Amazon PPC for New Sellers: A First-Month Sponsored Products Plan

A new listing has no sales history, no reviews and no data about which shoppers it suits. Advertising can buy that data quickly, but only if you spend with a plan. Without one, new sellers tend to do one of two things: switch on broad campaigns with generous bids and watch the money disappear, or spend so little that nothing is learned. This guide lays out a first month for Sponsored Products only, and the logic behind each step: a bid ceiling, a learning budget, an automatic campaign first, manual campaigns second, and a short list of numbers to ignore until they mean something.

What the first month is for

The first month is for buying information, not for profit. You want to learn which search terms produce clicks and orders, what your real conversion rate is, and what an order costs when you pay for the traffic. Profit comes later, once those numbers are known. To keep the experiment honest, decide on a learning budget: an amount you are willing to spend to find out, with an end date.

Check three things before the first dollar goes out. The listing is active and can win the featured offer, since Sponsored Products shows ads for offers eligible to win it. You have stock for the whole month, because ads that run into a stock-out waste the learning. And you know your profit before ads: what is left per unit after product cost, Amazon's fees and a returns allowance. Ads sent to a thin listing buy clicks that will not convert.

Set a bid ceiling when you do not know your conversion rate

A click is worth what it is likely to earn. Multiply your profit before ads by the conversion rate, meaning orders per click, and you get the most a click can cost before you lose money on average. If a product leaves $6 before ads and 10 clicks in 100 become orders, a click is worth $0.60.

The catch for a new seller is that the conversion rate is a guess. So treat it as a range and see how much the ceiling moves:

The ceiling triples between the low and high cases, which is why it is wise to start below the middle one. A starting bid of $0.50 sits under the $0.60 ceiling at 10% conversion. If your real rate turns out to be 5%, you are paying $0.50 for a click worth $0.30. That gap is the cost of learning, and the learning budget caps it.

Two settings keep the ceiling a ceiling. Choose the dynamic bidding option that only lowers bids, so Amazon can reduce a bid when a click looks less likely to convert but never raises it above your number. And leave placement adjustments at zero for now. Once you have data, replace the guess: your conversion rate is orders divided by clicks, and the ceiling follows from it.

Week 1: one automatic campaign and a learning budget

Automatic campaigns let Amazon match your ad to searches and products based on your listing. They offer four targeting groups: close match, loose match, substitutes and complements, each with its own bid. Start with all four switched on. Close-match searches are the most closely related to your product, so it is reasonable to give them the highest bid and the loose group the lowest, then let the data overrule you.

Set the daily budget from your learning budget. Divide the amount by the days in the month. A $300 learning budget over 30 days is $10 a day. At an average of $0.50 per click, that buys about 20 clicks a day, or 600 for the month.

Run the numbers both ways. At 10% conversion, the 60 orders earn $360 before ads (60 × $6) against $300 of spend, a gain of $60. At 5%, the 30 orders earn $180, so the month costs $120. Which one you get depends on a number you do not know yet, which is exactly why the budget is capped. The test also leaves out the organic sales that ad-driven orders can help build, so a small loss is not automatically a bad month.

Weeks 2 and 3: read the search terms, then build manual campaigns

After a few hundred clicks, open the search term report. It lists the shopper searches and product targets that triggered your ads, with clicks, spend and orders. Sort by orders. Terms that produced orders are candidates for a manual campaign, usually with exact match so you control the bid for that term.

Do not promote a term on one lucky order. One order from three clicks looks like 33% conversion, and it is noise. A simple rule is to wait for at least two orders, then work out the term's own ceiling from orders divided by clicks, shaded down because the sample is small. Add each promoted term as a negative exact in the automatic campaign, so the two campaigns do not compete for the same search and you keep control of the bid.

For the terms with clicks and no orders, ask how many clicks a typical order takes. It is roughly 1 divided by the conversion rate. At 10%, that is about 10 clicks per order, so judge a term only after two or three times that, say 20 to 30 clicks. At $0.50 a click, that is $10 to $15 per term before you add it as a negative. Cut sooner and you may drop a term that was about to convert. Cut later and you pay for nothing.

What to ignore in the first month

Early numbers are noisy, and reacting to noise burns the learning budget. Hold off on these:

• ACoS in the first week. With a handful of orders, ACoS can swing from nothing to several hundred percent on a single sale.
• Impressions and click-through rate on their own. They show whether your ad is seen and clicked, not whether it sells.
• The most recent days of sales. Orders are credited to an ad within an attribution window after the click, so the latest days always look worse than they will.
• Daily rank movements for a single keyword, which shift for many reasons.
• A competitor's price or bid seen on one day.

Expect low volume, a few quiet days and conversion that starts below where it will settle, since a new listing has no reviews yet. The aim of this month is a clear set of numbers: your real conversion rate, your best-converting terms and a cost per order you can judge against your profit.

Where to start

• Confirm the listing is active, in stock and able to win the featured offer.
• Work out profit before ads, then a bid ceiling at a low, middle and high conversion rate.
• Set a learning budget with an end date, and turn it into a daily budget.
• Launch one automatic campaign with all four targeting groups and bids under your middle ceiling.
• Review the search term report after a few hundred clicks. Promote terms with at least two orders and add negatives only after enough clicks.
• Replace the guessed conversion rate with your own data and plan month two from it.