When you are a leading global retail brand, you are always looking for new insights on different markets in order to keep up with the local trends, interests, and cultures of your target audiences. And when you have to move fast, which is almost always, your brand doesn’t have the time to rely on traditional market research studies. Audience analysis is the solution that delivers.

Our client, the digital marketing team of a retail industry leader, wanted insights on yoga lovers in Singapore to test the efficiency of its tactical marketing actions. They used Profiler, Synthesio’s audience analysis tool, to better understand who their targets really are: what they like, how they think and behave. In just a couple of hours, they were able to get powerful insights.

Let’s deep-dive into how they used Profiler and what they learned. 

Step 1: An unrivaled audience sample size

They wanted to know more about yoga lovers in Singapore. It could seem like a tough challenge to find a representative sample. Fortunately, they created a study in Profiler allowing them to easily analyze a statistically significant sample of their target audience.

The brand used the following criteria to identify its target audience: 

  • Location: Singapore
  • Age: 18 – 65+
  • Interest: Yoga

They observed that Profiler could analyze 1,300,000 people that are interested in yoga in Singapore, a sample size that no other market research methodology can offer. Moreover, consider the fact that absolute percentages are not enough to assess whether a result is significant, and therefore actionable. An advanced audience analysis tool, therefore, should enable you to compare your target audience with a reference audience. Profiler does that.


For example, let’s say that 16% of your target audience has an interest in the brand Huawei. What does that information tell you? On its own, almost nothing, as you do not have a point of comparison to know if this percentage is high or low.

However, what if you also knew that 25% of your reference audience has an interest in Huawei? With this information, you can conclude that your potential customers have a low affinity for Huawei.

Step 2: Choosing the right reference audience

Again, when analyzing data, it is not enough to just look at absolute percentages. You need to have a reference, which is a way of understanding whether these percentages are significant or not.

In our example, you can see that our client chose the overall Singapore population as a reference audience. This reference contains more than 4,500,000 people, including people in the target audience, making it possible to draw comparisons. 

At first, our clients wanted to compare yoga lovers in Singapore (i.e. target audience) to those in Japan (i.e. reference audience). Methodologically speaking, that did not make sense. People’s affinities to interests can only be evaluated as significant or not when compared to the affinities of their peers. For each point of interest, Profiler shows if the target audience has a positive or negative affinity compared to the reference audience. If the reference audience does not include the target audience, Profiler cannot measure the affinity differences. Therefore, the target audience always has to be included in the reference audience.

Note: Please note that our clients first decided to launch a basic study. They then decided to link their Facebook Business Manager Account to Profiler. This enabled the team to build more complex audiences and launch studies. Should you want to know more about that, ask your questions here.

Step 3: Using Audience Analysis to Precisely Quantify Affinity

Beyond just looking at an interest in a brand, our client wanted to align their campaign messaging with a persona for their well-being line of products. Our client reasoned that a promising persona that might be interested in buying these products were people interested in yoga in Singapore. The next step was to determine their interests using audience analysis.

The team wanted to learn more about how Profiler can detect that people are interested in yoga. Profiler pulls data from Facebook suite (Facebook, Messenger, WhatsApp & Instagram), the largest data set in the world. What Facebook knows about its users is beyond what you could imagine. Thus, having an interest in yoga does not necessarily mean that you’ve liked a yoga page on Facebook. It may come from behaviors like reading articles about yoga positions, using a yoga app, or frequently visiting yoga studios. Facebook is able to track these aspects to classify user interests.

In the overview section of Profiler, we display the highest represented interests to draw a typical target audience profile.


The most valuable interests are those with the highest affinity score, represented by the blue points you can see above. This affinity score is a ratio between the shares of your target audience having an affinity with an interest compared to the percentage in the reference audience.

The more an interest or hobby is over-represented in your target audience, the higher the affinity score will be.


In the example above, our clients were able to find that yoga lovers in Singapore have a high affinity towards environmentalism.

Thanks to the typical target audience profile, they got a first glance of what their target audience looks like.

Step 4: Browse the results with the sunburst

They then used the sunburst data visualization to deep-dive into the different categories of interests: Brands, Media, Behaviors, People, Hobbies & Activities, as well as Arts & Entertainment.


Despite being very sporty and outgoing, audience analysis also discovered that the group is also concerned with their appearance, as you can see that yoga lovers in Singapore have the highest affinities with clothing and cosmetics brands. In campaign messaging, the brand can strategically mention the impact of athleticism on appearance.

As they were planning to launch test campaigns, they wanted to know which media outlets to avoid. The brand displayed the categories by negative affinity to find the platforms its audience has the least affinity with. It was surprising for them to observe that their audience has a negative affinity with Pinterest, as their audience was largely female, as is Pinterest’s user base. Before analyzing its Profiler study, the brand had considered a campaign launch on Pinterest.


Step 5: Measuring Results from Audience Analysis

The test campaigns they launched based on their Profiler insights were a success. Within a month, the team saw a twofold sales increase, as well as a cost per reach that was 1.5 times lower than expected. Thanks to Profiler, the team was able to discover many other interests to refine their knowledge about their audience. Convinced the tool would increase ROI, they decided to keep the tool to enrich their knowledge of their different target audiences.

They also built precise reporting in the Reports section of the tool around content creation and media planning. Should you want to know more about the Reports section, feel free to request a customized demo. We will be happy to show you what it looks like.