Impressions
What it means
Impressions is the number of times your Amazon ads were served — on a search results page, a competitor's listing, anywhere Amazon placed them. It counts serves, not people: the same shopper scrolling past three times counts three. Nothing was tapped and nothing was charged, so it measures reach and nothing else.
Amazon, Google Ads and Meta each count only their own surfaces, under their own rules. The three numbers on this page share a label and nothing else. Never add them together, and never read one platform's total as better or worse than another's.
Show the math
Formula and a worked example
Total times ads were shown counts each serve on Amazon, across every campaign type and placement. Amazon publishes no share-of-available figure here, so this count has no measurable ceiling — the only reference point is your own history.
Worked example. Your ads are served 500,000 times and earn 5,000 clicks — a click-through rate of 1%. At €0.60 a click that's €3,000 of spend, and the other 495,000 serves cost you nothing.
Free is the point and the trap. Reach costs nothing directly, but it's bought with bids and budget, and if it reaches the wrong people, every ratio underneath sags.
It answers the question
Are enough shoppers seeing your ads? A flat sales month with Impressions down is a distribution problem, not a creative one.
Why it matters
Impressions are the footfall past your window — serves, not people. You need them before anything else can happen, and no ratio underneath rescues an empty street.
That makes this the first place to look when things go quiet. If Impressions fell before clicks and orders did, the cause sits above the ad: a budget that ran out, a bid that stopped clearing the auction, or a listing that lost the Buy Box and quietly stopped serving.
What good looks like
There's no benchmark worth quoting — the right figure depends on your category's search volume, your budget and how many SKUs you advertise. Judge it against your own trend and the CTRCTRHow well your ad grabs attention in the feed.0.5% or more is healthy it produces.
Reach that grows while clicks stay flat was bought from the wrong audience; a steady Impressions line with rising clicks usually means relevance improved, not spend. The Google Ads tab has a published ceiling to measure against and this one does not, which is exactly why the two counts can't be scored on the same scale.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Lift caps on campaigns that run out | Raise daily budgets where campaigns stop serving before the day ends | Impressions up the same day | 1–3 days | Spend rises immediately, and the hours you buy back are the least efficient of the day — reach improves while the cost of an order worsens. |
| Fast Bid back into top of search | Raise bids and placement modifiers on the terms where you've slipped off page one | Impressions recover within days | 3–7 days | CPC rises across the whole ad group, and once a competitor learns you'll pay more, the auction price stays there. |
| Slow Keep stock and the Buy Box | Protect availability and price on the SKUs that carry your ad budget | Impressions return without touching a bid — ads stop serving when the Buy Box goes | Ongoing | Deeper stock ties up cash and racks up storage fees, and defending the Buy Box on price cuts the margin the orders were meant to earn. |
| Slow Widen what you target | Add keywords, product-page targets and placements you're not bidding on yet | Reach grows into audiences you've never shown to | 1 quarter | New reach is cheap and often irrelevant. Click-through rate dips while the account learns, and weak relevance can raise your click price for weeks. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Impressions counts who saw the ad. The number beside it says whether the ad deserved to be seen.
Reach that lands
More shoppers saw the ads and more of them tapped. Targeting and creative are both working, so push budget while the click price is still reasonable.
Seen and ignored
You've widened reach into people the ad doesn't appeal to. Click-through rate is falling, and on Amazon poor relevance eventually raises what you pay per click too.
Tighter targeting
Less reach, more traffic. Negatives and tighter match types are doing their job — you're showing to fewer people who actually want the product.
Out of the auction
You've dropped out of the auction. A stockout, a lost Buy Box or a spent-out budget all stop ads serving, and none of them show up as a campaign error.
This pair separates a budget problem from an auction problem. Impressions falling while CPC climbs means competitors outbid you and the same money buys less. Impressions falling while CPC sits still means you ran out of budget, stock or Buy Box.
Ads for a listing you don't hold the Buy Box on stop serving, so these two move together in a way no bid change explains. When reach drops on a day your bids and budgets were untouched, this pairing usually names the culprit — and the fix is pricing or supply, not advertising.
Common misreads
They're free to receive and expensive to buy. Reach bought with loose targeting drags your click-through rate down, and weak relevance quietly raises what you pay per click across the ad group.
Demand is one of the least likely explanations. Check the budget cap, then stock, then the Buy Box — all three stop ads serving while shoppers carry on searching.
It means the ad was served. Plenty sit below the fold or four rows into a results page nobody scrolled, which is why 500,000 Impressions on one campaign can produce fewer clicks than 100,000 on another.
Also called
Ad impressions · total ad views · serves
See yoursYour Amazon ad impressions for the last 30 days, next to Clicks and Ad Spend.
