Cost-per-click bidding and Cost-per-thousand-impressions bidding at a glance
Under cost-per-click buying (CPCAverage cost of each click.), the advertiser pays only when someone clicks. The platform still has to show the ad many times to earn each click, so it prices that in: it favours ads and audiences that are likely to be clicked, and it ranks ads partly by their expected click-through rate (CTRShare of impressions that result in a click.). Search advertising is built on this model, and many social and display platforms offer it as an option.
Under cost-per-thousand-impressions buying (CPMCost per 1,000 impressions.), the advertiser pays for each thousand times the ad is served, whether or not anyone responds. Display, video, connected TVProgrammes and streams watched through an internet-connected television, with advertising in the breaks, usually sold on impressions rather than rating points. and much of social advertising are priced this way, and programmatic auctions clear on impression prices, as the real-time bidding guide explains. The industry's standards bodies, such as the IAB, define how impressions are counted and when an ad counts as viewable, which matters because an impression is only worth buying if it had a chance to be seen.
The two are linked by one number. Multiply the cost per click by the click-through rate and by a thousand, and you get the effective cost per thousand impressions; reverse it to get an effective cost per click. The CPC calculator, CPM calculator and CTR calculator do the arithmetic. Buying models beyond these two, such as cost per view, cost per action and flat fees, belong to the media buying models guide.
Compare table
Cost-per-click bidding vs Cost-per-thousand-impressions bidding
| Row label | Cost-per-click bidding | Cost-per-thousand-impressions bidding |
|---|---|---|
| What you pay for | A click on the ad | Each thousand impressions served, viewed or not |
| Who carries performance risk | Mostly the platform: an ignored ad costs the advertiser nothing directly | Mostly the advertiser: impressions are paid for whether they earn a response or not |
| Best-fit goals | Website visits, leads, sales, app installs where the click is the first step | Reach, frequency, brand awareness, video views, launches, recall |
| Effect of click-through rate | A higher rate helps ad rank and delivery, so strong ads win more volume at the same bid | A higher rate lowers the effective cost per click, so strong ads get cheaper visits |
| Control over reach | Indirect: delivery follows likely clickers, so reach can be narrow | Direct: reach and frequency can be planned and capped |
| Platform availability | Standard in search; offered by many social and display platforms | Standard in display, video, programmatic and much of social |
| Reporting | Reported as cost per click, clicks and onward conversions | Reported as cost per thousand, impressions, reach, frequency and viewability |
| Translating one into the other | Effective CPM equals cost per click times click-through rate times a thousand | Effective CPC equals cost per thousand divided by a thousand times click-through rate |
The most important row is the second. Choosing a pricing model is choosing who absorbs the cost of ignored ads, and the platform's delivery system behaves differently depending on that choice.
When to choose Cost-per-click bidding
Pay per click when the campaign exists to send people somewhere. If success is measured in visits, leads or sales, the click is the first event that has any value, and paying for it directly aligns the price with the goal. Search is the clearest case: the person has typed a need, the ad answers it, and an impression without a click usually means the ad was not relevant to that searcher.
Click buying also suits advertisers who cannot yet predict their click-through rate. A new campaign with untested creative may earn very few clicks per impression; buying on impressions would mean paying for that uncertainty, while buying on clicks hands it to the platform. Small budgets benefit for the same reason: every unit of spend buys a visit rather than exposure that may not convert.
The trade-off is control. Because the platform earns only when people click, it steers delivery towards the people and placements most likely to click, which are not always the people the brand most needs to reach. Accidental clicks, curiosity clicks and placements that attract them can look efficient on cost per click and still produce poor visitors. That is why the click should never be the final measure; follow it through to conversion before deciding a click campaign is working. Where a campaign has enough conversion data, the next step is often bidding on the outcome itself, which the target CPA versus target ROAS comparison covers.
When to choose Cost-per-thousand-impressions bidding
Buy on impressions when being seen is the job. Launches, awareness campaigns, video storytelling and campaigns that aim to build memory over time all depend on reaching the right people often enough, and impression buying lets the advertiser plan reach and frequency directly rather than hoping a click-based system distributes the ad widely.
Impression buying also rewards strong creative. If an ad reliably earns clicks, buying impressions turns that strength into cheaper visits, because the advertiser captures the benefit of a high click-through rate instead of the platform. Experienced advertisers with proven creative sometimes switch campaigns to impression buying for this reason, then check that the effective cost per click and the cost per conversion actually improved.
The risk is paying for exposure nobody noticed. Impressions served below the fold, in cluttered placements or to the wrong audience cost the same as impressions that land. Use viewability measurement where the platform offers it, cap frequency so the same people are not paid for repeatedly, and set a review that ties impression spend to a downstream signal: brand search, site visits, assisted conversions or a lift study. Reading the cost per thousand together with click-through and cost per acquisition, as the note on reading CPM, CTR and CPA together sets out, keeps a cheap impression price from hiding an expensive outcome.
A furniture brand tests both models on the same audience
Oak & Elm, an online furniture brand, runs a social prospecting campaign at a CPM of £8 with a click-through rate of 0.9%, an effective cost per click of about £0.89. A parallel ad set bought on clicks costs £1.10 per click. The impression-bought ad set looks cheaper per visit, but its visitors convert at 1.1% against 1.6% for the click-bought set, so cost per order is roughly £81 against £69. The brand keeps click buying for prospecting, moves its proven video creative to impression buying for a separate reach campaign with a weekly frequency cap of three, and judges that campaign on brand-search growth rather than on cost per click.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Search campaign aiming for leads or sales | Cost per click | The click follows a stated need and is the first valuable event |
| Product launch that needs broad, repeated exposure | Cost per thousand impressions | Reach and frequency can be planned directly |
| New creative with an unknown click-through rate | Cost per click | The platform carries the risk of a weak response |
| Proven creative with a consistently strong click-through rate | Cost per thousand impressions, tested | Strong response turns into cheaper visits |
| Video, connected TV or audio campaign | Cost per thousand impressions | The inventory is priced and measured on exposure |
| Small budget that must produce visits | Cost per click | Every unit of spend buys a visit rather than exposure |
| Campaign has enough conversions to bid on outcomes | Neither: move to outcome bidding | Bid on the result the business values |
Common mistakes
- Comparing a CPC with a CPM directly, without converting both to the same unit through the click-through rate.
- Judging a click campaign on cost per click alone, while the visitors it buys never convert.
- Buying impressions without viewability or frequency controls, so a low CPM pays for ads nobody saw or the same people saw too often.
- Using click buying for awareness, then wondering why reach is narrow and concentrated on habitual clickers.
- Switching models mid-campaign without a test, so any change in results cannot be attributed to the pricing model.
- Treating a high click-through rate as success on an awareness campaign whose job was memory, not visits.
- Ignoring placement quality. A cheap price on either model often reflects placements that attract accidental clicks or unseen impressions.
