We launched Flex, our new series of vibrators, to learn and share more about what turns people on — in this case, specifically around vibration patterns. We know that there’s a lot of diversity in individual sensation preferences (which we applaud!) but it can be hard to collect both quantitative and qualitative research data to help create products to better meet those desires.
With Flex, early project participants had access to an online application that allowed them to customize vibration patterns and download them directly to their vibrators. At the end of their exploration, we asked them to submit their favorite vibe pattern and complete a survey, giving us insight on why that pattern made them tick. We were able to collect data on how many times a person had altered their vibes, how each pattern changed from the factory default settings, and the score for each new setting. We are working with these results to learn more about what people want, and while it’s still early in the project, we wanted to share a bit of what we’ve seen so far.
The constant/steady mode (pictured above) is a staple pattern for any vibrator. When it comes to programming this pattern, it’s pretty simple because there is only one variable that can change: power. After filtering all the constant pattern submissions based on a grading system we put in place, we charted power based on the frequency it was submitted as a favorite. One surprise for us was how many people loved having a very low setting! Some people responded that they enjoyed this very low setting as a teaser. Our lowest factory setting was set at 25% power, yet we are seeing a trend for having an even lower setting. Also, notice how many people loved the highest power setting!
Because the other settings all have multiple variables, such as power, speed and wave amplitude, it’s much more complicated to analyze the results. The pulse pattern, for example, has two variables: power and speed. To make the data for these patterns easier to read, we used a statistical technique called “k-means clustering” to organize and separate the data into distinct buckets. Each bucket contains the subset of data that most closely resembles the other data points within this group to help identify patterns. When you plot the results, it looks something like this:
After filtering the pulse pattern and running our k-means cluster analysis on this subset of data, we get some clearly defined pattern clusters. The two most populated and dense clusters are represented in the blue diamonds (high power with fast pulsing speed) and purple X marks (lower power with fast pulsing speed). The other two clusters are less tight: we can see that the green cluster, while consistently is in the lower speeds, does not clearly represent just one power setting, and the red cluster represents higher speeds at a broad range of power settings.
Fascinating, right? We are still very much in progress on our data project, and we’ll be using these learnings to program our “Greatest Hits” Flex vibrators, which are still available through our crowdsourcing project until the end of the month. While we don’t intend to proclaim one universal “right answer” as to what turns people on, we’re excited to see how this research helps us make better vibrators going forward.
Christine Concho is the lead engineer on the Flex project.