Open Ads Analytics →Impressions
What it means
Impressions is the number of times one of your Google ads was shown — a Search result, a Shopping unit, a Display placement. It counts serves, not people, and nothing was clicked or charged. This is the one platform of the three where the app also tells you what you missed: Search ISSearch ISSee the full entry. is the share of the Search impressions you were eligible for that you actually won.
Show the math
Formula and a worked example
Total times ads were shown counts each serve on Google. The denominator that makes it readable is the eligible pool — every impression your keywords and products qualified for. Divide your count by Search IS and you have that pool.
Worked example. Search takes 200,000 impressions in March at a Search IS of 40%. The eligible pool was 200,000 ÷ 0.40 = 500,000, so 300,000 impressions were available and went elsewhere. Those 200,000 earned 6,000 clicks — a rate of 3%. Winning the rest at the same rate would be another 9,000 clicks.
Search and Display are separate auctions with separate pools, which is why the app carries both Search ISSearch ISSee the full entry. and Display IS rather than one figure.
It answers the question
How much of Google's available exposure did you actually take? On Search the answer has a known maximum, so a low count can be read as headroom rather than as a mystery.
Why it matters
Impressions is the top of every Google Ads number underneath it. ClicksClicksRaw traffic from ads; only valuable if clicks convert, so read beside CTR. is a slice of it, CostCostRead against conversions and value-per-cost, never alone. is what that slice was billed at, and ConversionsConversionsSales or actions Google drove; pair with cost-per-conversion. is a slice of the slice. A month where cost held and conversions fell reads very differently once you know whether exposure moved.
It also tells you which kind of problem you have. Impressions falling with Search IS steady means the market searched less. Impressions falling with Search IS falling means you lost position or ran out of budget — the same symptom, two different weeks of work.
What good looks like
There's no published band, and there shouldn't be: the right count depends on how many keywords and products you run and how much people search for them. Judge it against your own trend, then against Search IS for the part you can control.
This is also why the Google count can't be scored against the Amazon or Meta tabs. Here the ceiling is measured and quoted back to you; on the other two the app publishes no share-of-available figure at all, so a bigger number on one platform says nothing about performance on another.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Raise budgets where campaigns stop early | Lift the daily cap on campaigns that spend out before the day ends | Impressions up the same day | 1–3 days | Cost rises immediately and the impressions you buy back are the ones the auction ranked last, so Cost/Conv. usually worsens before it settles. |
| Fast Bid into low Search IS | Raise bids on the campaigns and keywords where Search IS shows the most headroom | Impressions and Search IS climb together | 1–2 weeks | Avg. CPC rises across the whole ad group, and the impressions you gain are the ones you were previously outranked on — the hardest, most expensive slice of the pool. |
| Slow Clear what Product Diagnostics flags | Fix the feed problems on the Product Diagnostics table so blocked products can be shown at all | Impressions return on products that were serving nothing | 2–6 weeks | Feed work is slow and invisible, and the recovered exposure lands on whatever margin those products carry — often the ones nobody was watching. |
| Slow Add keywords and products you don't cover | Extend the account into terms and catalogue lines you've never bid on | A bigger eligible pool, not a bigger share of the old one | 1 quarter | Search IS can fall even as Impressions rise, because you enlarged the denominator. New terms also convert worse at first, so Cost/Conv. climbs while they settle. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Impressions says how much exposure you took. The metric beside it says how much there was to take.
Together they recover the eligible pool, which neither shows alone. Impressions flat with Search IS falling means the pool grew and you didn't grow with it — competitors moved in while your numbers looked unchanged. Impressions falling with Search IS steady means the pool itself shrank, and there is nothing to fix.
Search IS is the share of exposure you won; Click Share is the share of clicks. Winning a high share of impressions but a low share of clicks means you are showing up and being passed over — a copy and offer problem that no bid increase repairs.
Exposure is only worth buying at a price. Impressions up with Cost/Conv. flat is clean scale; Impressions up with Cost/Conv. climbing means the extra exposure came from further down the intent curve, and the next increment will cost more again.
Common misreads
Check Search IS before you touch a bid. If your share held while the count fell, you kept your position and the market searched less — raising bids there buys nothing that exists.
You are already taking nearly every impression you're eligible for on Search. Growth from here comes from qualifying for more — new keywords, new products, wider match — not from spending harder on the same pool.
They're different surfaces with different intent, which is why the app reports Search IS and Display IS separately. A Display impression from someone reading an article is not the same event as a Search impression from someone typing your product name.
Also called
Impr. · times your ad was shown · ad views
See yoursYour Google impressions for the period, next to Search IS, Clicks and Cost, split by campaign and by product.
