Learning phase — why nothing happens in week 1

You launch a new campaign and check the data after 24 hours — few impressions, a couple of clicks, no sale. Is the campaign broken? Short answer: no. Amazon needs time to learn.

What happens in the first 7 days

When you activate a new campaign, Amazon starts a learning process. The algorithm tests who should see your ad — and collects conversion data.

• Days 1-3:Low impressions, unstable delivery. Amazon doesn't know which shoppers will click your ad, so it's cautious in the auction.
• Days 4-7: Delivery becomes more stable. First sales (if your product converts). Amazon now prioritizes the shopper profiles that clicked and bought in the early days.
• Days 8-14: Steady state. The campaign delivers with consistent performance. Now the data is meaningful.

Why you shouldn't change anything in this window

Every change — bid up, bid down, change match type, add keyword — partially resets the learning process. Amazon has to collect new data and re-calibrate.

If you raise the bid on day 3 because you're impatient, by day 7 you don't have 7 days of data — you have 4 days of clean data. With 4 days of data you can't decide whether a keyword really converts.

What you can do in the first 7 days

Three actions are safe:

• Watch without changing. Note daily impressions and clicks, nothing more.
• If 0 impressions after 72h: delivery may be blocked. Walk through the Zero impressions diagnosis. A bid increase is ok if the bid is obviously too low.
• If your ad shows for clearly wrong search terms: adding a negative keyword is ok. It interrupts the learning process less than a bid change.

After day 7-14: now you can adjust

Once you have 14 days of data, it's meaningful enough for decisions. Rule of thumb:

• At least 30 clicks per keywordbefore you judge it. With 10 clicks and 0 sales, it's still too early to pause.
• At least 14 days of data before changing the bid significantly. With seasonal shifts (e.g. Black Friday approaching) shorter intervals are fine.

More on adjustment rhythm: First bid — how high?.

Three common beginner reactions during the learning phase

"No sales yet, I'm pausing everything."

Patience. For a new product with under 20 reviews, it can take 14 days for the first ad-driven order. Only pause if after 14 days + 30 clicks per keyword no sales have come in.

"I'll raise the bid 50% to make something happen."

Big bid jumps during the learning phase waste budget. If your bid truly was too low, a 20-30% increase on day 4 is enough — then wait again.

"I'll launch two campaigns with the same strategy to compare."

A/B tests on Amazon ads are tricky. Both campaigns compete against each other in the auction, which skews the result. Run one campaign for 14 days, then start variant 2 based on the data.

What Sellerwerk shows you here

In the campaign detail view you see the learning-phase status: whether the campaign is still in its first 7 days and which actions wouldn't be meaningful yet. Sellerwerk does not suggest automated actions during the learning phase — we wait for valid data before firing recommendations.

Automated rules (see Automation basics) should also have a minimum click threshold of 20-30 clicks so they don't fire too early during the learning phase.

One exception: critical campaign fires

If you're burning $100+ in the first 3 days with zero sales, the learning phase doesn't matter — pausing is the right call. Common cause in such cases: too-broad match types (broad without negatives) or too-high opening bid.

More on auction mechanics: What is Amazon PPC?. For questions: support@sellerwerk.de.