3 things to weigh up on video advertising benchmarks

Video benchmarks are not a separate discipline from PPC ad optimization. They belong in the same paid media workflow as paid search, shopping, and social. Use the same account structure, audience segmentation, bid rules, and cost per outcome tests. Size video tests with these ranges, then compare them with your own PPC cohorts. One standard governs every paid channel.

Three things to weigh up on video advertising benchmarks are the denominator, the view definition, and the full cost per outcome.

What to take away

Why this matters for PPC optimization

Video advertising benchmarks are not a separate universe from PPC ad optimization. They are another place where the same discipline applies. In paid search you already split by campaign type, device, match type, and market before you judge a cost per click or conversion. Video ranges need that same treatment.

A completion rate without its denominator is like a conversion rate without its conversion definition. Use these video ranges to plan tests, then compare them with your own PPC cohorts. The goal is not to copy a benchmark. It is to make every paid channel, including video, answer to the same cost and outcome standard.

Benchmark ranges by format and device

The table gives typical planning ranges for US campaigns. They are rounded, not audited medians, and they move with audience, season, and supply. Use them to size a test, not to settle an argument.

Every rate sits in one of six layers: delivery, opportunity, attention, response, outcome, and economics. Mixing layers is the most common benchmarking error.

[figure 1]

The viewability column is comparable because of one public rule. The MRC video viewability standard counts 50% of pixels in view for two continuous seconds. Display uses one continuous second, and accredited vendors quote that scope. Completion is not comparable until you know its denominator.

A completion rate over starts and a completion rate over impressions can differ by more than double. Both can be reported honestly. They answer different questions.

Format choice follows the audience job, so a video advertising plan built on that job keeps the comparison honest. Ranges never set a budget on their own.

Why the same word means different numbers

Google documents a skippable in-stream view at 30 seconds of watch time, or the full ad when it is shorter. A 6-second bumper counts as viewed when it plays through.

Meta counts a 3-second view, and ThruPlay for videos of 15 seconds or longer. TikTok has counted an in-feed view at roughly one second of play. Three help pages, three definitions.

Record format, definition, billing rule, and report date before comparing anything. Build the cohort in this order:

Example: cost per completed view

A plan with a $30 CTV CPM beside a $0.02 YouTube CPV looks like a 1500x gap. It is not.

At 90% completion the CTV spot costs about $0.033 per completed view. The YouTube buy at 65% completion costs about $0.031. Normalize before calling one side cheap. Worked format comparisons appear in video advertising examples.

Canada: currency, tax, and market size

If you buy Canadian inventory, campaigns are quoted in CAD. At roughly 1 CAD = 0.70 to 0.74 USD, a C$25 CPM is about US$18 to US$19.

Canadian CPM figures often land in the same numeric band as US ones before conversion. Audiences are smaller and premium inventory is thinner, so the spread is wider. GST is 5%, and provincial HST runs 13% to 15%.

The Canada Revenue Agency explains GST/HST rules for businesses, including input tax credits for advertising expenses (CRA GST/HST guidance). Budget the gross figure and claim the credit on the return.

Quebec requires French-language commercial advertising under the Charter of the French Language. Consent rules under PIPEDA shape what you can track, so a Canadian cohort can carry more missing data than a US one. Display units obey the same rules, and programmatic advertising benchmarks set out the 2027 view on distributions and full cost.

What to check before you use a benchmark

A reviewer should reproduce the reasoning without relying on memory.

[figure 2]

The GAO evaluation design guide ties an evaluation question to the evidence it requires, which is the test to apply before calling an observed lift causal. The NIST guidance on experimental design starts from the objective and the practical constraints, which is why a benchmark comparison and a controlled test are different instruments.

Benchmarks size the test. Cohorts and settled outcomes set the budget. An economics-first paid media strategy development checklist keeps media, production, and labor cost in one place.

Common questions

What is a good video completion rate?

There is no universal rate. A 6-second bumper typically completes on 85 to 95% of starts. A 30-second in-stream ad often completes on 55 to 75%. State the format and the denominator before judging either.

Which CPM should go into a 2027 budget?

Start from the format ranges above. Convert to CAD for a Canadian buy and add GST or HST to the gross figure. Replace the range with your own cohort median as soon as you have one.

Can Google, Meta, and TikTok views be compared directly?

Only after you match definition, initiation, thresholds, filtering, and report date. Three view definitions produce three different rates for the same audience.

Do I still need my own data?

Yes. External ranges size a test. Budget decisions belong on your own cohorts, full costs, and settled outcomes.