Hey everyone, I recently built this FB Messenger bot to help me exercise more regularly. I found that I was pretty off-and-on-again for exercise and wanted to be a lot more consistent. I feel so much better when I exercise every day. I originally built this for just myself, but I was asked by a few other people if they could use it - so I went ahead and launched it publicly.
I know, it's pretty basic. With the amazing exercise apps and services out there, what use is a silly little bot like this? I can't really give you a solid answer - but I will say it's absolutely helped me exercise significantly more than I had before. The daily reminder in the morning gets me in the mindset of wanting to do something active each day.
Looks very nice and a nice idea :) you could expand the idea to other fields (like going over bills, etc.) Makes me think of a backward Tamagotchi where the computer takes care of the human :)
"backwards Tamagotchi where the computer takes care of the human" has got me down a deep dark hole of thought about computers taking care of humans now haha. Simple statement, but lots to unpack there.
Having "robots", whether chat bots or something else, look at your data and make suggestions on how to improve is something I find super interesting. One for your finances certainly is a n interesting idea.
Interesting that the chat dynamic worked out so much better for this use case than an app with notifications. In theory, you don't really need an interactive chat agent for any of these use cases, but I guess what's being posited (by you and the rapidly growing 'chat bot industry' at large) is that personifying a UX is more engaging than the alternatives.
Do you think this working better for yourself (and your friends, I take it) is a result of motivational agents being inherently more effective as chat bots than traditional push-based systems (SMS/emails/app notifications)?
So this the crux of this whole project. You nailed what is "interesting" here. This should, by no means, be more effective for anyone who's currently using it. This bot is pretty damn simple. But the personification is seemingly the reason it "works". I did actually spend a reasonable amount of time writing (what I think) are kind of goofy/silly/funny/motivational responses to things which I think has made this project fare much better than simple "You worked out? K thanks." style responses.
Example: My girlfriend joked recently, after starting to exercise a lot more after a short hiatus, that she was "rekindling her relationship with Jumper." I found that statement pretty deep - that's not something she would have said about a standard exercise app (I think?).
Thank you so much! I've been operating Jumper for ~4 months with a small group of people and the conclusion most people have made: It's useful because it's very lightweight and the personal style interaction helps with motivation more than say, a RunKeeper iOS push notification. Whether that conclusion holds true for lots of people, who knows! But it's sure been a fun project so far that's helped some people get into better shape. :D
Have you giving any thought of using a different platform? I like the idea of writing a chat bot myself that would work iOS built in Messages but I can't seem to think of a way around it short of using SMS and Twillo which would start to get expensive. Anybody have any suggestions?
I actually started out on SMS with Twilio! I switched to Facebook Messenger simply because it was starting to get expensive and I really started to miss the niceties of using a proper chat platform: thing like predictable message delivery, getting someone's first name without asking, and easy/free image embedding.
I'm going to try and do Google Hangouts and Slack next - unless someone has other good ideas for platforms to support?
One of the most annoying things about physical fitness is you can't really outsource it. In the end, you have to do the work, with a trainer or whatever. There's just no way around it short of perhaps steroids speeding things up.
For cardio, the most tolerable way I've found to do it everyday is to get a heart rate monitor, get a machine that can adjust difficulty based on a target heart rate and then do that for 30 minutes with headphones and a good podcast or audiobook on. I try to ignore everything but the podcast. Things seem to work OK that way.
I combine my meditation with my cardio. I try to focus on rhythmic breathing and not letting my thoughts wander (especially not to "intellectual" topics since that is my default mode all day).
I use Zwift as well! It makes indoor riding nearly fun! (so long as you have a decent bike trainer, of course).
Ultimately there are no magic bullets, you have to do the exercise. And that's the idea with Jumper - motivating you to get out there every single day and get some exercise.
I've just recently gotten into http://zwift.com/, it's the only thing I've found that makes riding indoors tollerable, almost fun. The power meter on the bike connects to zwift, the more power you put out, the faster you go. There are a few virtual locations you can ride around, ride with other people. There are also workouts, where you need to hold a certain power for certain time. The workouts are scaled based on the power you can do. I prefer working out with power than heart rate since it responds a lot faster and is less affected by other things going on in your life. Mind you, a bike, a trainer and a power meter is pretty expensive if you don't already have them.
I am having the same thoughts. As lazy engineers, we try to optimize and automate everything. Heck, even food is optimized with things like Soylent, but for workout and muscle building there just isn't anything yet. There are so many things I want to do in my day, and putting an hour away to go to the gym and do a few reps here and there is time that I would really love to invest into something else.
I'm wondering if it's really that difficult to 'invent' something to make it easier. Recalling my school biology lessons, muscle build because of ripped muscle fibers that grow back together, just stronger. It sounds so simple to build something artificial that does this automatically to your muscle zones without working out.
On the other note, maybe it's good that we don't have that yet. A good physique, big arms and things like visible abs are a great trophy you get for investing hard work and time. Not everyone has it, which makes it so desirable.
