👨‍⚖️ KOL : Compliance Score

The launch of the Compliance Score follows the introduction of new laws in France that mean brands and communications professionals are now liable for illegal content posted by influencers that originated from a collaboration. Among its key provisions, the law sets down clear labeling rules for any paid or seeding-generated content.

The effect of these laws means that it is more important than ever to make informed decisions when choosing which influencers to partner with. That is why we have developed the Compliance Score, which allows you to assess at a glance the transparency of a KOL's sponsored posts, in respect of the French laws on influencer marketing.

After successfully testing our Compliance Score Feature in France, we have now
adapted the algorithm to German SRO recommendations for publicity and sponsored content.
As well as Spain with the EASA recommendations.
Italy with the AGCOM recommendations.
England with the ASA recommendations.

How it works

The algorithm developed by Kolsquare assigns a score ranging from "Very good" to "Very bad" to KOLs with over 30% of their audience in France or in Germany, and therefore subject to legislation or guidelines to follow, based on all their Instagram posts and Reels published in the past three months.

To determine the promotional nature of each content, the algorithm analyzes several semantic criteria, including:

• The presence of mentions of brand accounts, in comments or direct links,
• The presence of certain specific keywords, from a list of about a hundred terms, such as "sponso," "ad," "free product," and others,
• The use of collaboration or sponsorship features provided by Instagram.

Once the content is estimated to be sponsored, it is then subjected to an analysis of compliance with the French transparency laws or the German SRO recommendations. This analysis is based on two key elements:

1. The activation of Instagram's collaboration or sponsorship feature, or the presence of one of the two legally mandatory keywords, namely
   1. For France, "collaboration commerciale" or "publicité"
   2. For Germany, "werbung" or "anzeige"
   3. For Spain, "colaboración" or "en colaboración con"
   4. For Italy, "pubblicizzato da" or "in collaborazione con"
   5. For UK, "in collaboration with" or "Sponsored"
2. The mention of the brand's account, which must also be present.

By analyzing the compliance of each of a KOL's Instagram posts over the past three months, the algorithm then assigns an overall score based on the following compliance thresholds:

• "Very good" = 80% to 100% of sponsored content is compliant,
• "Good" = 60% to 80% of sponsored content is compliant,
• And so on, until "Very bad" = 0% to 20% of sponsored content is compliant.

In order to facilitate the interpretation of this score, we display two additional pieces of information in the KOL profile page:

1. The percentage of content (posts and Reels) evaluated by Kolsquare as sponsored compared to the total analyzed contents,
2. The average number of brand mentions appearing in all these contents.

The limitations of the algorithm and its evolution

In this first version, our algorithm has several limitations.

First, it is specifically designed for Instagram, focusing only on posts and Reels, excluding Stories for the moment. However, in a future update, we plan to extend its application to TikTok.

Second, it does not take into account where the keywords are (should preferably be written at the very beginning of the description). And it can sometimes make errors when evaluating a publication. These errors usually occur due to a misidentification of brand mentions or a complete absence of relevant keywords in the publication, resulting in the publication not being taken into account in the score (what we call a "false negative"). As a result, this can lead to a low percentage of sponsored content but with a high number of brand mentions.

However, since our main goal is to provide an overall estimation based on all a KOL's publications, these errors should be mitigated, and the calculated average score should reflect reality, provided there is enough content in the past three months. If you identify any errors when viewing KOL profiles, please feel free to report them to us. We will continue to feed and improve our algorithm to reduce the risk of errors.

Finally, it is important to note that France & Germany are not the only countries dealing with regulation of the influencer marketing, and we plan to soon expand our algorithm to cover other countries and languages in order to better take into account different regulations and practices.