Open Google Ads →Impressions
What it means
Impressions is the number of times your Meta ads were shown across Facebook and Instagram placements. It counts serves, not people: one person shown the ad five times counts five. Meta reports the people separately as ReachReachYour true audience size; compare with frequency for saturation., and the average serves per person as FrequencyFrequencyA read on ad fatigue in your audience.Under 2 is healthy, so on this platform the three are locked to each other by arithmetic.
Show the math
Formula and a worked example
Total times ads were shown counts each serve on Meta. Its own denominator is people: Impressions = Reach × Frequency, which is why Meta can grow this number without reaching a single extra person.
Worked example. March takes 400,000 impressions from 100,000 people — a Frequency of 4.0. In April you raise budget and take 600,000 impressions, but Reach only grows to 120,000, so Frequency is 600,000 ÷ 120,000 = 5.0. You bought 50% more serves and 20% more people. The rest went to people who had already seen it.
Impressions can be added across days, ads and campaigns. Reach cannot — the same person appears in more than one bucket, so summing double-counts. That is why the Ad Formats table carries Impressions and leaves Reach out.
It answers the question
How much exposure did the budget actually buy on Facebook and Instagram? On its own it says nothing about how many people that was, which is the question most readers think they're asking.
Why it matters
Impressions is the denominator under CTR (Link)CTR (Link)How enticing your ad is to the people who see it.1.5% or more is healthy and the cost-of-exposure columns on the Ad Formats table. When it moves, several figures move with it for reasons that have nothing to do with the creative.
It's also the early warning on audience size. Spend can keep buying impressions long after it has run out of new people to show them to, and the count itself never shows that — it just keeps going up while the same audience sees the ad again.
What good looks like
No band applies. A count like this scales with budget, audience size and how many placements you allow, so a figure that's healthy for one account is a slow week for another. Read it against your own trend, and always beside Frequency — Impressions alone can't tell you whether you found new people or repeated yourself.
The Google Ads tab has a measured ceiling in Search IS and this one does not: the app publishes no share-of-available figure for Meta. Comparing the two counts, or adding them, produces a number with no definition behind it.
How to improve it
| Lever | What you do | Expect | How long | Watch out for |
|---|---|---|---|---|
| Fast Widen the audience before you raise the budget | Loosen targeting on ad sets whose Frequency has been climbing | Impressions grow through new people rather than repeats | 3–7 days | The new people carry less intent, so CTR (Link) usually dips first and the cost per order follows it down the same path. |
| Fast Add budget only where there's room | Raise spend on ad sets whose Frequency is still low, not on the ones already repeating | Impressions up roughly in line with spend | 1–2 weeks | With the audience unchanged, most of the extra serves land on people who already saw the ad — Frequency climbs and response falls while the count still looks like growth. |
| Slow Open placements you've switched off | Allow the placements and surfaces you currently exclude, then read them on the Placements table | More exposure at a lower cost per thousand | 2–4 weeks | Cheap placements pull the blended cost of exposure down while sending less qualified traffic, so Impressions rises faster than Link clicks and the averages flatter you. |
| Slow Put genuinely new creative in | Build new concepts rather than variants, and judge them by format on the Ad Formats table | Exposure holds without Frequency climbing on the same tired ad | 1 quarter | New creative competes with your proven creative for the same budget, so the account's blended result usually dips while you find out which one wins. |
Every lever costs something somewhere. The last column is the one to read twice.
Read it with
Impressions counts serves. On Meta the number beside it decides whether those serves reached anyone new.
Divide one by the other and you have Frequency. 400,000 impressions from 100,000 people is a campaign finding new eyes; 400,000 from 40,000 people is the same ten people over and over. The two call for opposite decisions and the impression count alone cannot tell them apart.
This is the pair that tells you when a budget increase stopped working. Impressions up with Frequency flat means you bought more people; Impressions up with Frequency climbing means you bought the same people again, and the response rate will fall behind it.
Impressions is the denominator underneath CTR (Link), so exposure that grows into a weaker audience shows up here as a falling rate before it shows up as a worse cost per order. It's the earliest read you get on creative or audience fatigue.
Common misreads
Reach counts the people. Impressions can double with the audience unchanged and Frequency doing all the work — which is the single most common way a Meta budget increase buys nothing.
It's deliberate. Reach and Frequency count unique people, so they can't be added across ads without counting the same person twice. The table leaves them out rather than print a total that would be wrong.
Three platforms counting their own surfaces under their own rules, and the same person can be served on all three. The sum is not reach and it is not a total — it's a number with no definition. Read each against its own history.
Also called
Ad impressions · serves · ad views
See yoursYour Meta impressions by creative format, next to the Spend and Link clicks behind each one.
Open Ad Formats →