For this reason, we conduct non-ordinary playtests and ad creative tests. The problem of any questionnaire is, participants’ observations are filtered through their subjectivity, and we can’t access their experience directly. Were they angry at the train hitting the character, or were they aroused by it? Were they sad reading the dialogue, or were they aggravated? It’s hard to answer these questions yourself, and we don’t expect study participants to have this level of self-awareness.
The subjectivity of answers сomplicates turning them into actionable data. Moreover, it’s harder to connect them to metrics. That’s why we aim to scientifically break down players' experiences using AI neurobiological tools and behavioral research methodology. This way, we leave no room for guessing and connect all dots between emotional perception and metrics.
In this article, we’ll explain how we utilize the power of AI and who will benefit from our services.
How we Track Emotional Responses
We utilize quantitative and qualitative research methods. Quantitative research includes a huge number of participants. It’s required if you’ve already got hypotheses to check on a statistically significant audience. This method is used mostly in ad creatives analyses.
Qualitative research, on the other hand, doesn’t provide big numbers, but it digs deeper into players' perceptions. Our researchers ask participants a wide range of questions, gathering not only their emotional responses on core mechanics and meta, but also feedback on interface usability. It’s needed to form hypotheses and is mostly used in playtests.
Now let’s move to the research itself. Everything starts as usual: we record user interaction with the object of study and then ask a series of questions.
Then, the magic happens. Apart from playtests, we armed ourselves with two powerful neurobiological AI research tools:
- facial coding
- eye tracking
Each tool follows its purpose. Eye-tracking lets us create heatmaps that follow where users are looking, defining their points of attention. Heatmaps are used in ad creatives & ASO analysis, where it’s important to ensure that players pay attention to certain details and are not disturbed by flickering popups.
Facial coding is the most powerful tool in our arsenal. It tracks emotional responses, showing us where players' attention weakens or sharpens, which enriches our analyses with valuable insights.
Are players intrigued by the new mechanic or simply unable to understand the flow? Are prizes in Battle Pass appealing enough to join the run? Are players engaged enough to buy a bundle on this stage of the game? Answers to these questions can be found in this research. Let’s dig a little deeper.
What Exactly do we Track, and How to Convert it to Relevant Metrics
In a nutshell, facial coding works like this: we observe how players perceive the object of study during an interaction, collect the participants' data, and build an arousal curve. Arousal is the level of emotional engagement of the audience during interaction with content.
Let’s split the process stage by stage.
First, we measure a regular player’s facial expression to find its neutral state. Then we analyze their reaction throughout their interaction with the object of study with face-scanning AI.
We recognize seven emotions: neutral, happy, surprise, anger, sad, fear, and disgust. Just like in "Inside Out."
These emotions we distill into the most important engagement indicators. Polarity of emotions isn’t of concern to us. Whether we managed to scare the player or please them, we hooked them on something; we kept their attention. As long as it’s engaging, every emotional response is good (if it wasn’t caused by a critical bug in interface making a player lose their progress).
Take Fall Guys' success, for instance. There are many frustrating moments out of players' control. Nevertheless, it hooks you on emotional contrast, drawing you into their core loop.
Base on highest peaks of those indicators we build an arousal curve. It helps to measure general level and distribution of engagement at different stages of interaction. When it drops lower than needed or rise extremely high, we examine the in-game reasons behind it and provide recommendations based on our observations.
Emotion tracking opens us up to many possibilities. Not only can we tell exactly how engaged they were during each second of interaction, but we can also tell whether they were frustrated, surprised, angry, or intrigued. For example, we examined LiveOps, in which a character was ultimately hit by a train. It seems like it should have aggravated players. But our data showed us that they were actually aroused by it, leaving no room for guessing.
How you can Benefit from Sensemitter
Our study provides us with clear, unambiguous data on players’ perceptions, which we wouldn’t be able to retrieve through participants’ answers. We can tell why conversion from tutorial fails, which game mechanic is more appealing, and which hooks in a first-play session engage players to stay. We can compare your mechanic with similar from your competitors and tell where exactly to look to improve its performance. And it’s only a tiny part of insights you might get.
Here we listed just several ways to utilize our service:
Eye tracking reports allow app publishers to find hotspots that lead to better engagement. With the help of heatmaps, developers can simplify navigation, choose the most appealing placement for products in the in-app shop, etc. Moreover, eye tracking paired with emotional feedback can tell publishers if their monetization strategy is clear and appealing.
- Improve the game’s mechanics and complexity levels to maintain optimal engagement, immersion, and interest.
- Clarify the scoring system, navigation, and communication of achievements.
- Test hypotheses for boosting user acquisition and retention.
- Build a research-based foundation for personalized in-app recommendations.
- Assess the most suitable times for offering paid level-ups and items.
If you're thinking about how our tests can be applied to your products, let us know. We can demonstrate how our technology works, discuss which types of tests will bring you the best results, and devise a test plan to concentrate on the most important metrics.