Lottery Campaigns and the ASIN Bucket System

A large Merch catalog cannot be tested one product at a time.

By this point, the seller has prepared the catalog, understood royalty math, learned why PPC must filter before it scales, and defined the role each campaign should play. The next step is to apply that structure to a real large-catalog problem: how to test many products without giving every ASIN its own expensive campaign.

This is where many Merch sellers get stuck. They may have hundreds, thousands, or tens of thousands of live products. Some are new. Some are old. Some are seasonal. Some have organic signals. Some have never received meaningful exposure. Some are probably weak. Some may be promising, but they are hidden inside a catalog too large to judge manually.

If the seller tries to build dedicated campaigns for every product, the account becomes impossible to manage. There are too many campaigns, too many bids, too many budgets, and too little data per product.

If the seller throws every ASIN into one giant automatic campaign, the opposite problem appears. The structure becomes too broad, the data becomes hard to read, and a few products may consume most of the exposure while many others receive no real chance.

The solution is not more chaos. The solution is a controlled discovery engine.

A Lottery campaign is not built to scale every product. It is built to reveal which products deserve the next decision.

In practice, the default Lottery loop is short:

1. Group ASINs that can be judged through the same economics, timing, and buyer context.
2. Buy low-risk discovery with a controlled Sponsored Products Automatic campaign.
3. Read the results at the ASIN level, not only from the campaign average.
4. Sort each product into the next decision: winner, promising/watchlist, underexposed/second chance, weak or no-budget, or seasonal retest later.
5. Move only the active buckets forward. The whole catalog does not stay in one weekly PPC queue.

A Lottery campaign is a low-risk discovery structure designed for a large Merch catalog. It is usually built with Sponsored Products, Automatic targeting, low bids, controlled budgets, and a defined group of ASINs. Its job is not to force sales across the entire group. Its job is to expose products to enough traffic that the seller can begin classifying them.

That classification is the real value. A Lottery campaign may reveal products that receive orders, products that receive clicks but no orders, products that receive impressions but no clicks, products that spend quietly, products that receive no exposure at all, and products whose performance only makes sense inside a seasonal window. Each of those outcomes means something different.

A product-level result needs a product-level decision:

• A product with repeated orders may deserve graduation.
• A product with one early order may be promising but not yet proven.
• A product with no impressions has not failed; it may simply be underexposed.
• A product with relevant clicks and no orders may have an offer problem.
• A seasonal product tested in the wrong month may need to wait for its real buying window.
• A weak product that already had a fair chance may need to stop receiving budget.

The campaign is called "Lottery" because the seller is not pretending to know every winner in advance. The catalog contains uncertainty. Some products will respond, many will not, and some will not receive enough exposure on the first pass. The point is not to guess perfectly before spending. The point is to spend carefully enough that the market can reveal useful signals without damaging the account.

This makes Lottery campaigns different from scaling campaigns. A scaling campaign is built around stronger evidence. It exists to increase controlled volume on a product, target, or signal that already has proof. A Lottery campaign comes earlier. It is built to find the signals that may later justify cleaner structures.

That difference changes how the seller should judge the campaign. A Lottery campaign does not fail because every ASIN does not sell. That was never the job. The better question is whether the campaign produced useful product-level evidence at controlled cost:

• Did any products show demand?
• Did any products receive exposure but fail to convert?
• Did some products receive no impressions?
• Did seasonal products need to be separated?
• Did Amazon concentrate traffic too heavily on a few ASINs?
• Did the campaign reveal enough to make the next decision?

The rest of the chapter explains how to keep that loop controlled: what a Lottery campaign is, how to set up a low-risk Automatic test, how to group ASINs so the data stays readable, and how the bucket system turns campaign noise into next actions.

The chapter does not yet go deep into harvesting, negatives, or full winner graduation. Those decisions come next. Before the seller can harvest search terms or graduate products safely, they need to know what the Lottery campaign actually revealed at the ASIN level.

The goal of this chapter is therefore simple: turn a large catalog from a pile of untested products into a set of product-level decisions. The seller should finish the chapter knowing which products deserve more attention, which products need a cleaner test, which products were never truly exposed, which products should wait for seasonality, and which products should stop receiving paid support.

A Lottery campaign does not remove uncertainty. It organizes it. That is what makes it useful.