A simple guide to understand Synetiq’s engagement metric
At Synetiq we use biometric sensors to detect and understand people’s emotional reactions. Based on the data, we create our metrics to ‘interpret’ these biological changes to understandable and actionable insights. In the series Measuring emotions, we will give insight into the scientific background of each of our metrics.
“The ad was engaging” - But what does that mean?
Our Engagement metric shows people’s level of involvement in the story they are seeing, or, in some cases, the amount of cognitive effort it takes to understand it. Although it is possible to measure narrative engagement with different methodologies, heart rate variability is the most commonly used biometric indicator. It complements our other biometric data sources (Approach, Excitement and eye tracking data) well, and together they create patterns which allow us to dig deep into the complexities of human perception.
As studies have shown, changes in heart rate variability correspond to self-declared levels of engagement and cognitive effort, so we observe heart rate variability in order to create this metric.
In this article you will learn:
What is the connection between heart rate and engagement?
How do we measure heart rate variability?
How do we visualize and interpret heart rate data? - example
1. What is the connection between heart rate and engagement?
The human heart is not a metronome. In fact, a healthy heart exhibits complex patterns of variability in order to allow us to react to anything that happens in our environment. Heart rate is an aggregated metric that only tells us how many beats our hearts pump out in a minute, and does not reveal sufficient information about changes. As so, it does not give enough information about the subtle changes that occur in response to stimuli. That is why we need to use a finer measure to observe this variability, and that is where the link between our heartbeats and our attention level comes into the equation. We observe the time that goes by between consecutive beats, on a millisecond level. In essence, we do not simply measure heart rate, we measure how fast it changes.
Let’s look at an easy example. Suppose that Bob is happily walking down the street, the sun is shining, his life is good. His heart rate is stable, not too fast. Imagine that a big dog appears out of the blue, and there is no fence between the dog and Bob. Bob’s heart rate accelerates to a stable, fast pace in order for him to rapidly react to this stimulus, the dog, aka run away. We at Synetiq try to pinpoint the moment he sees the dog and makes the decision to run, measuring the acceleration itself instead of the heart rate during walking or running from the dog.
Although our reactions are not so dramatic and intense for ads, we can still measure these small changes with our sensors. When the change is steady, that usually means high attention level - like in the case of our Bob above, his heart rate’s acceleration upon seeing the dog was a steady, controlled, but fast process, which allowed him to react to the danger. When the change is not steady - accelerations and slow-downs follow each other in a random pattern -, that means that the person’s attention is wandering, nothing is actively engaging them.
2. How do we measure heart rate variability?
We at Synetiq conduct standardized tests in a controlled environment. The controlled environment is crucial to erase most of the possible environmental variables, so we can distinctively measure viewers’ reactions to the stimuli we provide (in most cases, videos and ads).
Having a regular blood pressure monitor attached to viewers while they watch ads might be a bit distracting though, so we use a more creative and less invasive way to measure changes in heart rate variability. We attach a sensor to the tester’s earlobe, which is equipped with a small LED light. The LED light shines through the earlobe, illuminating the capillaries found there and it measures the changes in the reflecting light. Based on the bloodflow in those capillaries, the sensor is able to measure the variability of the heart rate.
3. How do we visualize and interpret heart rate data? - example
In order to be able to interpret audience reactions, the raw data from the sensor goes through a complex processing pipeline, which cleans, aggregates and visualizes it for analysis. Visualization is designed to be intuitively understandable: a baseline is constructed from the neutral state of the audience, and upward differences show understanding and attention, downward differences show cognitive load and boredom. Let’s see it in action!
As we said above, unsteady changes of heart rate imply that nothing is actively engaging the audience. For an easy to understand example, let’s see what happens if an ad’s visuals are so dark that it is hard to see the details. Look at this ad set in the dark Norse town of Tromso.
Dark scenes in commercial
Want to learn more how dark visuals influence our emotional reaction? Check our previous article!
Loss of attention, cognitive load of 18-59 y.o. general audience to dark scenes in commercial
Our visualization shows that upon seeing the dark scenes, viewers struggled to find focus - look at how scattered the eyetracker’s heatmap is on the scene! Consequently, their Engagement level started to decline. No particular thing really stood out from the darkness, so the visuals were hard to understand. This also means that viewers did not have one particular thing that caught their attention.
On the opposite side, consistent heart rate variability signals stable changes - e.g. steadily increasing heart rate, think about Bob and the dog. To illustrate this, let’s look at an action-packed horseback chase scene from a commercial!
Horseback chase in a commercial
Increase of attention, understanding of 18-59 y.o. general audience to action scenes in a commercial
As you can see from the clip, viewers had no problems following the protagonist and his horseback adventure. The movie-like scene with the linear story was easy to understand, and it was interesting enough to grab attention. Our visualization shows that viewers were Engaged with the content they were seeing, they were interested in the outcome.
All in all, many things influence understanding and the level of engagement, even technicalities like lightning differences can push viewers from the edge of their seats to the edge of sleeping. If you want to know more about the factors that play an important role in making engaging video content, follow us and get in touch with our experts.
written by Szimonetta Ézsiás-Nagy
If you liked this short explanation on how we calculate insightful metrics from biometric data, stay tuned for the next articles and subscribe to our blog!
