A brand tracker in one country is a well-understood job. Global brand tracking, the same tracker run in twelve countries, is a different problem, because every choice has to hold up in every market at once. If you are new to the method, start with what brand tracking is, then come back here for the multi-country parts: questionnaire design, sampling, weighting and reporting at scale.
Key takeaways
- Global brand tracking works when the core questionnaire, metric definitions and method are identical in every market. Only the language and brand list change.
- Each market needs its own quota sample with a base large enough to read change.
- How you weight markets into a regional or global total changes the answer. Choose market-size or equal weighting for a stated reason and hold it across waves.
- Test wave-over-wave change inside each market first. A global number can hide a real move or invent one.
- One data model, with consistent codes and nets, makes hundreds of tables manageable every wave.
What is global brand tracking, and why is it harder?
Global brand tracking is the same brand tracker run in many countries on a schedule, so you can compare markets and watch each one over time. It is also called multi-country or international brand tracking. The goal is results that mean the same thing in São Paulo, Seoul and Stuttgart.
It is harder than a single-market tracker for four reasons. Languages and brand sets differ, sample sources differ, and people use rating scales differently from country to country. The output is also no longer one table but hundreds, refreshed every wave.
In multi-market trackers, the first problem is usually not statistics. It is that two markets quietly measured slightly different things.
How do you harmonize the questionnaire across countries?
Write a master questionnaire once, in one language, with each question’s purpose documented. Markets inherit it and do not rewrite it.
- Translate, then back-translate. One translator renders the questions into the local language. A second, who has not seen the original, translates them back. Compare, resolve gaps in meaning, then pilot with a few local respondents.
- Build local brand lists. Core brands appear everywhere so you can compare them. Add the local competitors that matter, each tagged to a global brand code so it rolls up correctly.
- Use scales that travel. Prefer simple, labeled scales and plain wording, without local idioms.
- Lock the definitions. Awareness, consideration and preference, and their bases, are defined once.
For the questions themselves, see brand tracking survey questions. For the metrics they feed, see brand health tracking metrics.
How do you set samples and quotas in each market?
Each market is its own tracker with its own quotas, usually on age, gender and region. The aim is the same kind of person everywhere, such as category buyers aged 18 and over.
Base size decides how much change you can read. These are approximate 95% margins of error on a result of 50% from a simple random sample:
| Respondents in market | Margin of error (points) |
|---|---|
| 100 | 9.8 |
| 300 | 5.7 |
| 500 | 4.4 |
Set a minimum base per market, and agree how to report a market that falls under it. Keep bases comparable too: if one market reports awareness among all adults and another among category buyers, the gap means nothing. The same logic applies to brand awareness tracking.
How do you combine markets into a regional or global total?
You weight each market, then add the markets up. There are two common ways, and they answer different questions.
- Market-size weighting counts each market in proportion to its size, such as its category buyers. The total says what share of buyers across these markets know your brand.
- Equal weighting counts every market the same. The total says how the brand does in the average market.
Here is a worked example with illustrative numbers. You track aided awareness in three markets with 500 respondents each. Category buyers are 60 million in Market A, 30 million in B and 10 million in C, so the size shares are 60%, 30% and 10%.
| Market | Sample | Awareness | Size share | Size-weighted contribution | Equal-weighted contribution |
|---|---|---|---|---|---|
| A | 500 | 70% | 60% | 42.0 | 23.3 |
| B | 500 | 50% | 30% | 15.0 | 16.7 |
| C | 500 | 30% | 10% | 3.0 | 10.0 |
| Total | 1,500 | 60.0% | 50.0% |
Size-weighted: 0.6 × 70 + 0.3 × 50 + 0.1 × 30 = 42 + 15 + 3 = 60.0%. Equal-weighted: (70 + 50 + 30) / 3 = 50.0%. The same fieldwork gives a 10-point difference, so state the basis on every chart. With equal samples, simply pooling all 1,500 interviews also gives 50.0%, which is equal weighting by accident.
Weighting has a cost. Under size weighting a respondent in Market A counts six times as much as one in Market C, so the effective sample size, (sum of weights) squared divided by the sum of squared weights, falls from 1,500 to about 1,087. The approximate 95% margin of error on the total is about 2.8 points under size weighting and 2.4 under equal weighting.
Pick the basis that fits the decision and hold it constant. If you update market sizes, do it on a documented schedule and restate the history.
How do you handle differences in how countries use scales?
In some markets respondents lean toward the top of a rating scale. In others they avoid the extremes. Acquiescence, the tendency to agree with a statement whatever it says, also varies by market. So a 7.8 in one country is not automatically better than a 7.1 in another.
You cannot remove the problem, but you can contain it:
- Compare trends within a market first. Same market, same scale, same method, wave after wave.
- Prefer categorical measures for cross-market gaps. Aided awareness is less exposed to scale style than a 0 to 10 rating.
- Compare relative position. Rank your brand against competitors within each market, which cancels much of the local scale effect.
How do you test wave-over-wave change by market?
Test inside each market, using its own bases, then look at the totals. Suppose the next wave gives awareness of 72% in Market A, 50% in B and 36% in C, with 500 respondents again. These numbers are illustrative, and the test assumes a simple random sample with no design effect.
| Market | Wave 1 | Wave 2 | Change | 95% threshold | Verdict |
|---|---|---|---|---|---|
| A | 70% | 72% | +2 | 5.6 | Within noise |
| B | 50% | 50% | 0 | 6.2 | No change |
| C | 30% | 36% | +6 | 5.8 | Just significant |
The totals tell a quieter story. The size-weighted total moves from 60.0% to 61.8%, which is +1.8 against a threshold of about 3.9. The equal-weighted total moves from 50.0% to 52.7%, which is +2.7 against about 3.4. Neither is significant.
The one real signal, Market C, is nearly invisible in the global number, so treat it as a lead to confirm next wave. Run enough tests and a few will cross the line by chance; for the logic, see statistical significance.
How do you govern and report a tracker across dozens of markets?
Governance is what keeps wave 20 comparable with wave 2.
- One data model. Every market delivers the same question IDs, answer codes, brand codes, wave labels and weights.
- Consistent nets. Top-2-box, aware and consider are defined once, centrally.
- Versioned questionnaire. Every change is logged, with a bridge wave or a break marker where it affects the trend.
- Access by market and team. Local teams see their market and the comparison set. Global teams see everything.
Reporting at scale is a production problem. With dozens of markets and hundreds of tables every wave, banners, tests and formats must be templated, so people review exceptions instead of rebuilding tables. That is also a tooling question, covered in how to choose brand tracking software.
How do you run global brand tracking in practice?
For global brand tracking in mTab, Halo Reports handles crosstabs, weighting and significance testing on wide global trackers. On top of that, a ranked view of what changed monitors your trackers and surfaces meaningful movement by market. You can also track brand health and competitors continuously across the markets you run.
What are the common mistakes?
- Switching weighting between waves. A move from size to equal weighting looks like a brand change. It is a method change.
- Translating once and trusting it. Without back-translation and a pilot, two markets can answer different questions.
- Letting the sample drift between markets. Different panels, quotas or bases by market make gaps unreadable.
- Ranking countries on raw scores. Scale style makes level comparisons risky.
- Reporting only the global number. It can hide a real move in one market.
- Changing wording in one market mid-series. Without a bridge, the trend breaks there.
What should you do next?
Take your current global brand tracking program and write down, for each market, the base, the quota, the weight and the definition of your top three metrics. Wherever two markets differ, decide whether the difference is deliberate and documented. That one-page audit usually finds the problems before the next wave does.