When I was a kid, I wasn’t what you’d call … an athlete. 12-minute mile in gym class? Not a chance. First pick for dodgeball? Not a chance.
I was much better at dodging the undead on my Commodore 64.
It wasn’t until I entered adulthood that I began to take my health more seriously and, in doing so, establish more concrete fitness goals. Of course, like anyone with a fitness goal knows, turning that“goal” into a “routine” presents a “fun” new set of challenges.
My struggle: replacing happy hours in the bar with power hours in the gym. I needed help from a community that shared my goals, which wasn’t easy to come by circa 1999.
Luckily, several years and many hard-earned healthy habits later, I was able to channel that frame of mind to create Jolt.ai. Jolt is a health and fitness chatbot scientifically engineered to keep people on track with their fitness goals through a sequence of behavioral triggers.
That said, in the beginning, things weren’t always so … scientific.
Sometimes the Obvious, Isn’t
When first building Jolt, it was clear group competition was going to be an essential component of our motivational engine. Why? Because it made sense! Who hasn’t succumbed to peer pressure at least 1000 times in their adolescent (and, let’s get real, adult) lives? “They did X, so I need to do X even better.”
So, we launched Jolt Teams. Teams consisted of groups of like-minded strangers with similar goals who would compete on weekly leaderboards. The goal of these leaderboards was to earn more MovePoints (based generally on calories burned from an activity) than other members on your team. The users who engaged with this feature were hooked. But for important segment of our user base, engagement was falling flat.
Our theory? We didn’t know our users as well as we thought we did.
Here was our mistake: We built Teams without ever having real, thoughtful conversations with our users. Call it laziness. Call it hubris. Call it anything you want. The point is, we had to change course and the only way to correct our mistake was to start taking to people, and fast.
Let’s just say, these conversations led to a lot of “duh” moments. But there is something remarkable that comes out of being able to admit that. We often work in silos and, the deeper we get into those silos, the more committed we get to our ideas, blinding us to the obvious. By having the humility to say, “we should have seen that,” we were able lift our blinders and hear what our users had to say in their purest form.
One particularly important “duh” moment was realizing not everyone liked the idea of competing with strangers. A large cohort of our users, it turned out, preferred a more collaborative environment that consisted mainly of friends. To be fair, we had toyed with this idea prior to building teams, but the technical investment to build it was large. So, we didn’t. (Hindsight, right?)
Because we’re overachievers, we decided to take this period of self-reflection even deeper and shared these conversations with actual behavioral scientists to understand the “why” behind this feedback. Here’s what we learned:
There are three different types of motivational orientation that exist when people try to accomplish their goals.
- Task Orientation
If you’re the kind of person who loves a good checklist, this orientation often presents as a desire to learn new skills, master challenges, or achieve self improvement compared to past performance.
2. Performance Orientation
If you’re a competitive type, this can be recognized as the desire to win and beat others, measure success against others, or demonstrate superior ability.
3. Social Orientation
For the more collaborative among us, social orientation is driven by establishing an affiliation to a group, desiring social connection, or gaining approval/acceptance from others.
Most people will not fall squarely into one box, but as we dove deeper, it became clear the majority of those who were dissatisfied with Teams fell more into the social orientation category.
To that point, we had framed so much of the product around task orientation (Want a personal progress report or challenge? We have you covered!) with a touch of performance via Teams, but very little catered to the collaborative, socially oriented of the bunch.
Teams, though attempting to appeal to this social need, were falling flat because they remained highly individualistic. Members still focused on their own personal goals and had limited communication with others on the Team, which made it particularly challenging to form meaningful bonds.
We had to reorient our product to better serve the social needs of our users. So, that’s what we did.
A Bite of Humble Pie
Learning we were out of touch with our users when we had so confidently built a product based on what we (real people!) knew in our bones would work … sucked. But it was important for us to let go of ego and realize we needed to be more rigorous in the way we test and release features going forward.
Sometimes, humble pie is bitter. Sometimes, it’s sweet. How it tasted was up to us.
To start, we decided to incorporate findings from user interviews to build a collaborative group feature we call “Co-op.” This allowed users to tap into deeper motivations of social orientation. Here’s how we did it:
- Collaboration. Instead of living in silos of personal accomplishment, we built a way for users to join what we call “Squads.” Like Teams, Squads consist of other Jolt users who have similar goals. The difference here is, although each person has their own personal goal, they are now also accountable to a collective mission (e.g., achieve “x” MovePoints in one week). One person’s failure can lead to the failure of the entire Squad. One person’s overwhelming success can also carry the Squad to a glorious finish. That said, it’s up to the collective Squad to ensure a successful end result.
