Enhancing responsiveness with bot analytics
Increasing user engagement is critical for delivering customer experience in numerous industries. In our increasingly connected world, being able to monetize that data is a game changer for the communications industry. Dashbot is a bot analytics platform that offers metrics like engagement, retention, demographics and sentiment. It enables marketing managers and customer support teams to guage user experience or campaign performance in real-time.
We recently spoke to their founder Arte Merritt ahead of his panel session at Mobile World Congress.
Tell us about Dashbot?
Dashbot offer bot analytics with additional tools to take action on the data. In addition to traditional metrics like retention and engagement, we provide bot specific ones like sentiment analysis, conversational analytics, AI response effectiveness, and even the full transcripts. We also have tools to take action on the data like our Live Person Takeover of chat sessions and Push Notifications for reengagement.
We support Facebook, Alexa, Google Home, Slack, and Kik natively, and have a generic API for any conversational interface. We launched about 10 months ago, have over 2,200 bots on the platform and have processed nearly 900 million messages.
The three co-founders are serial entrepreneurs – myself, an MIT alum, I founded the mobile analytics platform Motally, which I sold to Nokia. Jesse Hull, an MIT alum, co-founded Mesmo.TV which he sold to the Game Show Network. Dennis Yang, a Cornell alum, co-founded Bureau of Trade which he sold to eBay. We’re funded by ffVC, Bessemer Ventures, Samsung, Scrum Ventures, BDMI, and Rembrandt Venture Partners.
What kind of insights can bot analytics offer?
Bot analytics are quite a bit different than traditional web and mobile analytics. The data is much richer and more actionable. There are four main reasons why bot analytics are different:
- The tracking mechanisms are different. Anything that is click stream or event based tracking loses the richness of the messaging.
- The content is different. Users are sending unstructured data to bots – images, audio, video, location maps, and more importantly, their own voice, their own words saying what they want from the bot, and what they think of the bot, which sometimes may not be pleasant.
- The processing is different. Conversational interfaces can be multi-user and asynchronous – this is quite a bit different from web and mobile. It opens the door to different types of users and sessions depending on how active or passive they are.
- There are new types of reports. With conversational interfaces, now you have new types of reports like sentiment analysis, AI response effectiveness, conversational analytics, the message flows through the bot, transcripts, and more.
Given the data is much richer, there’s an opportunity to take action on the data in real time to improve the user experience and retain users. For example, with our Live Person Takeover, if a user is having difficulty with the bot, you can pause the bot, insert a live person, and help lead them through to conversion.
How can brands get (and offer) more value from this kind of platform?
Our platform helps brands increase user engagement, retention, and monetization. With our reports, product managers can learn more about their users and see how users interact with their bots including the common messages, user flows, demographics, and more. Marketing managers can use our CRM tools to re-engage users and bring them back to the bot. Customer support can use our Live Person Takeover to help users have a better experience and lead them to conversion.
Given users send in unstructured data, it’s important to look at the messages and see if your bot is responding properly. We have reports for Top Messages in and out from the bot, as well as our Message Funnel which allows customers to see the flows through their bots. With this information, they can see where the bot is breaking down or not responding appropriately and improve the responses. They can also see the types of messages users are sending in which could lead to new features and functionality.
For example, one of our customers was initially not responding to the images users sent to their bot. After seeing how many images there were in the reports, they added support for images – acknowledging that users send them in. This created a personality for the bot which resulted in overall increased engagement. Another customer added multiple features based on the top messages users were sending to their bot. Their bot provides cricket scores to fans. They noticed users kept asking for player info as well, so they added a feature to enable users to follow players. They also implemented a mute functionality for when users’ teams are losing, they don’t have to be reminded of the score – this has lead to increased user retention.
What does it mean for customer acquisition and CX?
Dashbot helps track the effectiveness of ad campaigns to improve user acquisition and provides tools and reports to increase user retention and engagement. On the customer acquisition side, in the case of FB Messenger, we can help track referrals so our customers can identify which channels and campaigns work better.
In regards to retention, our Push Notification feature enables customers to segment their audience and send broadcast messages to re-engage their users – all based on the data. Through our Top Messages in and out, Message Funnel, and full transcripts, developers can better understand how their users are using their bots and improve the overall user experience.
With our support for voice-enabled interfaces like Google Home and Alexa, customers can also track the intents and contexts of users messages and improve their AI response effectiveness to increase engagement. As mentioned earlier, our customers who noticed new types of messages coming in and/or the different content types of messages, and added support for those messages, saw the engagement of their bots increase.
What’s the technology behind it?
We’re hosted on Amazon AWS and use a variety of their features including Lambda functions, EC2, S3, and more. Our code base is primarily Node.js on both the front and back ends, with React on the front end as well.
Are we at the tipping point for Natual Language Processing?
It’s definitely looking like this is the year for voice-enabled interfaces. We see a lot of interest in Amazon Alexa and Google Home from brands and developers. The NLP is pretty good and will only continue to get better. These voice-enabled devices are providing the “iPhone moment” for bots. You may recall the videos of two year olds knowing how to swipe on an iPhone or iPad. We see similar occurrences with voice enabled devices. In fact our co-founder Jesse has two young boys who already know how to interact with Alexa – “Alexa play Star Wars,” “Alexa play the happy song.”
How can developers get started?
It’s very easy to get started – developers can sign up for free at www.dashbot.io. It’s super easy to integrate – just 2 to 3 lines of code. Developers can be up and running in minutes! At a high level, the way it works is a copy of any message into, or out from, the bot gets sent to us. It’s all asynchronous and non-blocking. You can learn more at www.dashbot.io/sdk or sign up for free at www.dashbot.io