Poncho the Weather Cat, An Interview with Greg Leuch
(Part of the Bot Master Builders Series)
There has been a lot of bot hype in the last 12 months and only a handful of good bots. Poncho the Weather Cat has stood out as one of the great early chatbots. He delivers information about weather, and if you want it, horoscopes and some light social games.
The Poncho team’s broader goal is to understand how best to deliver ‘thin content’ via chat and notifications platforms, pushing what it means to be a content company the way the 24/7 cable TV Weather Channel did decades ago. Poncho is currently available via SMS, email, Slack, Facebook Messenger, Kik, Viber, iPhone, and Android. The 11-person company was launched out of the Betaworks accelerator and it raised $2 million in seed funding from a group of investors led by Lerer Hippeau Ventures.
Below is my interview with Greg Leuch, the Head of Product for Poncho.
What was the original design vision for Poncho? Just for weather notifications, or something else?
Poncho began his career sending text messages to his friends in New York City and began to chat and make friends all over the world!
Kidding aside…Poncho was started by Kuan Huang as one of Betaworks’ Hacker-in-Residence incubators. The inspiration for Poncho came from his mother. While in college, Kuan’s mother would text him daily with reminders (as many of our mothers tend to do), with messages to wear a jacket because its cold or bring an umbrella because it will rain later. By taking these helpful reminders and combining them with humor, pop culture, and weather in a short-message format, Poncho was able to have a engaging purpose that differentiated itself from generic weather messaging services. From there, Poncho expanded city-by-city across the US, eventually going international, while along the way launching email, iOS, Android, and chat services to better showcase our editorial and visual content.
Team — how did the Poncho team come together, and what are the roles?
We’ve been fortunate to have a great team of smart & talented individuals with backgrounds in comedy, hacking, math, poetry, and gaming. We’ve strived for being diverse & inclusive, for which the results have allowed us to build a company culture is unique and fits perfectly with our brand. Our team of 11 employees is a mix of editorial writers, engineers, designers, and startup veterans.
How do you measure success; what are your metrics?
For our chat products, we study our daily retention metrics with focus on D7 and D14 [7-day and 14-day], which we’ve been able to hit a 60% retention rate for D7 with our key audiences. Our success in this metric has been A/B testing on-boarding and content offerings, but also just being awesome with our editorial content and various functionalities. We also use unique and total sessions to understand trending growth and engagement.
What are successful interactions? What are failed interactions?
Poncho does an amazing job with a wide-range of weather queries. We’ve trained our models on many permutations of weather conditions, locations, timeframes, slang terms, etc. We’ve also done very well with games and off-topic content, seeing high-use with seasonal or limited-time games & content flows. And while we’re doing a great job, we’re currently English-only. We have many users that are not native or fluent English speakers, so some content may not be understood, or on some platforms we’re just not able to reach those audiences. We’ve also spent a lot of our time on user retention, and only now have we started efforts to improve sharing and viral components of Poncho.
What are the most common things users ask outside of weather — do they ask for jokes or other advice?
A bi-weekly horoscope was a big feature we offered until recently. It was not something all users wanted and there were very polarized audiences that were for or against it receiving horoscope content. Our limited-time/seasonal games also got a lot of attention and tends to really increase daily sessions. For example, when we launched our SCREAM therapy engagement during the US presidential inauguration, we saw our daily sessions nearly triple from our normal amounts. All of this is possible due to the nature of our content management setup and our talented, nimble-minded editorial & design teams.
What other bots have you looked to for inspiration — who else out there is making a decent bot?
Too many to decide! Although they just launched and are very early, Hugging Face has been fun to watch. And the Dexter team recently launched their platform for building bots, with some influence by Poncho on the development of their editor interface and syntax language. But mostly, the PlusPlus++ Slack bot has been a killer fave in any Slack team I’m in.
User acquisition strategy — how do people hear about Poncho and start using him?
Up to this point, Poncho is all based on word-of-mouth. We’ve focused our efforts on making it engaging, playful, and useful, for which our reward has been high retention and respect from our users. That respect has in turn generated positive publicity, press mentions, and plenty of examples & citations by others designers/developers/botmakers for best-practices in chat interface design. But with a focus now on engagement and sharing, we’ll start to see a lot more acquisition coming from sharable content.
Re-triggering and re-use strategy — as many bot developers know, how do you get initial users to engage Poncho again without annoying them — what have you learned?
Honestly…be direct, be mindful, and don’t be annoying. Our litmus test for many things we do at Poncho has been “would my friend send/do that?” We want Poncho to be your friend, the one who is respectful and informative but not too annoying either. Re-engagement with this focus has done well for us.
Thoughts on the different platforms? Why has FB Messenger been the main one and have you done much on other platforms? Voice platforms?
Along with Messenger, we’re also available on Kik, Viber, and Slack. We’re also very experimental with our content, tone, & features. We’ve prototyped on nearly all platforms, however we’ve have chosen these so far based on user audience, product offerings, & expected adoption rates. Voice platforms, as well as video/VR/AR, have been a struggle from a character development focus. We’ve yet to be satisfied with how we think Poncho should act & sound on these platforms. But we have a slew of new features, expansions, and new launches planned over the coming months, so stay tuned…
Monetization? So far, Poncho is a great free service — how have you tested monetizing it?
We’ve experimented with mixtures of sponsored content, app inclusions, marketing emails, and co-branded partnership. We’ve been satisfied with these results and we’re actively engaged in deeper integrations & partnerships.
There has been much chatter in the Silicon Valley developer and UX communities on how terrible the Facebook Messenger bot platform is, and how Facebook execs may be abandoning it. That poses a big risk to Poncho and other bot developers, as FB Messenger is a big delivery platform. Do you have other ways of staying in touch with users? What have you discussed to be more platform independent?
Poncho is based in New York City, so we’re not always privileged on the latest gossip. With that said, Poncho is focused on providing friendly & witty weather, lifestyle, and workplace content across many mediums, including Slack, iOS & Android, as well as our bread’n’butter of email and SMS. We’re not constrained to a single platform or product focus. In fact, I’m excited to see how Facebook continues to improves the Messenger Platform with input and feedback from the bot-maker community.
