5 Ways to Improve Data Quality in Brand Tracking: Latana

Table of Contents

Introduction: Why Data Quality Matters

For many companies, brand is one of their most valuable assets. Yet accurately measuring whether marketing changes consumer perception remains difficult. Brand tracking tools are widely used, but poor data quality can lead to unrealistic estimates, unexplained fluctuations, large margins of error, and ultimately misguided business decisions.

While perfect measurement is impossible, five key improvements can make brand tracking significantly more reliable.

1. Avoid Survey Fraud

Online research panels typically reward respondents with cash or gift cards. While incentives increase participation, they also attract people who are motivated by rewards rather than honest feedback.

Common problems include speeding through surveys, straightlining answers, duplicate registrations, false qualification responses, and automated survey bots. Latana’s internal research found that around 20% of incentivized responses were fraudulent, with some industry estimates even higher.

A better approach is to design surveys people are willing to answer without incentives. Questions should be simple, quick, and easy to understand. For example, each question should take only a few seconds to answer, require no scrolling, and avoid unnecessarily long answer lists.

When surveys are designed around the respondent experience, people can engage voluntarily, reducing fraud while improving authenticity.

2. Reach Everyday People

Traditional research panels often represent only a tiny share of the population. A relatively small group of “professional respondents” may complete many surveys every year, creating self-selection bias and demographic skews.

This becomes especially problematic in smaller markets, where limited panel sizes make it difficult to recruit fresh and representative samples.

One alternative is interactive ad-based sampling. Survey questions are embedded directly into digital ads across mobile apps and websites, allowing people to respond without registering for a panel.

This makes it possible to reach a much broader population, including groups that are traditionally difficult to recruit, such as younger adults, seniors, and high-income audiences.

3. Keep Sample Composition Stable

Brand tracking depends on consistent sample composition over time. If the proportion of different demographic groups change between waves, KPIs can move even when real consumer attitudes have not.

Traditional quota systems try to control this, but become expensive and complex when many variables need to be considered.

A more flexible approach combines answer-based quotas with Multilevel Regression and Poststratification (MRP). Instead of creating rigid quota cells in advance, the system gathers sufficient answers for relevant audience characteristics and statistically adjusts estimates to reflect population demographics.

This allows steady daily data collection while improving stability and reducing artificial fluctuations between waves.

4. Avoid List-Based Bias

Brand tracking surveys often ask respondents to choose familiar brands from a list. However, this can distort results.

Awareness may depend on which other brands appear in the list, and the number of options shown can also influence selection rates. Adding or removing brands may therefore create artificial changes in historical data.

A stronger approach is to present each brand individually. This “siloed” format avoids list-based bias, allows brands and statements to be added or removed without disrupting historical comparisons, and improves the overall survey experience.

5. Maintain End-to-End Control

Large or complex tracking programs often rely on multiple sample providers and subcontractors. Different markets may also use different methodologies.

This fragmentation can create inconsistent quality standards, different methodological biases, and difficulties when investigating suspicious data.

Centralized ad-based sampling provides greater control over the full respondent journey. It enables consistent methodology across markets, direct visibility into data collection, access to a large respondent pool, and reduced reliance on subcontractors.

Conclusion

Improving brand tracking data quality requires more than simply increasing sample size. The strongest systems rethink how respondents are recruited, how surveys are designed, how samples are stabilized, and how data collection is managed.

The five key improvements are:

Remove incentives to reduce fraud.
Use ad-based sampling to reach everyday consumers.
Stabilize samples through answer quotas and Bayesian modeling.
Replace list-based questions with siloed brand questions.
Maintain direct control over data collection.

Together, these approaches can reduce bias, improve stability, and give organizations greater confidence in the brand metrics they use to guide growth decisions.

Continue the Brand Tracking Conversation at MRMW Europe 2026

How can these principles be put into practice? Join Dr. Nico Jaspers, Founder and CEO of Latana, at MRMW Europe 2026 for “How to Overcome the 5 Biggest Challenges in Brand Tracking”.

The session will explore non-incentivized sampling to reduce survey fraud, modular survey design to adapt to new KPIs and segments, and AI-driven analytics to turn complex tracking data into actionable insights.

📅 MRMW Europe 2026 | October 21–22 | Berlin, Germany
Explore the event →

Discover more sessions at MRMW Europe 2026: