On How to Save (the) Medium
Ev Williams, Medium’s CEO recently admitted that Medium needs a new business model. One that will transform not only their business but an entire industry. That’s a big goal but candidly it’s not just Medium that needs it. Here’s what he said in recent a blog posting:
“So, we are shifting our resources and attention to defining a new model for writers and creators to be rewarded, based on the value they’re creating for people.”
If Medium can do this, it will indeed be transformational. In Medium’s space no one has been successful — not the least of which Medium which has been trying for years. Part of the reason is that rewarding writers and creators for quality stories means you need to be making money off those stories. Medium and others like it struggle to solve this problem because they haven’t been able to get people to pay. That’s about the short of it.
Creating value from the Wall of Words
The problem for the Mediums of the world is that there is so much free stuff online that people don’t feel a need to have to pay for it. People find almost any distraction they want online immediately and at no cost. Who doesn’t love that!
It reminds me of the conundrum that we had when I was running Nook Press for Barnes and Noble. Nook Press is a self-publishing platform and as such we encouraged and indeed had thousands of authors uploading hundreds of thousands of books, most of it free. While we built a great business with Nook Press we knew that the real value of it was literally buried. Buried under ten thousand copies of some classic. Buried under the umpteenth version of a really poorly written romance novel. Buried under the sheer weight of success with “user-generated content”.
What the publishing industry as a whole suffers from is the same thing: a wall of words. A daunting, ever growing, seemingly insurmountable wall of freely available, snack-sized, bite-sized, right-sized, any-sized package of words covering any topic by anyone with an opinion.
At the end of the day the only way out of this conundrum is to use that wall of words to create value for the reader. An individual. Someone who has their own distinct likes and dislikes. It’s about mining the wall of words. To help readers find just the right stuff that’s going to be valuable to them.
What I’m talking about is good old-fashioned merchandising. If you’re in the business of “user-generated content” and you don’t have a scalable system for helping you identify what you’ve got and what you’re receiving from creators, then you can’t properly merchandise anything. And by merchandise I mean getting the right stuff in front of the right people so that they’ll pay. If you can’t merchandise, then you can’t reward those creators monetarily. And if the creators who are creating valuable content for you aren’t being compensated then they won’t continue to create that value.
Therefore if you want to “create a new model for writers and creators to be rewarded, based on the value they’re creating for people” then you must create a scalable solution where the business can classify every story. The classification needs to be done to such a level of detail that you can build on-the-fly, personalized merchandising that gets the right content to the right consumer at the right time. You have no choice. This has to be done. Full stop.
At the end of the day we’re talking about “big data”. This is big data analytics on your stories. And it sounds an awful lot like another company’s game plan that is successfully disrupting the television and movie business: Netflix.
Netflix as your model
People like to think that Netflix is just a streaming movie service which of course it is but that’s not what makes it disruptive. Netflix, at its core, is a data analytics company. What they’ve successfully done is reinvent television and movies by leveraging the data on their shows and viewer behaviors to drive a subscription model that is turning the industry on its head. They are leveraging their “big data” (detailed metadata that they have on each movie) to better recommend shows to their customers. In other words they are using data analytics to add value to their customers. Sound familiar?
Of course being a clever data-driven company, Netflix is further leveraging their data-dominance by using it to inform themselves of what shows they should pick up or develop. “We have a lot of female viewers watching reruns of Full House. They’re looking for wholesome, nostalgic, family comedies. If we remake Full House but feature women we’ll see a 17% increase in engagement among our core female audience aged 35–50.” Imagine that, using data analytics to help you maximize value with your customer by feeding detailed data about viewer behaviors back up stream to your content creating partners. That also sounds like a good plan for Medium too.
To make any of this work, you need a lot of data on your content. The more metadata you have the better merchandising you can do. Netflix does this manually. They literally have people fill out forms for every show. This data is then entered into systems and Netflix’s machine learning algorithms do the rest. This manual data entry works fine for Netflix because the number of shows is dwarfed in comparison to the number of books let alone Medium stories that are created each year.
For such a classification and meta-tagging system to work for Medium, you need a scalable, AI powered system that can classify millions of stories in minutes. Such a system exists and it’s called Intellogo.
Creating more, better “Big Data”
Intellogo is a cloud-based solution for generating big data from any kind of text. It was specifically designed for long form content like books but has grown to include web articles, Wikipedia and even YouTube videos.
Let’s hypothetically say that Medium is pumping all its stories by Intellogo. It all starts with Intellogo analysing a story and then auto-tagging it to an unprecedented depth. Every story must be auto-tagged. The tagging cannot be done by humans as that’s error-prone and incomplete. Intellogo ignores any metadata that is already present. In Intellogo’s world we call tags Insights and they can be grouped into things like: tone, themes, formats, styles, etc. Intellogo is an open-ended system so Insights don’t have to belong to any of those groups and can exist on their own.
Here’s the actual Insight cloud Intellogo generated for Ev’s original Medium story. Larger words means Intellogo finds the story to be more of that word. Intellogo accurately tags the story as about Strategic Management, Leadership and Breaker and Shakers. But it also gets the fine details right too like it being about Business Planning, Business Operations, Executive Innovation and Dynamic Creative Problem Solving. It correctly identifies the point of view as being Executive, Analytical, Thought-provoking and Cutting Edge. It also detected a bit of Despair.
So, half of the equation is generating lots of metadata on each story. The other half is tracking what everyone is reading and creating predictive reading profiles for each reader. Here’s my reading profile as tracked by Intellogo:
Because we now have such a high level of detail on each story, we consequently have a high level of understanding of each reader. In my case, you can see that 61% of all stories I read are ‘Concise’ and ‘Analytical’. But Intellogo goes deeper and can tell you that when I read about science I tend to like stories to be written plainly, be about the cosmos and be thought-provoking. By the way, Intellogo predicts that Ev’s story is a 72.89% match with my reading profile and that’s pretty accurate.
Conclusion
The bottom line is this: If you want readers to pay for content you need to create value for them. In order to create value for individual readers, you need an in-depth understanding of what they like to read and have accurate predicative analysis. And in order to have accurate predictive analysis, you need to have data, a lot of data on your stories. When you have lots of data on your stories you can do the best job of matching up content with readers. More importantly, you can better inform story creators with what readers are interested in right now.
This is a model that works. Netflix has proven it. It’s the way to realize Ev’s vision of:
“…building a transformational product for curious humans who want to get smarter about the world every day.”
The industry needs an automated, scalable system for generating lots of data from text: A way to generate the big data that we need to empower business innovations; that unlocks hidden value for readers; that ultimately results in products and services they are willing to pay for.
