Measurement
Part of A method for measuring and reporting short form video performance that avoids vanity metrics
Where can you find reliable short form video benchmark research?
Short form video benchmark research is scarce and often misread. Here is how to judge a benchmark study, apply it to England, and report it honestly.
What to take away
- LinkedIn's own advertising resources, reviewed 11 September 2026, set out video formats and campaign objectives but publish no universal benchmark for views, watch time or cost per result.
- The reason is structuralbenchmarks depend on placement, objective, creative length and audience, so a single number rarely transfers between accounts.
- UK law shapes what you can collect and report. The Data Protection Act 1998 has been replaced by the UK GDPR regime, and the ICO's international transfers guidance governs moving that data offshore.
- The practical answer is to build a local baseline from your own account data, then use published platform guidance to sanity-check it.
- Record the date, source and definition behind every number you quote, or the benchmark becomes decoration.
Why published benchmarks are thin
Most platforms publish guidance on formats, objectives and creative best practice rather than performance averages. The LinkedIn Success Hub is a useful example: it explains how to structure video campaigns and which objectives suit them, but it does not promise a benchmark figure for completion rate or cost per thousand impressions.
That is not an oversight. A benchmark is only meaningful alongside the conditions that produced it. Placement, objective, creative length, audience size and season all move the number. A study run on one placement in one quarter tells you about that placement in that quarter.
So the honest position for a marketing team in England is that no single published figure can stand in for your own baseline. Treat any vendor who offers one without conditions as a sales document, not research.
What counts as good research
Good benchmark research states its method before its numbers. Look for the sample period, the account or campaign mix, the metric definition and the geography. If a study reports average watch time, ask whether it excludes views under three seconds. If it reports cost per lead, ask which objective and bidding strategy produced it.
What good benchmark research states
- Sample period
- Account or campaign mix
- Metric definition
- Geography
- View threshold (e.g. three seconds)
- Objective and bidding strategy
Definitions matter more than sample size. Two studies can both be accurate and disagree because one counts a view at one second and the other at thirty. When you compare your account against an external figure, compare definitions first.
The measurement and reporting guide sets out how to align definitions across channels before you benchmark anything, which is the step most teams skip.
Building a local baseline
Start with your own account. Pull twelve months of data by placement and objective, then group campaigns by creative length. You will usually find that the spread within your own account is wider than the gap between your account and any published average.
Building a local baseline
- Pull twelve months of account data
- Split by placement and objective
- Group campaigns by creative length
- Compare spread within your account
- Label every figure as measured or illustrative
Label every figure you present. If you estimate that a team spending £4,000 a month on vertical video might see a 15% swing in cost per result between placements, say that it is an illustrative example, not a measured outcome.
For a worked method, see the common measurement mistakes in England, which covers the errors that make a baseline unusable before the first comparison is made.
Data protection duties
Benchmark work often involves exporting audience and performance data, sometimes to a platform or agency outside the UK. That processing sits under the UK GDPR regime, which replaced the Data Protection Act 1998, a statute whose text remains on the legislation.gov.UK record for historical context.
If personal data leaves the UK, the transfer needs a lawful mechanism. The ICO's guidance on international transfers explains the safeguards and the assessment you should document. Build that check into your reporting cycle rather than bolting it on at audit time.
Reporting a benchmark honestly
State the source, the date and the definition in the same sentence as the number. If you cannot, drop the number. A benchmark without provenance invites a challenge you cannot answer.
Separate observation from interpretation. "Completion rate fell two points in the quarter" is observation. "Creative fatigue set in" is interpretation. Both are useful, but they carry different weight in a board pack.
Finally, date your benchmarks and review them quarterly. A figure from eighteen months ago may still be the best available, but the reader deserves to know its age.
Decision table
Choose
- No internal baseline yet
- Twelve months of your own account data, split by placement
- Vendor supplies a benchmark
- Ask for method, period, definition and geography
- Data exports leave the UK
- Document the transfer mechanism per ICO guidance
- Reporting to a board
- One dated figure with its definition
- Comparing across channels
- Aligned metric definitions first
Avoid
- No internal baseline yet
- Quoting a vendor average as your target
- Vendor supplies a benchmark
- Accepting the figure without conditions
- Data exports leave the UK
- Assuming the platform handles compliance
- Reporting to a board
- A slide of undated averages
- Comparing across channels
- Ranking channels on mismatched metrics
Common questions
Are published short form video benchmarks reliable?
They are reliable only within the conditions stated. A figure without a period, placement and metric definition cannot be checked, so treat it as indicative at best.
How often should a benchmark be refreshed?
Quarterly is a sensible cycle for most marketing teams, because placement mix and creative formats shift within a year. Refresh sooner if you change objective or platform.
Can I use a US benchmark for an English audience?
Only with care. Audience behaviour and platform mix differ, so a US average is a starting hypothesis, not a target. Validate it against your own account data before quoting it.
Does benchmark research involve personal data?
It can, if you export audience-level or lead-level records. Check the ICO's international transfers guidance before moving that data outside the UK.