- Communication. Prior to building this feature, users had limited ability to communicate with like-minded individuals in Jolt. We needed new avenues for people to build sincere bonds that ensure lasting social connection. By joining a Squad, users can now monitor the progress of others and respond to their failures and successes in real time. We also decided to build an avenue for users to invite friends to join their Squad for additional support. This last part, though more challenging to build, is one of those essential missing pieces we collected from our user feedback. It simply came up too often to be ignored.
- Rewards. Though for some users, simple acknowledgement of success (or fear of failure) can be enough of a motivator to keep the ball moving. For others, we found the most effective mechanism to get the entire Squad to rally behind a collective cause was to offer a shared reward for collective achievement of a group goal. We are still experimenting with what kind of reward will produce the greatest success, but it’s clear working toward a tangible reward energizes Squads in a way reward-free collaboration does not.
By listening to our users and deepening our knowledge of group psychology, we have been able to introduce a feature that people not only love, but, so far, use at a higher rate than we were ever able to achieve with standard Teams.
Don’t Let Me Down, Bro
While exploring the mechanics of how Co-op would work, we encountered research on a little thing called “cohesion.” I know … we all know what this word means, and it should come as no surprise that cohesion of varying degrees is always present in groups.
What we found interesting was a study pioneered by a group of researchers led by the late Albert Caron.
They defined cohesion as:
“a dynamic process which is reflected in the tendency for a group to stick together and remain united in the pursuit of its instrumental objectives and/or the satisfaction of member affective needs” (Carron, Brawley, & Widmeyer, 1998, p. 213)
Something that stuck out in this definition was the word dynamic.
From personal experience, I knew how easy it was for small changes (e.g., group size, group leadership, etc.) to alter the cohesion of a group. Ever play a competitive team sport without an even number of players on each side? Great way to test the strength of your team and potentially lose an eye.
For that reason, it made sense to view group cohesion as something multidimensional and complex.
Source: Gill, Williams, & Reifsteck, 2017
In the figure above, you can see there are two factors of cohesion: task and social.
Each of these factors are then broken down into two perspectives: group integration (the overall unity of the group) and individual attraction to the group (the motives, interactions, and emotions of each individual toward the group).
Depending on which camp your group falls into, cohesion can take on different forms. For the more task-oriented groups, it is essential for all members to manage their individual tasks and not create undue burden on the rest of the group. If one member of the team falls out of line with a task, unity is compromised and creates a ripple effect that can negatively impact the cohesion of the group.
For more socially oriented groups, it is essential for members to form strong relationships. They must have shared interests and participate first as members of the group and second as individuals. If individual members begin to make choices that do not reflect the interests of the collective group, it creates animosity that weakens the social fibers essential to this form of cohesion.
Ideally, we wanted to frame our Co-op Squads in a way that served the needs of both task and social cohesion. Part of doing so was determining how large each Squad should be. How do we connect enough people to create an interesting social experience while eliminating the risk of “too many cooks” in the Co-op kitchen? Here’s how we figured that out.
That’s right. And smaller, when it comes to group cohesion, is better (Ewys & Brawley, 2018). Why? It all comes down to productivity. Too big, and productivity takes a nosedive. Too small and, well, you’ve barely gotten started.
Source: Carron & Hausenblas, 1998
Based on this fact and additional research, we decided to cap our Squads at 3–8 people. This is our sweet spot that allows members to hold each other accountable to weekly tasks while enabling a stronger sense of social value among individuals in the Squad.
Co-op Squads are still relatively new, so the jury is still out on whether or not we hit the nail on the head. However, I’m more confident than ever that we are on the right path. Why? Because we took the time to ask questions. Our users told us what they wanted and we listened. Nothing can ever substitute hard data to confirm whether or not a course of action was correct, but I feel confident knowing we based our development on real qualitative analysis.
So, what now?
Based on past experiences, we know how important it will be to not simply accept this research as fact, but continually monitor user behavior and collect quantitative data to either confirm or deny our hypotheses. So far, we are pleased with the results and look forward to what is to come with a user base that is more connected to our product.
But don’t take it from me. You can try the Jolt Co-op experience for yourself here. Go ahead and set a goal, give our pal Jolt some time to observe your behavior and place you in a Squad, and let me know what you think. I promise, in this game of dodgeball, no one is ever last pick.