I worked at Betaworks this past summer. Pursuing an opportunity to work there was probably one of the best decisions I have ever made. The experiences were invaluable, and I have already documented two of the major lessons I picked up (Permission vs Advice and Influence vs Authority).
A highlight of the summer was getting to work on the botcamp program at Betaworks. My responsibilities quickly progressed from tweeting to bigger, more impactful projects, and I spent a little over a month working with the botcamp teams building messaging products and chatbots in both the consumer and enterprise space.
Burley has a lot of experience in the media, social, and advertising worlds and a distinct vision for what chatbot-powered journalism and storytelling will look like in the future. He talks about reading habits, consumer behavior, newsletters, and messaging as a creation channel and shares some interesting insights.
Sar Haribhakti: Tell us a little bit about your product, how you got to building it, and why you think messaging platforms are ideal for delivering that product experience.
Cam Burley: We spent the past two years monitoring the social graph and how users from across the web engage with everything from memes and GIFS to long-form articles and videos. We learned quickly that when users are in their news feeds and find something compelling, they look at the headlines/image and then go straight to the comments section. It seems so obvious that it’s easy to overlook. We all do it. I may click into an article I think looks interesting but four times out of five, I’ll go to comments, first, to see what the community is saying about it.
It was such an interesting insight, I wanted to learn more about the behavior. Looking at or reading “lean back” content is a one-to-many kind of relationship. But everyone can participate in a debate on a topic. There’s something about the many-to-many exchanges that have captured our attention since the inception of IRCs and internet forums. I think it’s only continued to pick up in social environments.
Our goal was to capture this behavior in it’s most streamlined, simple, and organic form by creating a space where the focus was around enabling the community to quickly see a buzzing topic and comment, first. How do we strip away the fluff in a news feed, in a long-form story or five-minute video and create an environment where people can contribute ideas or simply consume them, in real time? The idea is simple: Drop a bite (a snackable piece of content) and let the community react.
This made the most sense on messaging platforms because that’s what people are already doing in texts — they’re sending short bursts to individuals and groups. It’s the most endemic way to consume and share content on mobile devices. It’s almost so obvious that it’s easy to overlook.
Haribhakti: A lot of big and small publishers have been experimenting with chatbots for pushing out their content. But none of them seem to have gone mainstream yet. What do you think is working and not working with news bots?
Burley: This is such an interesting question because it gets at the larger question of what exactly is a chatbot?
It’s way too early to define the entire scope of what it is, but I can talk about what it’s not. A chatbot is not a simple RSS reader. Sure, you can programmatically accomplish this, but why? I don’t understand the value prop here. I do understand why some publishers have taken that path. They want to try something and iterate on learnings. I get that. But, like any other piece of technology, it needs to solve a specific problem w/ greater effectiveness and efficiency than the substitutes or its value is an open question.
Similarly, a chatbot is not a tool to drive inbound traffic from links. Why is that a better experience than the web? Or mobile apps? Or etc., etc.? There will be edge-cases, but chatbots are more effective when the experience is local to the messenger platform — when the messenger is in and of itself a channel for consuming content or creating content.
This seems to be the high-order bit. I’m not sure I’ve seen a ton of publishers pursue bots in this way.
The second piece that’s worth considering is the question around do people really want to talk to machines? Until publishers figure this out and just how much users want to engage with a machine that will respond with pre-programmed data, it’s going to be hard to drive the type of engagement and value they’re looking for. Is it faster? Is the experience more delightful? Is it natural? Are the interactions compelling?
We’ve certainly placed our bet on people wanting to talk to other people. And a smart A.I. technology can make that experience more efficient or insightful, but at the core it’s still humans communicating with other humans. I think when Publishers surface more human-centered bot products, they’re likely to see an engagement lift.
Haribhakti: Are there any chatbots that you particularly like or use regularly?
Burley: There are a lot of chatbots I think are cool. Selfishly, I wish some were less like a ghost town and more open to social interactions and exchanges. I like Shelfjoy a lot. I’m excited to see them grow their product. You gotta love Poncho. Their approach to weather plus media is singular and it’s fun. It’s short, snackable, service content wrapped in color. PennyCat is also cool. They are leaders in building chat-based games/quizzes.
Haribhakti: We all know how Facebook’s news feed algorithms govern the distribution of media companies. Google recently announced that its ranking algorithms may not favor web pages with interstitials. Do you think a combination of these two factors will influence the outlook of publishers on chatbots in general?
Burley: Perhaps it speeds the sprint to chatbots more than anything else. The fact is, news feeds are choking off publisher’s access to their own audience. I remember seeing Elite Daily on Facebook and a post would have 100,000 social interactions, routinely. Now, even with more than 1 million fans on their page, they might reach 2,000 on a good day. If that’s true, it means that the cost to reach one of their users is exponentially increasing, while revenue from that one user is not. That’s unsustainable. The contrapositive is also true. Using a subscription-based chatbot gives publishers 100 percent deliverability at little to no cost. So, in many ways, it’s a no-brainer.
That’s why we’ve focused on a newsletter product where publishers can schedule digests to their audience that are super-interactive and social. The one thing social media, specifically Facebook, does a great job with is driving echo referrals. If you share a cat video and your friend then shares it, your friend’s friends see it in their news feed. There’s a network effect, and that can be free…emphasis on can. It’s not guaranteed. It’s a little casino-like. You’re boosting content and hoping it catches a viral wave. That can be really cool and addicting.
Our focus will be to continue to better connect users around experiences they love. And, of course, this includes echoing content to users who wouldn’t otherwise have access to it. News feeds like Facebook solve a specific use case. Ours is different. We think most publishers in the space are clever and will see the opportunities chatbots like ours provide and will decide to try it to see if it’s a great solution for them.
If you are interested in learning how chatbots are not just software products and can serve as a medium for a commerce business, check out my interview with Rabi Gupta. He is building a gifting network using messaging as a channel. It’s interesting how Burley is trying to deliver the entire product experience from production to consumption in the chatbot itself while Gupta is trying to use chatbots merely as an entry point for kick-starting his logistics-driven, commerce business for nurturing relationships between professionals.
Source: New feed