Miso.ai

Public speaking is always a bit nerve-racking. But FIPP is easily one of my favorite media conferences every year. I mean the walk to the venue along the ocean definitely doesn't hurt. But the people who attend it are the real draw. Something about being in the sunshine makes everyone very open to sharing what they're doing, what's working, and what's keeping them up at night. It's just a great community. And as Miso's Answers platform continues to scale with our amazing partners, we wanted to share what we've learned from the last 3M Answers we've delivered. I'll save the slide deck for another time and just stick to the highlights: Insight # 1: Readers Just Want Answers ??♀? Answers are often seeing 10-50x higher daily usage vs legacy keyword search. Which isn't a shock honestly. What would you prefer, a bunch of links or a contextually rich trusted answer from a trusted source? Insight # 2: Meet Readers Where They Are ?? If you put AI search on your news site and expect overnight adoption, it's going to be a long wait. We instead contextually insert in every article interesting questions to explore in Answers, which we nickname..."Explore" questions. These Explore questions usually drive 80% of the usage of Answers, and are 5-8x higher CTR than article recs on the same page. Insight # 3: Answers Can Be Default Profitable ?? AI isn't cheap, but we can still put money back into the newsroom, versus taking it out. Whether it's contextual affiliate ads, sponsorships, 1st Party Data, personalized display ads or an integrated subscription funnel, we're finding new and creative ways to have Answers pay for themselves and the journalism that's powering it. Insight # 4: READERS DON'T CLICK ?? I should caveat that Miso's Answers are running on partner sites, versus as a bundled search engine, but we're still only seeing 10-15% clickout rates on sources. Readers really do seem to be taking this as "Trusted Answers from a Trusted Source." Insight # 5: Yesterday's News Isn't Dead ??♂? Even with a focus on Answers that are up-to-date, older content is still important for providing important "how did we get here?" context and insights on an Answer. Up to 80% of Answers contain an insight from longtail content. We all need to start thinking about the value of our archives differently. Information isn't just for human consumption anymore. It's for machine consumption as well now. Go get those licensing deals, but also get royalty deals for the long tail value of your work. A big thank you to Alastair Lewis, John Schlaefli, Cobus Heyl, Dr. Bastian Schwithal, Yulia Petrossian Boyle and Jeremy Walters for a great event. And many thanks to new and old friends Roxanne Fisher, Aled John, Stefan Seghers, Richard Lander, Madeleine White, Maanas Mediratta, Eirik N?sje, Jean-Paul Reparon, David Buttle, Preethi Nallu, Christoffer Birch-Jensen and Srinivasan B for the great chats. See you all next year! ??♂?