Jumper actually started out as SMS. It operated solely on SMS for about 2 months, but the unpredictable message delivery, high cost or messages (esp. images) and lack of access to user profiles made me switch.
Is there another chat platform you'd prefer to use? Slack? Google Hangouts? Telegram?
Not the person you're responding to, but aren't there service plans with unlimited texting and images? I'd definitely use this service if I could text.
I'd be fine paying for it to cover the cost of SMS. Like parent, I'm not willing to use facebook to get access, even for free.
One option would be to re-enable the SMS option for people who deposit money to cover the SMS fees. Mark them up 100% (so 2c per SMS or something) and you'll have a built-in profit margin.
Somewhat off topic - has anyone used a desk bicycle (pedals) or treadmill for a significant period of time? Being able to do that for an hour every workday would be an amazing health improvement basically for free (in terms of time), but I have difficulty imagining anybody sticking to it.
Yes, one of my co-workers had a spin desk. He said you couldn't really ride with a heart rate above 100 bpm and still concentrate. He was doing a huge amount of very low intensity riding, it didn't have much benefit, but hugely increased appetite.
Is the chatbot implementation intended as a prototype or the final product? I think it keeps the barrier to entry low and it's got a bit of fun novelty value to it, but the value in your (great) idea comes entirely from the fact that it's quick and painless to use, and it could be 10x quicker and less painful with virtually any interface besides a chatbot.
Mobile keyboards suck, and nothing highlights that more than seeing a cool idea that could deliver a good value to me by simply asking me to tap one or two large-ish buttons once per day, but that I instead need to chat with.
Edit: I realize there's a large personal preference component to this as well. I personally would rather have a button or two to tap, but I do see how the chatbot-powered anthropomorphization could be a very effective motivator for people who like that.
So the "friction" conversation about a product like this is pretty interesting imo.
One one hand, the bot is easy to get started with: Almost everyone already has the software (assuming Jumper works on multiple chat platforms) to get started right away with very little friction or commitment (no "signup" required).
On the other hand, once you are started - the chatbot interface is certainly higher friction than a traditional app would be (say, with two big buttons as you suggest).
But the while the app approach is lower friction once you've started, getting people to download and install an app is quite tough. Have you seen the latest data on how many apps the average American downloads in a month? It's zero!!! Crazy eh? So that barrier is one I wanted to avoid.
In the end though, the reason I chose the chat platform and the reason it is the final implementation for Jumper is that I've been super interested in the power of conversational interfaces and their ability to change human perception. Don't get me wrong - the "chatbot" craze is pretty exhausting to me too and I feel most chatbot products/services could have been built in a more traditional interface and have been far more successful. In fact, I built a failed chatbot for analytics not long ago. But when it comes to motivating people to exercise, I think having a conversation with an "exercise buddy" is more likely to get people to exercise. I mean, if I didn't believe that, I would have just used one of the currently available fitness apps and set up simple reminders. Of course, I may be very wrong here haha. But so far it's working!
Awesome, and good reasoning all around. This is indeed absolutely worth a shot; you may have found a really great fit for a chatbot. I'll be keeping an eye on this one, congrats on shipping!
If you're looking for useful features to build I would find a well being tracker very useful. Exercising isn't my problem. Injuries are. If the bot would manage my schedule so if my back is stiff it recommends certain exercises or if I haven't done cardio in a while it recommends that instead of weights etc it would be fantastic. With ML I can imagine being able to build a statistical model off the back of this that would be able to perfectly curate a program for optimal health based on frequency of exercise, type of exercise and reported well-being.
This sort of stuff is really interesting to me. I think any app/service can use ML though, it doesn't have to be a chatbot. I'm actually surprised we aren't already seeing more of this. I feel like most all fitness trackers are more concerned with the very specific details of your current workout (speed, heart rate, etc) but don't look at the wider trends across many days, weeks, months, etc.
If anyone has seen a service that does this, I'd love for you to share it here!
Interesting idea and execution. As mentioned above, I imagined a physical robot, not a chatbot, so might need some clarification in future messaging revisions.
Also interested to learn more about how you plan to utilise user data to make the platform smarter.
Awesome. I have been experimenting with Facebook Messenger bots as well. Did you use any existing framework/tool to create the bot?
Also, I am building a service that allows bots to gather actionable feedback from the users. I am looking for some testers. Let me know if you would be interested in helping out. :)
I am working on an app in a similar space; although as a bot is a novel idea I had not considered. But it looks like you're doing a lot of things right in creating an experience thats engaging and memorable. Plus you've got a great name!
Great job on that landing page too - crystal clear.
Thanks so much! Honestly, I think I'm more excited about the landing page than I am the chatbot at times haha. I built it using http://tachyons.io and I'm not sure I've ever had more fun writing HTML/CSS. Definitely check out that CSS framework.
