If your company isn’t on Slack – the ubiquitous messenger platform helping teams “be less busy” – you’re probably stuck in an email bog somewhere. If your company doesn’t have an analytics framework, your products are built off of hunches rather than data-driven decisions. But it’s a good thing you have both. Now what?
When it comes to launching a new product, or understanding that uptick in sign ups, Artyom Keydunov knows it’s crucial to pull data into the conversation thread while at work. That’s why he created Statsbot, the bot company that makes Slack bots for business tools like Mixpanel, cutting down on the ad hoc queries and helps people be more productive. But Statsbot isn’t just making Mixpanel users’ lives easier – it’s also a leading example of how bots-for-businesses will become as common as wearing hoodies to work.
Hailing from the Russian tech startup scene, Artyom Keydunov, chief executive officer, and Mike Melanin, chief product officer of Statsbot joined us at Mixpanel HQ where they imparted their wisdom on how bots are changing the workplace. It turns out, being data-driven is now only a Slack message away.
Christine Deakers: Where did you first get the idea to build Statsbot?
Artyom Keydunov: In my previous job, I noticed a pattern where all discussions and decisions were being made in Slack. We used several analytics tools, including Mixpanel, and my coworkers (including myself) would would go into Mixpanel, take a screenshot of the graph, and drop it into Slack as an example, like, “Hey guys, what do you think about this metric change?” Once I realized that pattern, I knew pulling the data into the channel (without having to query it myself) would be really useful when having these types of conversations.
Christine: So once you had that realization, what did you do?
Artyom: Over a weekend, I decided to build a bot to test this idea. It turned out to be super useful in Slack. At this point, I knew I could handle the development of the bot, but I needed help on the design and UX stuff. This is when I called Mike [Statsbot’s Chief Product Officer].
Mike Melanin: Getting that call from Artyom was really exciting. I liked the concept right away, so I decided to try it at my own company. I was working at a bank at the time, so I connected the bot to our analytics platform to test it out. We even had our friends who were working at different companies try it, too. I think that was the secret to our early success – people trying it out early so we could learn where and how we could make it better.
Artyom: Definitely. Before our public release (where we launched our bots for business that support Mixpanel, Salesforce, and Google Analytics) in January 2016, we had a soft launch and did a lot of customer development, conducted interviews, and asked questions around how people used it, why they use it, and what other features we could we add.
For instance, people liked the digest feature. You can set your bot to receive daily or weekly digests at a specific time. (And that’s been a feature has been supported by Mixpanel since day zero.)
Christine: So how did you start building the Mixpanel Statsbot?
Artyom: It’s a funny story. Suhail actually reached out to us via email. At first, I thought someone was playing a practical joke, and then I saw his email was legit. He was really excited about what we built for Google Analytics and thought our bot would be perfect for Mixpanel, so we started working one-on-one with him.
Mike: Suhail’s partnership was really important in order for us to learn all the ins-and-outs of Mixpanel. Building a Statsbot for Mixpanel was going to be more complicated than doing one for Google Analytics, considering Google Analytics operates with a predefined domain and every Mixpanel project is customized. However, after doing a closed beta with several startups in the Bay Area, we got the overwhelming feedback that everyone in San Francisco is into Mixpanel, so this was worth the investment of resources. We didn’t immediately know what the best pattern or the best use case was for Mixpanel when building a bot, so we worked closely with Suhail, asked questions to thoroughly understand the product. Eventually, we left Moscow to visit Mixpanel to do all the final touches in person.
CD: So what were some of your biggest lessons you learned while building this bot?
AK: When it comes to botmaking for businesses, that deep product knowledge is essential, which is why Suhail’s partnership was so important. You have to know what “value” you’re delivering to a customer, and how can that value or information be expressed via bot.
CD: We learned from Kik’s bot evangelist that messenger bots can rocket a product’s engagement rates. Do you think messenger communities are new distribution channels for a product?
AK: Absolutely. Products can have alternative distribution channels, and a business bot can be the perfect setting, especially when conversational applications, like Slack, and analytics solutions, like Mixpanel, meet at a crossroads. Bots are awesome for those everyday, small tasks, that aren’t too difficult, but still take time. The magic of it really comes through experience – like when you’re in Slack having a conversation and you just need the data right away.
