Channel and measurement choices by market
A multi-market plan fails in two predictable ways. It buys channels that are unavailable, ineligible or unmeasurable in some markets, and it spreads the budget so thinly that no market produces a readable result. The framework below runs five gates per market and channel, funds only the lines that pass, and assigns each line a measurement method that can answer the budget question at that spend level. Platform attribution, lift tests, geo holdouts and marketing mix models answer different questions, and each needs a different amount of data.
Executive findings
- 1
TikTok publishes which placements are available in which locations, so the same campaign set-up can have different inventory per country.
Documented: official source. Scope: TikTok Ads Manager placement availability, as of . [1]
- 2
Google requires publishers serving ads to users in the EEA, the UK and Switzerland to use a Google-certified consent management platform, which shapes the signals available for measurement there.
Documented: official source. Scope: EEA, UK and Switzerland; Google consent management requirements, as of . [2]
- 3
Google (Meridian) and Meta (Robyn) both publish open-source marketing mix modelling tools; Meridian documents a geo-experiment integration to anchor the model in causal tests.
Documented: official source. Scope: Open-source MMM documentation, as of . [4]
- 4
On iOS, app-install attribution runs through Apple’s privacy-preserving frameworks, so app campaigns report differently from web campaigns in the same market.
Documented: official source. Scope: Apple ad attribution documentation, as of . [7]
- 5
In our illustrative allocation, 30,000 spread across six markets leaves every market below the two-channel minimum; the same budget across three markets gives each a readable test.
Illustrative model, not a benchmark. Scope: FOCUSADS worked model with labelled assumptions, not a benchmark, as of
Method and limitations
- Question
- How should a business decide which channels to fund in which markets, and how should it measure each line?
- Written for
- CEOs, heads of growth and performance leads planning paid media in two or more countries.
- Decision at stake
- Which markets and channels receive budget this quarter, and which measurement method decides next quarter’s split.
- Sources
- Platform documentation on placement availability, consent requirements, lift measurement, attribution settings and app attribution; Google and Meta documentation for their open-source MMM tools; IAB guidance on modern MMM.
- Period
- Documentation as published on 17 October 2026.
- Markets
- All 32 FOCUSADS markets. Paused markets (China, Russia) do not pass the eligibility gate.
- Definitions
- Gate: a pass/fail check that removes a market-channel line from the plan regardless of its score.
- Readable budget: spend high enough for the chosen measurement method to detect the effect that would justify the line.
- MMM: marketing mix model, a statistical model of outcomes against spend by channel over time.
- Limitations
- Minimum readable budgets depend on conversion volume and noise in each account. The figures in the model are placeholders.
- MMM needs a long history with variation in spend; new markets rarely have it.
- This framework does not replace legal advice on consent or sanctions in any market.
Five gates per market and channel
FOCUSADS view
Availability asks whether the platform sells the format in the country, from its own pages. Eligibility asks whether this advertiser, with its entity, category and payment method, can buy it. Demand asks whether there is evidence of buyers: search volume, marketplace listings, distributor interest. Measurement asks whether the line can be measured under the market’s consent rules. Minimum budget asks whether the line can be funded at a readable level.
A line that fails any gate leaves the plan, however well it scores on fit. The FOCUSADS planner applies the same rule: hard exclusions override the numeric score.
Worked model: concentrate before you spread
Illustrative model, not a benchmark
A business has 30,000 per month for paid media and six candidate markets. Each market needs at least two channels to compare, and each channel needs a minimum readable spend.
| Line | Spread: six markets | Concentrated: three markets |
|---|---|---|
| Budget per market | 5,000 | 10,000 |
| Two-channel minimum per market | 6,000 | 6,000 |
| Markets with a readable two-channel test | 0 of 6 | 3 of 3 |
| Budget left for a third channel or a holdout | 0 | 4,000 per market |
Assumptions
- Minimum readable spend of 3,000 per channel per market is a placeholder for the planner’s channel minimums and your conversion volume.
- The three concentrated markets are the ones that passed all five gates with the strongest demand evidence.
- Markets left out are revisited after one full test cycle, funded from the lines that proved incremental.
Measurement trade-offs
Documented: official source
Platform attribution is fast and available everywhere, and each platform counts from its own point of view, so totals across platforms overlap. Lift tests such as Google’s Conversion Lift and Meta’s incremental attribution estimate what the ads caused, at the cost of holding out part of the audience. Geo holdouts compare regions and work for offline media too, but need comparable regions. Marketing mix models, including the open-source Meridian and Robyn, estimate the contribution of each channel over time and need a long, varied spend history.
Consent requirements in the EEA, the UK and Switzerland, and Apple’s app attribution frameworks on iOS, reduce what user-level attribution can see. Aggregate methods such as geo tests and MMM are less affected, which is one reason to plan them early in those markets.
What we would test, in order
| Step | Measured by | Counts as success when |
|---|---|---|
| Run the five gates for every candidate market and channel and record the evidence for each pass. | Lines passing all gates, with a source for availability and eligibility. | No funded line lacks a verified availability source. |
| Fund the passing lines in the strongest markets at readable budgets, each with a named measurement method. | Incremental outcomes per line from lift or geo tests. | Each funded line produces a result that can move budget next quarter. |
| After two or three cycles, fit an MMM with the test results as calibration. | Model fit against held-out periods and agreement with test results. | The model’s channel estimates stay within the ranges the tests found. |
Measurement method selector
Pick the method by the decision it has to support and the data you have.
| Method | Answers | Needs | Main weakness | Use for |
|---|---|---|---|---|
| Platform attribution | Which ads were present before conversions | Tag or API, consent | Overlap between platforms; credit is not cause | Daily bidding and creative choice |
| Platform lift test | Incremental conversions from one platform | Holdout audience, enough conversions | One platform at a time | Budget level per platform |
| Geo holdout | Incremental outcomes in exposed regions | Comparable regions, regional sales data | Region noise; few regions in small markets | Offline media and whole-market tests |
| Marketing mix model | Contribution of each channel over time | Two or more years of varied spend, or calibration tests | Slow; weak on new channels | Annual and quarterly mix |
How this connects to our method: Assign every channel a role
References
- Placements and Available Locations, TikTok Ads Manager Help, accessed
- Google consent management requirements for serving ads in the EEA, the UK, and Switzerland, Google AdSense Help, accessed
- About Conversion Lift, Google Ads Help, accessed
- Meridian: open-source marketing mix model, Google for Developers, accessed
- Robyn: open-source marketing mix modelling package, Meta Open Source, accessed
- About attribution models and attribution settings, Meta Business Help Center, accessed
- Ad attribution, Apple Developer, accessed
- Modernizing MMM: Best Practices for Marketers, IAB, accessed