What 10 Billion Messages can Teach us About Making Chatbots
Meetings with 5 Botfathers
I recently met with Dmitry Dumik CEO of Chatfuel, İlker Köksal CEO Botanalytics, Arte Merritt founder Dashbot , Alexander Gamanyuk founder at Chatbottle and Peter Buchroithner founder at Swelly.
Each of these companies has a top down view of the Chatbot ecosystem and collectively their platforms process more than 25 Million messages per day. Combined, they have analyzed over 10 Billion messages.
Here are some random thoughts, key insights, and a few tips you should consider when building your next bot.
5 Things your Bot Should Be Able to Do with Arte Merritt
Hi what’s your Name?
Your bot should be able to handle small talk. Most users start a conversation with ‘hi’ or ‘hello’ and some with the thumbs up emoji. This is a great way to create a good first impression and connect with your user.
Is it a Love Hate Relationship?
‘Hi’ was the most popular thing people say to bots which is not surprising, however what is surprising is the number of people that use strong emotional language with bots. About 10% of bots receive an ‘I love you’ or ‘I hate you’ type of message. Other popular emotional messages included, ‘thank you’, ‘thanks’, ‘you suck’, and ‘f*ck you’.
When sentiment was broken down into Love/Hate categories, about 30% of Bots received a ‘love’ message and about 15% received a ‘hate’ message. Even in Bots Love Trumps Hate :)
Multimedia Messages & Emojis
Your bot should be able to handle multimedia messages, emojis, videos, pics, location, and more. According to Dashbot, here are the top Facebook Stickers sent over the past three months:
12% Want to hear a Joke
It pays to make people laugh… Just like in real life..wait this is real life :) Your bot should be able to tell a joke or 2. About 12% of users asked bots to tell them a joke and here is how they did it:
Does your Bot have a personality?
Your bot should evoke a strong emotional response. Here is a look at the emotions that garner the highest number of views and shares across social media:
The best way to evoke emotions is through personality. Start looking at your Bot as if it were a character and consider what emotions you want that character to embody. In the end, what you are really selling is an emotion!
To Learn More About using Creating Personalities, Using Emotions & Psychological Hooks in your Bots check out:
Most Successful Bot Use Case with Dmitry Dumik
Currently, the bots with an already built in audience are enjoying the highest engagement rates. A number of types of bots fall into this category most notably, Celebrity Bots.
Celebrity Bots
Redfoo is an engaging bot that you can have a small talk conversation with. The bot circles around and introduces you to his content, likes, playlists, videos and more. It has well over 1 million messages.
Another interesting Celebrity bot is, Christina Milian. The bot has an interesting super power… she treats different users differently by segmenting the audience!
Gaming Bots
Gaming bots are also doing well. Swelly has over 850K users and engages users via trivia style ‘Hot or Not’ play. Another popular format is a ‘chose your own adventure’ Humani: Jessie’s Story, which was created by Hollywood screenwriters.
Transactional Bots
One of Dmitry’s favorite use cases was that of a Waterpark Bot. The Waterpark was naturally getting a lot of questions on FB regarding working hours, etc. One day they decided to create a bot that would answer these question. The bot has worked so well in answering common questions, that most users did not realize that they are talking to a Bot.
How PennyCat Increased Engagement by 87% with İlker Köksal
PennyCat is a popular bot that helps you find discounts.
A lot of people were giving them the good old try, but most of users never came a second time. PennyCat suffered from the, all too common ‘one and done’ syndrom.
After analyzing all of their messages, with the help of Botanalytics, they realized that users fell into different sub-groups based on their behavior and interest.
To re-engage them, PennyCat started sending out very targeted messages to each individual group with a focus on the person's individual interests and voila… They engagement increased by 87%!
Bots & Scalable Personalizations
Most of us are building bots the same way we built Apps and Sites; we are still building products for our entire user base.
Instead of A/B testing product with our entire user base and arriving at an average solution, bots allow us to do A/B testing with individuals and offer the best personalized solution possible.
Great bots won’t have 1 or 2 flows, they will have multiple flows and copy each tailored specifically to a person’s preferences.
To learn how to Segment an Audience with Zero Coding Check out this article:
Tips for Bot Builders with Alexander Gamanyuk
The hype is over and it’s time to and it’s time for good bots people actually use. More wisdom from Alex…
When Getting Started:
- Focus on a niche problem and do it great!
- Don’t use NLP in the beginning, start with a dumb bot :)
- Connect analytics from day 1.
- Don’t create bots for every possible platform. Focus on your target audience and the platforms they use.
- Communicate with users via your bot since this is best possible channel and it’s already established.
When Marketing Your Bot:
- Target your market as your would for a website or app
- Try to find repeatable customer acquisition channels.
- For Social Bots that can be used in group conversations, (like in slack or WeChat) group chats work the best. For FB Messenger Bots, it’s SEO, content marketing and bot directories.
How to Talk to Customers:
- Don’t Message users daily, instead 1–3 times per month or whenever you have something to say.
- Ask for advice, send surveys, etc.
- Be ready to connect live person if needed
Alex has recently launched Chatmetrics.io with the goal of improving engagement.
Hockey Stick Growth with Peter Buchroithner
Which dress should I wear? What movie should I see ? 🐳Swelly is a bot that helps people make better decisions by collecting their friends opinions.
To date Swelly is approaching 850,000 users and 60 Million messages! They are one of the most popular bots on FB Messenger and Kik. Best of all, they spend $0 on user acquisition! So how did they do it?
Multi-Platform Approach
Swelly’s main strategy was launching on multi platforms and thereby creating a massive social distribution channel. They started by creating an MVP Bot for FB Messenger and testing it. After the bot showed a lot promise they launched full versions on Messenger, Kik, and Telegram.
3 Marketing Tips:
- Product & Platform Fit: Make sure that your bot and the audience is a fit. Swelly and Messenger were a great fit; Telegram, no so much.
- Target Customers: Know who you target market is and go after them.
- Multi-Platform Approach: Launch on all viable platforms for maximum exposure.
Let’s Hack Chatbots Together
Creator of 10+ bots, including Smart Notes Bot. Founder of Chatbot’s Life, where we help companies create great chatbots and share our insights along the way.
Want to Talk Bots? Best way to chat directly and see my latest projects is via my Personal Bot: Stefan’s Bot.
About Stefan
Stefan Kojouharov is a pioneering figure in the AI and chatbot industry, with a rich history of contributing to its evolution since 2016. Through his influential publications, conferences, and workshops, Stefan has been at the forefront of shaping the landscape of conversational AI.
Current Focus: Currently, Stefan is channeling his expertise into developing AI agents within the mental health and wellbeing sector. These projects aim to revolutionize the way we approach wellness, merging cutting-edge AI with human-centric care.
Join the Journey on Substack: For exclusive insights into the development process, breakthrough experiments, and in-depth tutorials, follow Stefan’s journey on Substack. Join a community of forward-thinkers and be a part of the conversation shaping the future of AI in mental health and wellbeing.
Subscribe now to Stefan’s Substack and unlock the full potential of AI-driven transformation.
