The Wild West of Bots: 3 Key Takeaways from MobileBeat 2016
Last week, Verto Analytics participated in VentureBeat’s MobileBeat conference in San Francisco. This year, the two-day event focused on AI and the expanding bot-verse. The most compelling aspect of the conference was an overview of the various ways that chatbots can be used to handle mundane tasks and serve as a proxy for routine activities; anything from setting appointments to ordering flowers. Most everyone, including our CEO, agrees that chatbots could very well be the next huge wave in digital communications.
Bot Metrics: The Next Frontier
How we measure a successful chatbot is still largely undefined. In a session on chatbot analytics, panelists agreed that KPIs could vary depending on the purpose of a given chatbot. While engagement levels and MAUs (monthly average users) are metrics typically used for apps and services, a high engagement level can be a poor indicator of functionality for certain types of chatbots: eg., if a chatbot is designed to complete a quick transaction for the user. If a NLP-powered (natural language processing or voice-activated) bot is showing lengthy time-spent and a low stickiness index (metric for daily returning visitors), it may signal that the the voice-processing and matching features of the bot are not working efficiently or accurately.
But while transactional bots may minimize engagement, others, such as content-based bots, actually encourage it. For example, celebrity chatbots like FooBot (which rapper Redfoo of the musical duo LFMAO uses to communicate and engage with fans) might be measured on the level of intensity during a day. How often are fans engaging with the chatbot and at what point do they leave the platform?
Use Cases for Bots: Automating Date Night
When we already have a plethora of communications platforms, do we really need another? At MobileBeat, the answer was a resounding yes: we can use chat to get away from the pesky noise of SMS, emails, and app notifications. But the use cases can go even further: it’s about utility. What can a bot do for me that I do not have time, energy, or the brain capacity to do? Utility bots are purpose-drive and designed to keep friction low: if I want to order flowers from 1-800-Flowers every month for a different friend, I could just tap their address into the chabot and be done. It’s a simple request that can be set up and configured to be made repeatedly according to my specifications.
What if an ad turned into a bot? Here’s an interesting way that advertisers could nearly guarantee engagement for a time: grant an ad bot permission to ping me when it has found a good deal personalized to my interests. Publishers like Fandango and Shopify are already using bots to make it easier for consumers to buy movie tickets and shop. Bots can remember what you have done and customize and personalize products for your needs. This ability to microtarget will be a goldmine for advertisers.
The Bot Discovery Dilemma
According to VentureBeat’s founder, Matt Marshall, Facebook has over 11,000 bots on its platform. But discovery is an issue that many bot developers – just like app developers – will face. For now, many chatbots are being built upon existing platforms, despite the potential pitfalls that presents for bot developers. Just days ago, we published a messaging apps report which looked at the messaging app platforms that have the widest reach and the highest fan engagement.
We’ll continue watching Kik’s bot store as an example of what direction the bot ecosystem could take. They have the highest amount of user time spent in-app, and they already have a fairly active bot store that includes flagship ecommerce brands and clever bots targeted at niche audiences. For example, within Kik, you can tell the Sephora bot which product you want to buy, and you’ll receive a customized feed of reviews and details. Or you can ask the Spock Chat bot if it thinks you are logical.
Want more great insights on the future of messaging apps and bots? Stay tuned to our Twitter feed for the latest updates and research on this topic.