus on the top few
- Pick a forecast method that matches each SKU's demand pattern
- Calculate lead-time demand: average daily sales x lead time
- Set safety stock on purpose, sized to variability and supplier reliability
- Set the reorder point: lead-time demand plus safety stock
- Decide order quantity to demand, then check against MOQ, cash, and shelf life
- Forecast each channel separately, then sum
- Keep launches and promos in their own line, not the baseline
- Run the cash and expiry gate before every purchase order
- Review weekly at a glance, reforecast properly once a month
Frequently asked questions
Use a similar existing SKU as a proxy, forecast conservatively, and review weekly. A new product has no data, so your first job is to gather it fast and correct quickly, not to guess precisely on day one.
There is no single number. The reorder point is lead-time demand (average daily sales times lead time) plus a safety buffer sized to how variable your sales and supplier are. A local product needs a small buffer. A long lead-time import needs a larger one.
Enough to cover normal variation without tying up cash you need elsewhere. Start with a buffer of a set number of days of average sales, larger for risky or long lead-time SKUs. If you want precision, use peak daily sales times longest lead time, minus average daily sales times average lead time.
Per channel, then add them up. DTC, wholesale, and retail behave differently, and a blended average hides the pattern you actually need to plan around.
No. A spreadsheet and this checklist are enough at a small range. Software earns its place when the monthly upkeep across many SKUs and channels starts costing you more time than the planning itself.
Written by Seb Hoffmann, founder of OrderBee. OrderBee is a forecasting tool for small Shopify brands that shows you when to reorder, so you stop guessing and stop tying cash up in stock you cannot sell in time.