Q&A: Michael Sippey, Chief Product Officer at Medium
This week, The Idea caught up with Michael Sippey about product innovation at Medium. Subscribe to our newsletter on the business of media for more interviews and weekly news and analysis.
Can you tell me a little bit about yourself and your role at Medium?
I’m the chief product officer at Medium. I manage two teams: the product management team and the product design team. We’re responsible for the “what” and the “why” of what we build in our app, on our website, and in our email products for both sides of our marketplace which consists of readers, writers, and publishers.
How large are both the teams?
16 people overall. And, it’s split roughly 50/50 between product design and product management.
As a team tasked with answering the “what” and “why”, how would you describe the role of Medium in the publishing space?
Medium is an incredible platform that allows anybody to publish and have a voice online. We curate and discover the best content as well as publish our own original content and bundle all that into a compelling subscription for readers. Essentially, we’re marrying a content platform that allows anybody to have a voice with a subscription offering that uses the original content we’re publishing and content created in the community to deliver value to readers.
A lot of the things that we’re working on right now is finding the best way to surface those independent authors and independent voices alongside the professional content that our in-house team is producing.
Could you tell me a little bit more about some of the products and features that you’ve worked on?
I’ve been at the company for about two and a half years. When I started we had a very nascent subscription business. One of the things that we’ve been focused on over the past couple of years is better understanding how to deliver value to readers.
This means: 1) helping readers discover great content and original voices, 2) having a personalized and clean reading experience, and 3) building a relationship between Medium creators and readers over time through our website, email products, and app so that readers feel compelled to convert and become paying subscribers.
On the content creation side, we’ve been investing in our Medium Partner Program over the past couple of years. This enables any writer on the platform to sign up, connect their bank account, and get paid for their writing based on reader engagement. We also have a program called Amplify that helps writers get more distribution, offers professional editing help, and grants them exposure in the publications that we’re publishing ourselves.
Recently, we’ve been working on two big projects.
One project is the relaunch of four of our in-house publications. Four of our in-house brands (Gen, OneZero, ZORA and Level) now have brand new publication homepages that allow editors to be much more expressive on the site. So, it’s been fun to work with those brands to test some of the new tools that we’re building right now.
Our other project has been focused on improving our personalization and recommendations technology.
It’s a little bit less visible on an everyday basis but it’s been fun to work with machine learning engineers and recommendation engineers to continuously improve the algorithm that we use to recommend stories on our homepage, app, and in our email digests. We’ve been doing a lot of iteration and experimentation with the recommendation system. What I love about what we’re doing at Medium is that we’re trying to find the balance between giving professional editors, creators, and writers the tools to express themselves with crafting truly personalized experiences for readers based on the content that’s being created.
You get the yin and the yang of media here at Medium.
Given the range of products your team works on, how do you go about measuring impact?
It comes down to the scientific method. With everything that we build, we have a hypothesis that reads something like: “If we change this algorithm or we make this change in user experience, we believe it will lead to this type of outcome either for the business, for our readers, or for writers.”
Sometimes the outcome is measurable. We look at metrics like read through rate on articles which is the percentage of people that are reading an article. We also look at increasing the reading time of users overall.
For other experiments, our hypothesis might read as: “If we change this user experience, then it will unlock more creativity from our publication editors or writers.” These are a little bit harder to measure and we tend to ask ourselves, “Are we proud of this experience?”, “Are we putting something great into the world?” and “Do we think that it’s making the Medium product and Medium reading experience better?”
These are changes that take courage. The numbers may move in one direction or the other, but if it makes the experience better, then it’s an investment in the long-term health of the business.
How does your team prioritize reader and writer experiences when tackling problems?
Luckily, we have a big enough team and enough engineers to work on two parallel paths when it comes to building tools for writers and crafting experiences for readers. What’s interesting is that those incentives are aligned. Writers want to create great things for readers and want their stories to get viewed by the right audiences. And readers want to find the stories that they can connect with.
We’re not only interested in reader experiences of consuming Medium content, but we’re also interested in crafting a product that guides a writer or editor to create a great experience for readers. At the end of the day, the connection between a creator and their audience is what really matters and it’s really what we’re here to enable. You really can’t look at one experience in a vacuum without the other.
It’s really fun to think about different types of users and craft experiences where the two can meet in the middle.
How has your background in English informed your experiences as a product manager?
There’s been an interesting trend in the last 10 to 15 years in product management that has been around pure data-driven, product skills. What I’ve found is that the best product managers that I’ve worked with are people that are really good storytellers. They can take a lot of disparate information from users and customers, understand their problems, and turn that into a story that a team can understand.
And, this is why I still read a lot of fiction and care about the craft of storytelling. I believe in the power of the written word and this is one of the reasons why I ended up at Medium.
What were some of the most surprising takeaways you’ve come across at Medium?
One lesson we’ve learned is that bolder experiments are always better. You think that you can iterate through small A/B tests, but the way to really understand a user’s motivation is to be bolder in your experimentation. When we do bolder things with our experiments, we have deeper learnings.
At Medium, we like to say “M.O.L.O” which is Medium only lives once. So, there’s no reason to do small things. You’ll either succeed or fail spectacularly, but either way you’ll learn something.
The other lesson is that humans are weird. You can have a great and logical hypothesis, and still have results that you don’t anticipate. So, don’t be surprised when the results are surprising.
What is something interesting you’ve seen in media from an organization that’s not your own?
I’m always impressed with Spotify’s Wrapped Up campaign. It’s very on brand for them and they make it a really great user experience across their platform.
It’s also a reminder of the limits of personalization technology in my opinion. Computers are very good at counting things. They’re good at telling us what’s popular across a large population or even popular amongst people you share behavioral attributes. With Spotify Wrapped, it uses listening time as a proxy for “importance” for you. While those things might be correlated, they’re not causal.
For instance, I spent a lot of my year listening to podcasts on Spotify but those podcasts were not the most emotionally resonant things for me on Spotify. I also spent a lot of time listening to Taylor Swift this year on Spotify. And, while Taylor is definitely emotionally resonant, there are songs that had bigger impacts on me this year than her song “Lover.”
It’s a really great package, but it’s also a reminder of the limits of personalization technology. My 2019 Wrapped isn’t going to tell me what’s good, it’s going to tell me where I spent my time.
My best music recommendations don’t come from what the algorithm tells me, but from what my friends tell me. And, this is something we think a lot about at Medium — using technology to drive personalized discovery while enabling individual creators to have direct connections with their audiences.
Rapid Fire
What is your first read in the morning?
My inbox, unfortunately. And, then, I generally skim the home page of The New York Times and the home page of Medium on my phone while I have coffee.
What’s the last book you’ve consumed?
I just finished Exhalation by Ted Chiang. I’m currently reading The Topeka School by Ben Lerner. My favorite book of the year was Trick Mirror: Reflections on Self-Delusion by Jia Tolentino.
What job would you be doing if you weren’t in your current role?
I think I’d want to be a curator at the Museum of Modern Art in New York. That’d be a fun job.
This Q&A was originally published in the December 16th edition of The Idea, and has been edited for length and clarity. For more Q&As with media movers and shakers, subscribe to The Idea, Atlantic Media’s weekly newsletter covering the latest trends and innovations in media.
