Customers often want to know whether a campaign is "worth it" to measure. Two questions decide it, and both need a yes:

1. Is the product available in this market at all? Availability varies by region, and no amount of scale overcomes a product that does not operate somewhere.
2. Is the campaign large enough? Measurement is a statistical exercise, so a campaign below a certain scale cannot produce a conclusive result even where everything is available.

The matrix below answers both at once. Everything after it covers the second question in more depth.

🚧

Feasibility Does Not Guarantee Significance

Meeting these guidelines does not guarantee positive significant results. It is impossible to guarantee statistical significance, because whether or not a result is statistically significant depends on the magnitude of the result and the number of observations.

For example, if a particular campaign ran with a black screen as the creative, it would not be reasonable to expect there to be any lift, and thus would be impossible to guarantee significance even if the campaign had 1 billion impressions.

On the flip side, these are not hard thresholds, so if you do not meet these guidelines, it's still possible to observe significant results

Availability and Impression Guidelines

Each cell gives the minimum recommended impressions (in millions) for that product in that region, or states that the product is not available there.

Measurement

Footfall, Web, and App are the three attribution conversion environments. Impressions (IQ) is Mira's impressions product, which does not carry an impressions threshold of its own. Brand Lift is described in Brand Lift Surveys.

Data and Activation

Raw MAIDs and DSP Push are delivery methods for Automated Data Exports. Retargeting delivers exposed device identifiers to a partner for re-serving. These are data deliveries rather than measurement, so they carry no impressions threshold.

📘

What "Case by Case" Means

Not a soft yes or a soft no. Availability depends on specifics that have to be checked for the individual engagement - which conversion partners operate in that country, and what local privacy regulation permits for the data involved. Ask your account team early, because the answer affects what can be promised in a proposal.

Where a cell reads Available without a figure, the product runs in that market but no published impressions guideline exists for it yet. Your account team can advise on scale for a specific campaign.

If the estimated impressions fall below a stated threshold, it is still possible to observe meaningful results, so always feel free to reach out to our support team.

Region Definitions

Europe currently covers:

Germany, United Kingdom, France, Italy, Spain, Netherlands, Belgium, Sweden, Austria

LATAM currently covers:

Mexico, Brazil

If you are interested in a country that is not currently listed, reach out to your account team.

Reading the Matrix

Brand Lift travels furthest. It is the only product available in every region, because it does not depend on observing a behavioural outcome - it asks people directly. Where a market has enough location data to identify exposure but no usable conversion environment, brand lift is frequently the only measurement available.

Attribution narrows fastest. Each conversion environment carries its own dependency. Footfall needs sufficient location data density in-market. Web and app need conversion partners operating in that country and permitted to send data. Those dependencies are met in different combinations in different places, which is why the three do not move together.

Availability is not the same as comparability. Where a product runs in two regions, the methodology is the same but the underlying data is not. Panel density, conversion coverage, and regulation all differ, so results should not be compared across regions as though they were like-for-like. A multi-market campaign is usually best read market by market.

Coarse Location Data Within the USA

The USA column is the national picture, and it is not uniform. A growing number of states restrict the sale of precise geolocation data, and in those states location data reaches Mira coarsened rather than not at all.

Measurement continues there using Exposure Zones - larger zones drawn around each asset in place of a precise viewshed. This keeps campaigns in those states measurable, at a real cost: the exposed group takes on false exposures, results are attenuated, and a null becomes the expected outcome rather than a finding.

Because it is a different measurement rather than a degraded one, it is opt-in per Study and per export, and excluded by default.

🚧

Check the Current State List Before You Commit

Four states currently have restrictions in effect, with two more enacted and awaiting their effective date and others under active consideration. The list changes, sometimes without notice. Exposure Zones for Coarse Regimes carries the current list along with what the regulations restrict, and a campaign with meaningful spend in a restricted state should have its measurement approach agreed before the buy rather than after.

Conversion Guidelines

In addition to needing a certain number of impressions, we also want to ensure that the conversion environments are well-trafficked as well. For example, a customer wants to measure impact on a website that has 1 unique visitor per day, it is unlikely we will be able to observe a statistically significant lift, because the volume is so low.

Additional Considerations

Campaign Duration

We recommended a minimum campaign duration of at least 1 month. If the campaign impression count is below the recommended threshold, a length of longer than 1 month is recommended. Longer campaign durations are recommended (longer than one month), even if the campaign achieves impression recommendations.

Markets and Geographical Diversity

Using multiple markets will yield the highest number of unique devices. For campaigns in one or two markets, we are more likely to observe statistically significant results if the campaign units are distributed well across the markets, rather than concentrated in one area.

Explanation

OOH attribution is based on sampling the population and running a statistical hypothesis test. As such, to obtain meaningful insights we need to have enough observations.

While we can measure and report on any sized campaign, with very small campaigns it is unlikely that we will be able to statistically prove that the campaign had an effect on the population (that is, on all users, including those we did not observe). For example, if we only observe 100 users in each of the control and exposed group, we might observe a 15% lift in conversions, but because the sample is so small we cannot say it is statistically significant.

The statistical significance of the results we report typically depend on three things:

• the sample size (number of users in the exposed and control groups)
• the number of successful conversions (or: the conversion rate in conjunction with the sample size)
• the size of the effect (i.e. the magnitude of the net lift)

This is a complicated balancing act that makes it difficult to establish thresholds that we are certain will lead to significant results if met. For example, it's possible to observe a significant result with a small number of users if the lift is very large; but it is also possible to not observe a significant lift with enormous sample sizes. This is expected -- if there is truly no effect of the media on conversion, then we shouldn't be detecting significant differences between the exposed and control groups even if we observed the entire population.