How Storytelling Took Our Bot From 0–100 (messages) Real Quick.
“Marketing is no longer about the stuff that you make, but about the stories you tell.” — Seth Godin
You’ve probably heard the hype about bots.
The idea of a perfect conversational interface is irresistible, and the biggest of big companies are putting serious money into it.
The ability to cut away all the noise and develop a simple and genuine 1 to 1 communication channel with our users is something like a dream.
Well it still is a dream as our toddling AIs still need a lot of work before they can truly understand what we seek from them.
It all begs the question what are users, people really looking for?
In our journey to find these answers we’ve been testing and iterating on bot concepts that help people in three major ways: learning, creativity, and wellness.
It’s been a ride, and we learned quite a few hard lessons about what it takes to make a bot that users actually want to interact with.
Stanson: Ivy Educated Digital Tutor
In the late summer of 2016 Stanson was made to help us stick with the ‘How to Start a Startup’ course by Sam Altman (amazing course by the way). We tested it privately but, it was an uphill battle.
We quickly learned that less is definitely more and that emoji and gifs are essential to an engaging bot.
We upped our game something nice but it was still a major challenge to get people to read the messages, let alone stick with the course.
Adding some inspirational quotes helped, and we even had a trivia quiz, but something was missing.
Lessons Learned:
- Onboarding is pretty much the ultimate challenge in product development, its the part where someone goes from a viewer to a user and it’s so important.
- Once users are in, they need to have something that makes them want to stick around and look forward to your messages.
Litkit: Beating Writer’s Block With a Writer’s Bot
It was now October and in advance of National Novel Writing Month we were thinking about how to make something that would make writing super easy.
Litkit was pretty awesome, and we still have a few users but it was so tough to get people to go through it for longer periods of time.
The content was solid and we started getting some nice long message threads, but only for some.
Setting a writing habit is hard, so we didn’t expect to deliver a magic pill. This is stuff that is tough for a sufficiently motivated person to do, let alone a random user.
We also faced a major issue in that changing API rules were breaking our core feature: the minimalist editor which would extract your messages and place them into a formatted document.
Lessons Learned:
- Solving a hard problem is hard. Don’t expect to crack it on your first go.
- Make sure your bot isn’t brittle. With our bot built so deeply around the editor, it was a tough sell whenever the editor wasn’t actually working.
ESPER: Prologue
Leading up to our last bot, we we went through a little disillusionment.
Were bots really a good way for people to experience things? Without either a deep web of responses or a superior quality of AI many people were falling into the abyss of trying to talk to it and feeling kind of meh about it.
Borne from my efforts to write a story in bot form during NaNoWriMo, we steeled ourselves and made our wellness/life coach bot: ESPER.
We went ahead and posted it on Reddit (as with the others) and man, we were a little nervous. We really wanted people to actually use it!
Incredibly, we were amazed to see that 95% of visitors were clicking through the onboarding message and 10% of users were exchanging over 100 messages with our bot!!
We were simply elated to see this. ESPER is a creative piece so there’s always that anxiety if people will think its stupid, especially since this one was running with our brand!
It looked like all the lessons we learned from our bots as well as the community of developers paid off.
In live testing people were smiling and clicking through for minutes, they were enticed to find more story and secrets.
The things that made a difference can be summed up in three points:
1. Enticing Premise
Whatever your premise is, make it enticing.
The Christian Grey bot made it feel like you were texting with this creepy but interesting dude. The Persona bot for Westworld was amazing with the way it invited you into it’s world and made it feel so real with the use of your name and the timing of its messages.
Having an enticing premise got us an amazing boost in click through rates and had a lasting impact on the user’s curiosity in exploring the bot.
2. Choices
Giving people choices that affect the progression of the bot make it way more fun and increase the investment. In live feedback sessions, you could feel people opening up and becoming engaged.
3. Actions
The action of breathing with the light was something cool, it connected the bot experience with the real world. A very powerful crossover that made it a lot stickier in people’s minds.
With this we made the story about the user’s real lived experience, and thus built a real connection with them.
Conclusion
Bots are indeed awesome. Though they have a host of challenges, they also have a powerful ability to be in the most intimate contexts of the modern user.
This power cuts both ways, if you betray or simply bore someone in their messages, your bot will fall flat at best and at worst actually make your users angry.
So what are people really looking for?
Heck if I know, but one thing is for sure: people are looking. They are curious and want interesting stories that engage them and respect their choices.
Try our bot out for yourself at www.tryesper.com
Thanks for reading. If you liked it, hit that recommend button: it gives me fuel to keep writing!
