How our Chatbot Leverages Goal Setting Research to Help You Be More Active
As humans, we respond well to fresh starts and new challenges. Whether it’s starting a new job, taking part in a 30-day fitness challenge or putting a plan in place to read more, we begin with the best intentions.
Yet why is it we often fail to achieve the goals that we set ourselves?
The difficulty lies in setting the right kind of goals: ones that provide challenge, motivation, direction and that we can persevere through to the end.
Researchers often refer to this phenomenon as the intention-behavior gap, meaning that there is often a mismatch between the intentions we set and our actual behaviors.
So, in the context of exercise, if our current fitness habits/routine fail to match what we want to accomplish, that’s why we fail.
This is why goal setting is crucial. While it’s not the only factor in sustaining an exercise commitment, it does help us progress from where we are to where we want to be.
Top Articles on How Businesses are using Bots:
1. Higher Education Chatbot: Chatbots Are the Future of Higher Education
2. Chatbot for Healthcare: Chatbots Can Be Money-Savers for Hospitals and Clinics
Breaking Down the Theory Behind Goal Setting
When thinking about goal-setting, Edwin Locke and Gary Latham’s Goal Setting Theory (1990) is a great place to start. Their research was developed from nearly 400 studies involving 40,000 participants, in 8 different countries, performing 88 different tasks. Goals were either self-set, prescribed or a set collaboratively in a group lasting from 1 minute to 3 years.
In other words, it’s some pretty comprehensive research!
The overarching finding from their theory was that specific, challenging goals lead to increased performance opposed to no goals or vague goals (e.g., “do your best”).
This may not be a big shock to those familiar with goal setting but as I dug into their theory, I found the key components of their theory interesting.
Specifically, Locke and Latham identified several factors that explain the positive effect of goal setting on performance, including:
- Choice
- Effort
- Persistence
- Task strategy
- Ability
- Feedback
- Commitment
- Task complexity
- Situational constraints
This brings us to our motivational chatbot, Jolt.ai. Jolt was created to help people accomplish their fitness goals and to do so, we put this much of this theory into action.
For example, every week, Jolt does the following:
- Encourages users to commit to a goal that is both specific and challenging (commitment)
- Algorithmically creates a goal matched to the users’ level of activity (ability)
- Allows users to tweak the goal if they choose (choice)
- Ensures the goal is simple enough for them to understand (task complexity)
- Provide progress updates throughout the week (feedback)
- Allows the user to tune the the feedback, making the notifications easier to integrate into daily life (situational constraints)
Isn’t this this same as a“SMART” goal?
If you’ve made it this far in this post, you’re probably familiar with other goal setting frameworks. SMART (i.e., Specific, Measurable, Achievable, Realistic, Time-bound) is among the most popular.
But the majority of the research around SMART and other frameworks emerged from organizational and industrial settings. The practice of exercise, sport and physical activity are somewhat different.
Specifically, over the years, researchers in sport and exercise have classified three types of goal:
- Outcome goals (the overall outcome that you want to achieve)
- Performance goals (a result specific to performance)
- Process goals (the individual behaviors that happen during the performance)
The relationship between processes and outcomes is interesting.
Researchers suggest that short-term, performance-based process goals boost self-confidence and increase the likelihood of sustaining exercise over time.
Additionally, behavior change experts argue that building on small successes is the key to achieving larger, more ambitious goals.
Other sport and exercise psychologists have likened it to the process of climbing a mountain; it can seem daunting when you first see it from below. However, by setting short-term targets that increase as you go, what was at first seemingly unattainable suddenly becomes achievable.
Again we have implemented these concepts into Jolt.ai, in way that leverages this research but is approachable.
For example, one of the first things Jolt asks users is “What is your fitness goal?”
Options the user can chose from include “stay consistent”, “lose weight”, or “build muscle”, i.e. outcome goals. Jolt will then set a weekly process goal based around a simple points system that is an equation of Intensity x Time — the more intense the movement and the longer the period of activity, the more points earned.
Additionally, Jolt regularly shows you how you’ve performed vs. prior performances. This an example of a performance goal: How you performed on a given day relative to your performance on the same day last week.
Going to the next level: Adaptive vs. Static Goals
The idea of a goal that changes based on performance is an emerging concept within behavioral technology and is something researchers refer to as adaptive goal setting.
The more common approach modeled by most fitness trackers is static goal setting — this is when a goal is fixed and doesn’t change (e.g., 10,000 steps), or that increases from an initial baseline measurement (e.g., 8,000 steps + 500).
However, we think there’s a better way.
Let’s look at some research examples that have tested static and adaptive goals using technology.
The first study, published in 2013 in PLOS ONE compared static step goals (10,000 steps per day) vs. adaptive step goals over a six month period.
During these six months, twenty (20) inactive adults were randomly placed in to either a (1) static goal or (2) adaptive goal group. Both groups wore pedometers to track progress and before the testing, participants used the tracker for ten days, providing the researchers with baselines for participants steps per day.
Those in the adaptive goal group had their daily goal updated regularly based on how well they were doing. Those in the static goal group held the same step goal throughout the testing. Prompts, feedback messages, and financial rewards were used to encourage physical activity.
The result? The static group increased their step count by ~1,600 per day, while the adaptive group increased their steps by ~2,700 per day. The difference between groups was a significant 1,100 steps per day.
In other words: Goals that adapt regularly to performance can help to improve overall results.
Another study from the University of California tested the efficacy of a new fitness app, CalFit, over 10 weeks. The app combines a personalized goal-setting algorithm with self-monitoring and feedback to provide an adaptive exercise experience.
Researchers recruited 13 college students who wanted to become more physically active and placed them into a control or adaptive group. During the first week of using the app, all participants received the same daily step goal; this acted as a baseline measurement for the rest of the study.
After the first week, the control group received a goal of 10,000 steps and the adaptive group was given algorithmically-generated adaptive goals.
Results demonstrated that the static goal control group actually decreased their step count by 1,520 steps per day between baseline and 10-weeks, compared to the adaptive algorithm group who increased their step count by 700 per day. The overall difference between the two groups was significant at 2,220 steps per day.
So what can we conclude from this research?
It seems that personalized, adaptive goals result in more success than static goals.
Naturally, there are also likely to be several other factors in play. For example, despite the increase in physical activity across the studies, it’s reasonable to assume that daily prompts, self-monitoring techniques, feedback and rewards also helped participants in achieving their daily step goals. And of course the goals outlined above are step-related as opposed to calories burned, minutes active or another, varied measurements.
Nonetheless, these studies have influenced why we’ve built Jolt.ai the way we have. While users of Jolt don’t receive an adaptive goal every day, they do receive a weekly adaptive goal that is based performance of the previous week.
If a user had a bad week, their goal the following week will be lower making it more achievable and helping to build up confidence again.
So far, this technique is proving to help people regularly achieve their weekly process goals and, based on our weight loss study conducted with the training studio F45, also helping people accomplish their overall outcome goal.
Thanks for reading! If you enjoyed this write up, I’d appreciate it if you hit that clap👏 button a few times ( 👏👏👏👏) to share our insights :)