When you have a strong messenger community and an analytics platform, a bot is a perfect vehicle to deliver the data to your team, easily and quickly. For example, if someone wanted a summary report on “the number of ‘in app purchases’ today,” Statsbot could query on a user’s behalf, and deliver the graph within Slack. While it’s simple, it’s still a bit annoying to have to open a browser, query a simple report, and drag the report into Slack. But with bots for business, you won’t miss a beat. It helps keep those data-driven conversations flowing.
However, there are limitations to bots for business. Not every human activity can be replaced by bots, especially if the questions being asked are very complex. If you want to build a very complex chart in Mixpanel with a lot of conditions and properties, that’s a lot of work for the bot’s UI.
CD: Right. It’s easier for the user to go into Mixpanel to build more complicated queries than for a bot to do that type of analysis.
AK: Right. But we’ve also built the bot to understand the simplest syntax, so you don’t have to waste any time. Sure, you can ask Statsbot to pull data with a long sentence, or you can easily type in, “@statsbot number of messages sent today.” Statsbot will understand it just the same.
Typically though, you’re not going to have long conversations with this type of business bot. And that’s the whole idea. They can make an individual, team, or business more productive by saving time. We figure when you say less and do more you’re more productive, right?
CD: How do bots for business differ from other types of bots?
AK: The main purpose of a business bot is to help make us all more productive – to save time and money. That’s why mundane, simple, everyday tasks, like merging data from different sources, is a task that can be done by robots.
CD: Totally. Creating a frictionless experience via bots has been vastly overlooked by businesses, and not just in how the business can build a bot to service their customers, but how a bot can help teams internally at the company to make their jobs easier.
From my own experience, Mixpanel’s Statsbot helps keep the data front and center in our conversations, so decisions are made quickly and with all the right evidence before us. What I like best is how I can pull data immediately from Mixpanel but keep it in our Slack interface. Statsbot can do my bidding.
AK: Yes! A bot cuts down on those smaller tasks in order to get to the more complicated questions you’re trying to answer. That’s how bots can help from an individual level, but but bots can also help the organization-at-large, by facilitating a data-driven corporation.
CD: How do you think this is going to change the way people work? And what do you mean by “data-driven corporation”?
AK: People who own the data or have access to it are often the people who start the conversations or collaboration on an issue. When more people have access to the data, more team members can become problem solvers.
Mike Melanin: Definitely. I saw this at my previous job. Even as a designer, I didn’t have to go to the product manager to get access to the data to make a decision about a landing page. Instead, I just went straight to Statsbot to see how a new landing page performed in comparison to older ones.
If a business analyst, community manager, marketer, or whoever wanted access to the data, but didn’t have credentials to log into Mixpanel, then they could simply ask Statsbot a question via Slack, instead of asking the product team to query on her behalf. What this does is give everyone the same amount of access to the information as everyone else, so there’s equal opportunity to the data.
CD: It’s cool to see how Statsbot can help with the adoption of Mixpanel, or any other analytics solution. When data becomes accessible to everyone, and it’s ingrained in someone’s daily practice, then more people can make data-driven decisions. They can become closer to how the product is performing.
So speaking of product, what’s next? What’s on Statsbot’s horizon?
AK: We still have a lot we want to add to the Mixpanel integration. Right now we are working on smart alerts. It’s basically an alert system where a bot will ping you when you have major spikes or dips in your data. It’ll happen automatically notifying you when there’s something newsworthy. That way, you can see the topline change and dive in deeper, if it’s cause for any concern (or celebration).
The other thing we’re working on is programming Statsbot to merge data from different sources. For example, if a company was using MailChimp and Mixpanel then Statsbot could help correlate your email campaign activity with the revenue generated data. We want Statsbot to answer questions on not only one analytics system, but multiple, whether it’s Mixpanel with Google Analytics, MailChimp, Stripe etc.
CD: That’s going to make a lot of people’s jobs easy and save more time. Thanks so much for sharing your “bots-pertise” with us. For our Mixpanel users, can you explain how to integrate the Mixpanel Statsbot into their Slack channels?
AK: Thanks for having us. And sure! It’s super simple.
- First, go to Statsbot.co.
- If you already have Slack, click on the “Add to Slack” button on our homepage.
- Then, click on the “connect Mixpanel” button and copy the API key and API secret from selected Mixpanel project into the two fields.
- At that point, you should see a confirmation within your Slack channel where Statsbot confirms that it has successfully integrated with your account.
If you’d like to learn more about the Statsbot, Artyom and Mike are currently in the Bay Area raising a pre-seed round in hopes to expand their team and broaden their product line. If you’re interested in learning more about this bot-for-business company, drop them a line at firstname.lastname@example.org.