Welcome to this edition of our interview series, where we chat with the founders of tools for thought who are on a mission to help us become more productive and more creative while taking care of our mental health. In a world where we all struggle with information overload, the team at Summari is reinventing the way we consume online content. More than a simple read-later app, they provide high-quality, human-made summaries of articles, newsletters, and podcasts, all written by students at top universities.
Ed Shrager, founder of Summari, discusses how Summari helps you navigate the modern media landscape by consuming more content in less time. In this interview, we talked about the art of crafting summaries, solving the biggest media paradox, building an alternative to serendipity for content discovery, long form summarization as a research topic in artificial intelligence, the impact of distracting social media on our inspiration, and much more.
Save 15 minutes reading the summary as a “Jog” — the kind of summary you would get for the content you consume online through the Summari app:
Hi Ed, thank you so much for agreeing to this interview. There are many summary tools out there, but your approach is quite unique. What inspired you to create Summari?
Summari was born out of a pivot. The company started as a payments company with a mission to recreate a new payment network that eliminated fees for merchants. Unfortunately that didn’t work due to the pandemic, so we pivoted.
During the pivot process, we were doing a lot of research as a team. Team members would find something interesting and post it in a Slack channel. This became overwhelming really fast, and the problem we now solve was identified: there is an amazing amount of knowledge freely available and this is growing very fast every day, however, the time available to consume this is not growing. Finding signal from noise is becoming harder.
In addition, it was clear that the prior distribution model for media is breaking: big name brands are being unbundled, and as talent fragments, the number of sources one needs to discover and follow increases. This makes it even harder to get the information people need to function in their professional (and sometimes personal) lives.
The “aha moment” was saying to the team that if they find something they have consumed that interesting, they should then share a summary of the content with others. That meant the team could look at the summary and have a better understanding of whether to spend the time digging in to see the full content or not.
For content creators, it’s a struggle to get someone to invest their time consuming their content. It’s a media paradox: you don’t know if you’ll find value in the time spent consuming until after you’ve consumed. The movie industry solved this with trailers and most people wouldn’t go watch a movie without seeing the trailer first.
The same doesn’t apply for general content across the internet. So we thought about the problem and decided to build a software tool to help gather and organize content users find across the internet, but then to add a lot of value by producing very high-quality human-made summaries of the underlying content.
Can you tell us a bit more about those high-quality, human-made summaries?
The summaries are not a replacement for the content itself: they are meant to give the reader an insight into the content and whether they should dig in further. This should serve to increase the probability a piece of content is read, which should increase readership for publications themselves — and this is a top priority for us.
We want people to have greater confidence in consuming content so they can learn what they need and benefit from the inspiration that content will generate. We have also found that the summaries are useful as pre-written notes after consuming the full piece of content. These notes are great at “jogging” one’s memory and aiding recall.
Right from the beginning we felt that we had to ensure that any summary we produced was exceptionally high-quality to do justice to the original content. If we want to increase viewership, then we need to produce a compelling enough summary for people to get value from it. We hear from users all the time that they really loved the summary and it encouraged them to view the main content which they wouldn’t have done prior to the summary. Unfortunately, we did not get the type of quality that we wanted from AI technology. We believe there is an art to a great summary, it’s not just copying select phrases from the text, there is a deeper understanding required and it requires a human, at least for now.
We are constantly on the lookout for new technologies to make our summaries better and faster. Several AI companies that focus on natural language processing have asked us for access to our data set given how large and unique it is. Long form summarization is an exciting research topic and with thousands of long-form human generated summaries written at Summari, we believe we have a very exciting dataset for future use in training AI models to help scale our product.
Humans seem to be central to everything you are building Summari. Tell us more about your community… How do you recruit the “geniuses” behind those “jogs”, and what’s in it for them?
We went out to top colleges like Harvard and Stanford and found students who were interested in careers related to the types of launch content we thought our users would be consuming. The interest was very strong and we began onboarding a lot of very smart students to Summari. This community gets access to some of the most interesting content on the web and in return they summarize it.
The summaries are formatted in a very specific way — this is something that we train everyone on to ensure consistency and quality. Clear concise headers, and short bullet points. We limit formatting to maintain a level of consistency so a user can consume the summary as fast as possible.
The community loves working with Summari and we have recruited students from over 25 top schools completely organically. The students love the access to world-class content discovery, and we also compensate them on a per completion basis. Many have told us they would still do it without the monetary component, but we feel strongly we should remunerate people for high-quality work and we don’t want to take advantage of young people.
We interact with them on a regular basis. We hired a recent graduate from Georgetown to help coordinate all of the writers. He is responsible for checking work, approving it and coordinating with the community writers on a frequent basis to continually improve our system. Because of this, we have continued to scale the supply of summaries using the same, or less, resources, which has been terrific — and we have many ideas on how to improve further. We track supply capacity in a granular way to ensure we are always able to produce the highest-quality content possible in the quickest possible time.
That sounds terrific indeed. But there’s so much content out there. How do you select which content to prioritize?
There is some content which doesn’t require a summary. For example, we don’t summarize anything under five minutes, as we believe the summary will not be valuable or high quality. We also try to stay away from short-life news as this has limited value to users too. Pretty much everything else is summarized, whether it’s a podcast, a newsletter, an article, a video or anything else. Users can hit the “Request summary” button for any piece of content they want summarized.
