Planning a Seasonal Amazon Product: Demand, Lead Time and Cash

A seasonal product earns most of its year in a few weeks, and the mistakes made around those weeks are paid for all year: too little stock when demand arrives, too much when it leaves, cash tied up in cartons for months. The answer is to plan the product's whole year and not just its season. That means knowing the shape of demand, working back from the peak to the date you must order, measuring how much cash you are putting at risk, spreading your ad budget along the curve, and deciding in advance how to clear what does not sell.

Draw the demand curve before you order anything

Seasonal demand has a ramp, a peak and a drop, and the ramp starts earlier than most sellers expect, because many shoppers buy ahead of the occasion. Build your own curve from weekly sales in your reports for last season. If you have Brand Registry, add the search data in Brand Analytics for your main terms across several years. A free trend tool can supplement both. Look for three dates: when interest starts to climb, when it peaks and when it falls away.

The curve is usually steeper than a calendar suggests. Picture a hypothetical item that sells 50 units a week twelve weeks before its peak, 150 at eight weeks, 600 at four weeks, 1,000 in the peak week, 200 two weeks after and 20 six weeks after. Most of the season's sales sit in a narrow band, so a stockout or a late shipment in that band costs far more than the same problem anywhere else.

Turn the curve into a forecast with a range, not a point. Use last season's units as a base, then write a low, a base and a high case, and note what would push you toward each, such as a competitor entering or a product improvement.

Work back from the peak to the order date

Seasonal products usually cannot be reordered once the season starts, so your first order is the season. That makes lead time the central number. Add up production, transit, customs if the goods are imported, the time Amazon takes to receive stock, and a buffer for delay. Then subtract the total from the date you want stock ready to sell, which should be before the ramp begins, not on peak week.

In a hypothetical plan you might decide volume about nine months before the peak, place the production order at six months, ship at three to four months and have stock checked in well before the eight-week mark. Your own lead times will differ, so find them from your last purchase order, not from a template. Also check your restock and storage limits early. They can cap how much you are allowed to send ahead of a peak, and that limit may be tighter than the volume you planned.

Cash: money goes out months before it returns

With a seasonal product you pay for production, freight and duties long before the first sale. Sales proceeds then arrive on Amazon's payout cycle. The gap can be several months, so count your maximum exposure before you commit. Multiply the units you plan to order by landed cost. In a hypothetical case, 8,000 units at $6 landed is $48,000 out of your account.

Then test the downside. If only 5,000 of those 8,000 units sell, 3,000 remain, which is $18,000 of cash sitting in storage, and storage is not free. Amazon has historically charged higher monthly storage fees in the last quarter of the year, and aged inventory charges apply to units that stay long, so check the current fee schedule and your inventory limits. A sensible order is one you can survive selling at the low case. If you want the high case, ask your supplier whether a smaller second order is possible, or whether part of the stock can wait at a US warehouse and move to Amazon only as it sells.

Pace ads with the curve, not the calendar

A flat ad budget wastes money in the quiet weeks and starves the loud ones. Tie the weekly budget to the expected sales in that week. A simple way is to take expected sales dollars and multiply by the TACoS you plan to spend. For a hypothetical $20 item and a 10% plan, the peak week of 1,000 units is $20,000 in sales and a $2,000 budget. Four weeks earlier, 600 units is $12,000 and $1,200. Twelve weeks before the peak, 50 units is $1,000 and $100.

Treat that line as a starting point and adjust it. You may want to spend a little more than the line in the early weeks to build rank and reviews before the surge, with a cap and an end date. Click costs often rise as competitors pile in near the peak, so watch your cost per click and your stock. A campaign that is working when only a few days of stock remain should be turned down, not up, because ad spend that ends in a stockout is wasted.

Clearing leftovers and the off-season

Decide your clearance plan before the season starts, because choosing under pressure usually means choosing late. Set a date and a sell-through level that trigger it. After that, follow a ladder that moves from the gentlest option to the firmest.

Carrying stock to next season is a calculation, not a feeling. Suppose storage costs $0.25 per unit per month and you would hold the units for nine months, which is $2.25 per unit. If a liquidation offer pays $3.00 a unit, carrying wins only if you expect to net more than $5.25 a unit next season, after selling costs. Remember too that some products date, so a design that looks old next year carries a risk the arithmetic does not show.

Use the off-season for the review. Record sell-through, stockouts, lead times and ad results while they are fresh, and put them next to your forecast. The gap between the two is the single most useful number for next year's plan.

Where to start

• Build a weekly demand curve from last season's sales and any available search data.
• Write down your real lead time from your last purchase order, and work back from the date you want stock ready.
• Calculate cash at risk for a low, a base and a high case before you place the order.
• Check storage fees, aged inventory charges and restock limits for the months stock will sit.
• Set weekly ad budgets from expected sales and a planned TACoS, and cap early spending.
• Agree the clearance trigger date and the ladder before the season begins.