Lesson 7 of 10 · 4 min read
Your first campaigns
A starter structure: an auto campaign to find searches, a manual campaign to test them, the settings to pick on day one, starting bids, budgets, names, and what to expect in the first month.
By Cameron Scot, founder of Merch Jar · Updated
Short answer
- First check the product is in stock and winning the Featured Offer. If it isn't, the ad won't show.
- Run a Sponsored Products auto campaign to find searches and a manual campaign to test them as exact keywords.
- Start bids near Amazon's suggested bid, but no higher than what a click is worth to you (lesson 5).
- When you promote a search term, keep it running where it came from. Promoting and negating are separate decisions.
Start simple and make it more complicated as you go. You want a structure small enough that you’ll actually check it every week.
Before you start
Ads only show for a product that’s in stock and winning the Featured Offer (the Buy Box). If someone else holds it or you’re out of stock, your ad doesn’t run, no matter the bid. Ads also send traffic to the listing, so get the title, images and price where you want them first.
Discovery, then testing
An auto campaign finds the searches. Amazon picks where to show your ad based on your listing, across searches and product pages. This is where your listing’s indexing gets put to work. Because it shows up in so many places, keep it efficient by negating searches that collect about 20 clicks without an order, and searches that obviously have nothing to do with your product as soon as you see them.
A manual campaign tests what the auto campaign found. When a search term in the auto campaign gets a couple of orders, add it to the manual campaign as an exact keyword. When a product page (an ASIN) gets orders, add it as a product target. Now that search has its own bid, and you can push it or pull it back without touching everything else. You can copy them over by hand, or use Target Promotion in the ads console to move them for you.
Don’t wait for the auto campaign to fill it, though. Seed the manual campaign on day one with a handful of exact keywords you already know fit, like the main words in your title and the searches you’d type yourself.
Keep keywords and product targets in separate ad groups in the manual campaign. Amazon requires it, and it keeps the data clean anyway.
The settings to pick on day one
| Ad type | Sponsored Products (SP). It’s where most sellers start, and it’s what this course covers |
| Campaigns | One automatic, one manual, per product |
| Bidding strategy | Dynamic bids, down only (Amazon’s default) or fixed bids. Leave up and down until you have data |
| Placement adjustments | 0% everywhere to start (lesson 4) |
| Auto targeting | Amazon splits auto campaigns into four target groups (close match, loose match, substitutes, complements). Leave all four on at the same bid to start |
| Daily budget | Enough for 15 to 20 clicks a day at your starting bid, so a week gives you something to read |
Don’t cut off the source
A lot of guides say that once you promote a search term to an exact keyword, you should negate it in the auto campaign so the two don’t compete. I don’t. If the search is still selling in the auto campaign, it’s working there, and negating it throws away sales to tidy up a report.
Only negate it in the auto campaign if it’s selling there above your target ACOS while the exact keyword does better. Promoting and negating are two separate decisions with two separate rules.
One product per campaign to start
Start each product with its own campaigns, so its results are its own. Close variations of the same product can share, if you don’t need them split apart.
Starting bids
Your first bid matters less than you think. It’s your way into the auction so you can start collecting data. Higher bids collect data faster and cost more. Lower bids are cheaper and slower. Amazon’s suggested bid is a fair read on what the competition is paying, so start near it. If it’s more than a click is worth to you (lesson 5), start at what a click is worth instead and let the data move it.
Budgets are guardrails
A daily budget is there to stop a runaway day. If a campaign hits its budget every day, it’s missing traffic while you’re limiting it, so control ACOS with bids and negatives and leave the budget with some room.
If money is tight, test fewer things at once (fewer new campaigns, keywords or products you don’t know work yet) instead of starving all of them.
Name campaigns so you can read them later
A campaign’s name should tell you what’s in it and why, at a glance, a year from now. Include the product (ASIN or SKU), the ad type (SP for Sponsored Products) and targeting, the goal, and a short product name. Beyond SP, spell things out rather than using acronyms you’ll forget.
B0EXAMPLE1 - SP - Auto - Discovery - Insulated Water Bottle 32oz
B0EXAMPLE1 - SP - Manual Exact - Testing - Insulated Water Bottle 32oz
Your first month
For the first couple of weeks, most targets won’t have enough data for any rule in lesson 8. That’s normal. Look in twice during the first week for the obvious problems: no impressions at all, a budget that runs out by midday, searches that have nothing to do with your product (negate those right away).
Then leave bids alone until targets start reaching the numbers in lesson 6, usually somewhere in weeks three and four. Negating and promoting search terms tend to come first. Bid changes come once targets have a couple of orders.
Go deeper
Run this rule every day
This is a template in the Merch Jar Library. You preview exactly what it would change, it's saved switched off, and you turn it on when you're happy with it.
