Since starting out in 2003, we have been hard at work solving the challenges of understanding social media. In the course of executing thousands of projects, we have developed a strong set of tools which together form the WaveMetrix methodology.

Our approach to social media monitoring was developed, bit by bit, with an ambitious goal in mind: to create one methodology which could be used to provide a deep and accurate understanding of what people say online, for any topic whatsoever in any market.

This section describes our methodology and its key features. 

The WaveMetrix methodology:

Comprehensive understanding of sentiment, engagement and advocacy for any topic

WaveMetrix uses one comprehensive methodology to ensure consistent coverage of all sites in all languages.

We ensure a deep understanding of online discussion by having mother tongue analysts read every comment and code sentiment using our action based scale (+2/-2), which is standardised to reduce bias.

We don't just measure what is posted, but also what is actually read: our 'impressions' metric is designed to capture engagement and advocacy, including explicit and implicit recommendations.

For any topic we customise the coding schemes to extract insights such as brand values, use cases, purchase drivers and recommendations.

 

The six principles of the WaveMetrix methodology:

  • 1. Comprehensiveness. We use one methodology to provide comprehensive coverage of all types of sites across all languages
  • 2. Understanding. Mother tongue analysts read and analyse every comment
  • 3. Sentiment. Our action based normative scale of +2/-2 eliminates bias from different languages or writing styles
  • 4. Topic. We use custom classification schemes developed for each of our clients to code buzz in any dimension. This includes capturing brand values, use cases, purchase drivers and recommendations
  • 5. Engagement. Using 'impressions', we measure readership rather than just authorship
  • 6. Advocacy. Recommendations drive sales, which is why we focus on measuring both explicit and implicit recommendations

Further reading on the WaveMetrix methodology: