Careers
Part of Display advertising: signal versus noise
Display advertising questions: what people ask and what is true
Common display advertising questions cover formats, inventory, audiences, placements, viewability, attribution, frequency, creative, costs, ownership, and control.
What to take away
- Two beliefs fail most oftena click proves interest, and display only works for awareness. Both are false.
- Formats, placements and inventory are facts a platform can list. Guessing at them is unnecessary.
- The display viewability bar is 50 percent of pixels in view for one continuous second.
- Pricing is quoted in CPM. Open-exchange display in the US commonly sits between about $1 and $15 per thousand impressions as a working range.
- Frequency caps, attribution windows and placement exclusions are settings you control.
The questions people bring, and the short answers
Most display questions fall into a small set of groups. Display advertising covers how to tell signal from noise in the reports behind them. The table gives the short answer and flags the belief that usually turns out wrong.
Questions and short answers
| Question | Short answer | Belief check |
|---|---|---|
| Which formats exist? | Banners, native, in-feed video, interstitials, rich media | False: display means banners alone |
| Where do the ads appear? | Publisher sites, apps, exchanges, partner networks | True: placements can be listed and excluded |
| Who is the audience? | First-party lists, third-party segments, contextual targeting | Partly false: third-party reach has shrunk |
| How often does one person see it? | Whatever the frequency cap allows | False: platforms cap well by default |
| Is the ad viewable? | Only at 50 percent of pixels for one second | False: served means seen |
| What does it cost? | A CPM, sometimes a CPC or CPA | False: one benchmark covers all inventory |
| How is credit assigned? | Click and view-through windows, plus tests | False: last click is the truth |
| Who controls placement? | You, through exclusions and verification tools | False: the network decides alone |
Each question needs its own evidence set before an answer holds up.
Evidence needed per question
- Nativeformat, disclosure, context, contract
- Viewabilitybase, threshold, vendor, outcome
- Frequencyidentity coverage, reach, complaints, lift
- Direct dealssupply path, fees, placement, delivery
- Ownershipadmins, billing, assets, data, exit
Formats, inventory and placements
Display covers fixed-size banner units, native units that match a publisher layout, in-feed and outstream video, and full-screen interstitials. Format is only half the picture.
Placement type decides where the creative lands. Managed placements name the site or app. Open-exchange buys can land anywhere unless you add exclusions, and private marketplace deals sit between the two.
Supply quality varies across exchanges, so treat inventory as a set of sources rather than one bucket. Improving display advertising traces supply before spend, which fits this stage of the work.
What impressions and viewability figures measure
The Media Rating Council treats a desktop display impression as a filtered ad request recorded late in delivery, once the creative loads and begins to render. Its stated scope and its 2017 update matter. Other environments follow their own standards.
Viewability is a separate state with its own bar. The MRC and IAB standard for display counts an ad viewable when 50 percent of its pixels are in view for at least one continuous second. A served ad can miss that bar and still sit in a dashboard.
Check measurability on its own. When only part of the inventory can be measured, the remainder is unknown rather than clean.
Frequency, attribution and cost
Frequency caps set how often one person sees a campaign. Common practice runs from about 3 to 10 impressions per user per day per campaign. Lower caps protect the budget, while higher caps chase reach that may already be saturated.
Attribution needs a stated window. View-through windows of one to seven days and click windows of up to 30 days are both common settings. Compare results only within one convention, because mixed conventions produce confident and wrong answers.
Costs arrive as CPMs. Open-exchange display in the US commonly runs from roughly $1 to $15 per thousand impressions, with premium publisher and video inventory far above that. Ad serving, verification and data fees usually sit outside the media line.
Five steps keep an estimate honest:
- Define the audience and the reach you expect.
- Choose the buying pathmanaged placement, private deal or open exchange.
- Apply a working CPM and a frequency cap.
- Add the fees that sit outside media.
- Reserve part of the budget for an incrementality test.
The GAO evaluation design guide links evaluation questions to evidence needs and design choices. A display test needs the same discipline.
Who controls where the ads appear
Brand safety is a contract and a setting, not a promise. Buyers set account-level placement exclusions and blocklists in Google Ads, Display & Video 360, Amazon DSP and The Trade Desk. Verification vendors such as Integral Ad Science and DoubleVerify report on the placements those settings allow.
The ads.txt and sellers.json standards from IAB Tech Lab list who may sell a publisher's inventory. The Open Measurement SDK standardizes viewability measurement inside apps.
False signals get sold as inventory, and social advertising mistakes shows how deceptive interfaces and fake engagement work on the paid side. Display placements deserve the same skepticism.
Audience targeting carries privacy duties. The W3C Privacy Principles expects the system, not the individual, to carry privacy work. Apply that before widening a tracking-based audience.
Canada is the second market here. PIPEDA governs how personal information is collected and used for online tracking, and the Office of the Privacy Commissioner publishes PIPEDA compliance guidance on consent for that work. Quebec also requires French in commercial advertising under its Charter of the French Language.
Example: reviewing one placement report
Reviewing one placement report
- Pull the placement report for the period and sort by spend.
- Compare impressions, clicks and viewability for each site.
- Check the attribution window and whether view-through credit is included.
- Open the top three placements and read the page as a visitor would.
- Block what fails, keep what converts, and date the decision.