47 comments
[ 3.0 ms ] story [ 107 ms ] threadI know, it's pretty basic. With the amazing exercise apps and services out there, what use is a silly little bot like this? I can't really give you a solid answer - but I will say it's absolutely helped me exercise significantly more than I had before. The daily reminder in the morning gets me in the mindset of wanting to do something active each day.
Let me know what you think. :)
Having "robots", whether chat bots or something else, look at your data and make suggestions on how to improve is something I find super interesting. One for your finances certainly is a n interesting idea.
Do you think this working better for yourself (and your friends, I take it) is a result of motivational agents being inherently more effective as chat bots than traditional push-based systems (SMS/emails/app notifications)?
Example: My girlfriend joked recently, after starting to exercise a lot more after a short hiatus, that she was "rekindling her relationship with Jumper." I found that statement pretty deep - that's not something she would have said about a standard exercise app (I think?).
It looks very useful. Congrats on shipping this.
Have you giving any thought of using a different platform? I like the idea of writing a chat bot myself that would work iOS built in Messages but I can't seem to think of a way around it short of using SMS and Twillo which would start to get expensive. Anybody have any suggestions?
I actually started out on SMS with Twilio! I switched to Facebook Messenger simply because it was starting to get expensive and I really started to miss the niceties of using a proper chat platform: thing like predictable message delivery, getting someone's first name without asking, and easy/free image embedding.
I'm going to try and do Google Hangouts and Slack next - unless someone has other good ideas for platforms to support?
For cardio, the most tolerable way I've found to do it everyday is to get a heart rate monitor, get a machine that can adjust difficulty based on a target heart rate and then do that for 30 minutes with headphones and a good podcast or audiobook on. I try to ignore everything but the podcast. Things seem to work OK that way.
Works really well.
Ultimately there are no magic bullets, you have to do the exercise. And that's the idea with Jumper - motivating you to get out there every single day and get some exercise.
https://support.zwift.com/hc/en-us/articles/203968635--Setti...
I'm wondering if it's really that difficult to 'invent' something to make it easier. Recalling my school biology lessons, muscle build because of ripped muscle fibers that grow back together, just stronger. It sounds so simple to build something artificial that does this automatically to your muscle zones without working out.
On the other note, maybe it's good that we don't have that yet. A good physique, big arms and things like visible abs are a great trophy you get for investing hard work and time. Not everyone has it, which makes it so desirable.
Is there another chat platform you'd prefer to use? Slack? Google Hangouts? Telegram?
Maybe another SMS service offers that? If so, I'd certainly consider SMS again.
One option would be to re-enable the SMS option for people who deposit money to cover the SMS fees. Mark them up 100% (so 2c per SMS or something) and you'll have a built-in profit margin.
Not everything has to be edgy. Can't you simply say "Great job!" or "congrats" ?
Is the chatbot implementation intended as a prototype or the final product? I think it keeps the barrier to entry low and it's got a bit of fun novelty value to it, but the value in your (great) idea comes entirely from the fact that it's quick and painless to use, and it could be 10x quicker and less painful with virtually any interface besides a chatbot.
Mobile keyboards suck, and nothing highlights that more than seeing a cool idea that could deliver a good value to me by simply asking me to tap one or two large-ish buttons once per day, but that I instead need to chat with.
Edit: I realize there's a large personal preference component to this as well. I personally would rather have a button or two to tap, but I do see how the chatbot-powered anthropomorphization could be a very effective motivator for people who like that.
One one hand, the bot is easy to get started with: Almost everyone already has the software (assuming Jumper works on multiple chat platforms) to get started right away with very little friction or commitment (no "signup" required).
On the other hand, once you are started - the chatbot interface is certainly higher friction than a traditional app would be (say, with two big buttons as you suggest).
But the while the app approach is lower friction once you've started, getting people to download and install an app is quite tough. Have you seen the latest data on how many apps the average American downloads in a month? It's zero!!! Crazy eh? So that barrier is one I wanted to avoid.
In the end though, the reason I chose the chat platform and the reason it is the final implementation for Jumper is that I've been super interested in the power of conversational interfaces and their ability to change human perception. Don't get me wrong - the "chatbot" craze is pretty exhausting to me too and I feel most chatbot products/services could have been built in a more traditional interface and have been far more successful. In fact, I built a failed chatbot for analytics not long ago. But when it comes to motivating people to exercise, I think having a conversation with an "exercise buddy" is more likely to get people to exercise. I mean, if I didn't believe that, I would have just used one of the currently available fitness apps and set up simple reminders. Of course, I may be very wrong here haha. But so far it's working!
If anyone has seen a service that does this, I'd love for you to share it here!
Also interested to learn more about how you plan to utilise user data to make the platform smarter.
Also, I am building a service that allows bots to gather actionable feedback from the users. I am looking for some testers. Let me know if you would be interested in helping out. :)
Great job on that landing page too - crystal clear.