Is Your Amazon Product Seasonal? Run the 15% Test | Inventory Hero
·7 min readForecasting
Is Your Amazon Product Seasonal? Run the 15% Test
A flat business does 8.3% of annual sales per month. If one month clears 15%, your product is seasonal and your reorder math changes. Here is the test.
In Seller Central Business Reports, export units ordered by month for the last 12 full months, or 24 if available.
2
Clean the data
Exclude or normalize months distorted by stockout days, and strip one-off deal spikes like Prime Day or Black Friday if they do not repeat for this SKU.
3
Compute each month's share
Divide each month's units by the twelve-month total to get its share of annual sales.
4
Compare against 8.3% and 15%
A flat product does about 8.3% per month. Any month at 15% or more means the SKU is seasonal: forecast it on the forward curve with a seasonal index instead of a trailing average.
Frequently Asked Questions
How do I know if my Amazon product is seasonal?
Divide each month's unit sales by your trailing twelve-month total. A flat product puts roughly 8.3% of annual sales in every month. If any month carries 15% or more, roughly double a flat month, treat the product as seasonal and plan inventory on the forward curve instead of the trailing average.
What percentage of sales makes a product seasonal?
T. Brian Jones is co-founder and CTO of Inventory Hero. He leads the engineering behind its Amazon data pipeline, demand forecasting, and the AI platform that lets sellers talk to their live inventory, sales, and supplier data in plain language.
There is no official threshold, but 15% of annual sales in a single month is a practical operator's line. It is nearly double the 8.3% a flat business does per month, which is a big enough swing that trailing-average reorder math will understock the peak and overstock the trough.
How do I forecast a seasonal product without a year of sales history?
Triangulate three signals: whatever history you have, cleaned for stockouts and deals; the seasonality curve of the category and your closest competitors; and keyword search volume trends, which typically move weeks ahead of sales. A new product inherits its category's curve until its own history proves otherwise.
Does Q4 make every Amazon product seasonal?
No. Most products get some Q4 lift, but a broad November and December bump that stays under about 15% of annual sales per month can be handled with a modest buffer. Reserve full seasonal planning, with seasonal indexes applied to velocity, for SKUs where the test shows a true concentrated peak.
Is your Amazon product seasonal? Here is the test we use: a perfectly flat business does about 8.3% of its annual sales each month (100% divided by 12). If any single month carries 15% or more of your trailing twelve-month unit sales, that month runs close to double a flat month (1.8x the baseline), and the product should be planned as seasonal. One threshold, one report, and a clear fork in how you run your reorder math. For a SKU with clean history the test takes five minutes; a SKU with messy stockout months takes longer, because you have to reconstruct what it would have sold first.
This matters because the two planning modes are genuinely different. Evergreen SKUs can run on a trailing average. Seasonal SKUs planned that way stock out right before their peak and sit on excess right after it.
Pull twelve months of monthly unit sales and check each month's share of the annual total. Concretely:
In Seller Central, open Business Reports and pull units ordered by month for the last 12 full months (24 if you have them).
Clean the data: exclude or normalize months distorted by stockout days and strip obvious deal spikes like Prime Day and Black Friday week if they do not repeat for this SKU. For stockout months, estimate what you would have sold using the method in how to calculate lost sales; a month you spent half out of stock is a supply failure, not a demand signal.
Divide each month's units by the twelve-month total.
Compare each month against the two lines: 8.3% is flat, 15% is seasonal.
Worked example. A SKU sells 12,000 units in a year:
Month
Units
Share of annual
March
950
7.9%
July
1,010
8.4%
November
2,410
20.1%
December
1,890
15.8%
November carries 20.1% of the year and December 15.8%, both clearing the 15% line. This is a seasonal product, and its November velocity will run roughly 2.4x its flat-month baseline (2,410 units vs the ~1,000-unit flat month). No month over ~12%? You have an evergreen SKU with normal noise, and a standard reorder point with honest safety stock covers you.
To be clear about what the 15% line is: an operator's rule of thumb, not a statistical standard. Real "flat" SKUs are never perfectly flat; weekday mix, month lengths, and ordinary noise bounce them between roughly 7% and 11% month to month. The 15% line works because it sits far enough above that band that noise rarely crosses it, but low enough to catch real peaks while there is still time to plan for them.
The judgment calls live in the 12% to 15% band. A SKU with one month at 13% is not an emergency, but it is not flat either: give it a modest pre-peak buffer (an extra 2 to 3 weeks of cover into its strong month) without moving to full seasonal-inventory planning. If the same month clears 13% two years running, trust the pattern and treat it as seasonal.
Because a trailing average is a report on the past, pointed backward, at exactly the moment you need a forecast pointed forward. Days of cover is a forecast of the future, not a report on the past.
Take the SKU above in early October. Its trailing 30-day velocity is about 33 units/day (a flat September). Its actual November demand will run about 80 units/day. A reorder plan built on 33 units/day orders less than half the inventory November needs, and the resulting stockout lands in the most expensive week of the year to be absent. The same error runs in reverse in January, when the trailing average still reflects December and quietly orders you a pile of excess. This is the single most common seasonal planning failure we see, and it is invisible in the spreadsheet because the velocity number looks precise.
The fix is one multiplication: forecast_velocity = baseline_velocity x seasonal_index, where the index is the target month's share divided by a flat month's share. For our November SKU: 20.1 / 8.3 = an index of about 2.4. The full index math, including quarterly and weekly variants, is in our Amazon sales seasonality guide.
Triangulate three signals instead of trusting any one of them. This is also the fix when your history exists but is dirty:
Your history, cleaned. Same period last year, with stockout days and one-off deal spikes removed, using the same lost-sales reconstruction as step 2 of the test.
Competitors and the category. Your niche's curve tells you the true peak, not just the part of it you captured. Pull it from competitor sales-estimate history in tools like Keepa or Jungle Scout, or eyeball the BSR history of your closest three comps across last Q4. A new product inherits its category's seasonality until its own data proves otherwise, which is the working assumption behind new product demand forecasting.
Keyword demand. Search volume for your main keywords moves weeks ahead of sales. Rising search with flat sales usually means the wave is coming, and it is the earliest seasonal signal you can act on.
When the three disagree, weight the category curve for shape (when the peak lands) and your cleaned history for amplitude (how high it goes), and let keyword trends confirm timing.
Peak demand through the season, sized to sell through by season's end
For Q4 specifically, the baseline window matters as much as the index: measure baseline velocity on a flat stretch, March through June works for most SKUs, then apply your November and December indexes to that. The complete walkthrough, with the calendar attached, is in how to forecast Q4 inventory, and you can pressure-test the resulting order in the restock calculator.
The 15% test is the fastest honest answer to "is my Amazon product seasonal": any month carrying 15% or more of annual sales, against the 8.3% flat baseline, means you plan that SKU on the forward curve. Run it on your top ten SKUs this week. The ones that fail the test just told you, months in advance, exactly when their reorder math stops being simple.