We recently launched the ability for users to subscribe to automatic summaries of specific feeds of content. We call these “Channels” and we started with a small selection of some incredible newsletters. Users subscribe in Summari with one-click, and whenever a piece of content is published, our system automatically notifies a community writer who produces a summary within three hours.
This is very complex and we have figured out a system to make it happen to give our users the most value from the product. This has been incredibly popular and has contributed significantly to our growth and usage KPIs. A lot of newsletters are very long or contain a large amount of linked auxiliary content, and unfortunately people do not read them. With the summaries, users can have a quick glance and decide whether to dig in more later, and use Summari to save the content for later. These Channels give the user certainty on what is summarized which we believe is a big benefit and we are working on ways to take this Channel concept further.
A common limitation of read-later apps is that we don’t end up actually reading the content later. How does Summari address this challenge?
This is the number one insight we learnt from pretty much every single user we spoke to of apps like Pocket and Instapaper. These apps were designed to help us bookmark all the things we thought we should consume that would be interesting for us. Unfortunately, these apps have not evolved and are commonly called the graveyard for my content or a place where my content goes to die. People accumulate huge lists of unread content and this is a big problem.
The human brain is amazing at generating ideas. If you’re like me, the moment when an idea hits is very enjoyable. Getting inspired by something is an amazing asset humans have and capturing that inspiration is something that most people find very difficult. Why? Because we are distracted very easily by the ever-growing number of demands on our time.
Swiping through social media is a constant barrage of distractions and those moments of inspiration are mostly lost. If you are inspired by something and save it later but never revisit it, that is a missed opportunity to capitalize on what could have been a defining idea in your life. These missed opportunities happen all the time and the apps out there are just not designed to help with it.
We set out to build the table-stakes: save content from anywhere. This enables idea capture. See something, get inspired, save it, be reminded about it. Then we went a step further… We summarize what you save and notify you when the summary is ready. This notification allows you to dig in just a little bit further and capture more of that inspiration, versus forgetting about the content entirely. We also allow users to snooze content for later to avoid information overflow and managing a list of content in your inbox, as well as the ability to swipe to mark as read and archive to get something out of your queue.
These are basic features but important for reducing anxiety about a massive list of content to consume and ending up consuming none of it, because let’s face it… Who has the time? While we consider these features relatively table-stakes and basic, the magic and value-add of Summari is really in the summaries and their ability to help you capture more of those ideas that can lead to great opportunities in peoples’ lives.
How do you hope for Summari to impact the way we interact with knowledge?
There are many ways to discover content: the main ones being on social media or through friends. Discovering the right content on social media is a function of serendipity: was I at the right place at the right time when this was shared, and for those with limited time for social media, this is a myriad of missed opportunities. Friends are a good source of information but a lot of the time we’re not in the right space to consume and links are shared and forgotten, buried by the conversation being had at the time.
Newsletters are a great way for people to get access directly to information sources they care about, but delivery to an email inbox cluttered with other demands on our time is not a great mechanism. Unfortunately, newsletters build up and then are often ignored or deleted. Within Summari, users can subscribe to a summary of many newsletters and will automatically receive a summary of each edition as soon as it’s published. Users can visit their JogBox to view this content at their leisure, and of course the summary is there to help determine where to dig in. This taught us people like getting access to and subscribing to specific feeds of content from content sources they trust. More on this later as we discuss our plans for the future of Summari.
On the face of it, the summaries we produce enable people to consume more content in less time: the ultimate productivity hack for time-strapped individuals. Summaries also allow you to recall content easily without any effort of writing notes. Sharing is an important function too. People like to share great content with their networks: we enable this and make users look smart by not just sharing a link that demands 30 minutes of their network’s time to consume, but also a summary with the key takeaways to help their network save time and be more informed.
The value of sharing a summary is compounded in a workplace setting or any setting where you are sharing with more than one person. Sharing a long article is in effect an ask for that amount of the person’s time. If you share a 30-minute read with a 10-person team, that’s 300 minutes of productivity you have just laid a claim on. If instead, you share a 1-minute summary that captures the key insights of the article, you’ve made all ten people smarter while only taking ten minutes of their collective time.
We have many plans for how to incorporate note-taking into Summari and integrate with popular note-taking apps. However, we have an extensive roadmap we are working to get through and these features will come once we launch a few other things which we know users want.
What kind of people use Summari?
The product is not exclusively for those seeking to optimize productivity. Avid readers or consumers of content who want to learn as much as possible but have limited time are a strong audience for us. Professionals who need access to the information flow of those influencers across the globe in their spare time are also prime candidates for Summari, especially for access to Channels and some future products we’re looking at launching.
Talking about future launches, what’s next for Summari?
Access to the right information is hard to come by and sorting signal from noise is an ever-increasing problem. People trust brands, but these brands are being unbundled and talent is starting to create single person media brands. People are willing to pay to support these individuals and want to consume a curated feed of information from them.
We believe this is going to create an overload of information and people will struggle to keep up — this is why at Summari we believe providing people with summaries helps them scale their time more efficiently. We are working on a plan for influencers to connect with their audience on a deeper level, giving their audience access to the content they’re consuming — all summarized and distributed by Summari. So people can get the information they care about, and consume as much of it as possible.
Thank you so much for your time, Ed! Where can people learn more about Summari and give it a try?
Follow us @getsummari on Twitter and go to Summari to learn more and sign up. Or check out this short video about the product: