As far as fixing it, the obvious first step is to at least be more generous with the article count... and don't show it at all until someone has seen at least 5 articles on separate days that month. Count a session clicking through a few articles on the same day as at most 2 article views (maybe have a separate limit of something like 5 articles in the same day). Don't be a nag trying to upsell when people haven't even had a chance to get interested yet.
That at least would cut down on the sleazy feeling of the whole site. An easy next step would be hiring someone as an editorial curator. Most of their job to start would be to build a regex list of article title structures which are overly cliched, so that those articles get de-merited by the recommender algorithm
Another way to feed a meaningful recommender: Allow users to tag posts (and other more clever things like deriving tags by looking at, e.g., which sub-reddits link to them). Build user profiles that are similar to how StumbleUpon had a "Stumble DNA" for users.
I think the issue here might be of a more general nature.
Would it be fair to say that we usually expect blogs to be free to read? I think yes, because that’s how it has been for some time -- somebody wants to share their thoughts, you want to read them, no money involved.
Now, some people run paid newsletters, and can charge you for their content, and you pay them if you feel like you’re getting value out of it. Investment advice, curated content, you name it.
But that’s precisely the thing: if a blog is paywalled, you can decide if you want to pay this exact author. But Medium is a platform. And it’s weird: why should you pay money to some platform on which some author is publishing a seemingly free to read blog? Without the value clearly communicated, it feels like a weird form of extortion.
There might be something to do about it, like along the lines of “hey our crazy AI recommendation engine can find some articles on the same topic for ya, just for three bucks a month”. Okay, but why would you do that if you still have search engines and stackoverflow? Where’s the value? I don’t know vOv
An afterthought: there are platforms with “bunch of bloggers” that we can pay money to, and they are called old school newspapers. But the implication remains: with the newspaper, you know what they do and what you pay money for. What does Medium want to get paid for without being a traditional newspaper, with all the things like editorial policy, journalistic standards (whatever they might be) and so on? To me it’s unclear.
Based on what the CEO said, it sounds like they want to get paid for a recommendation engine that surfaces quality, substantial content. Upping the signal-to-noise ratio. Where traditional newspapers have editorial policy and journalistic standards, you could see that as an authoritative "these are the rules and follow them". I could see medium attempting a more 'free market system' where writers are incentivized to write quality content by getting views and money and the snuff sinks to the bottom. Essentially taking a more 'automated system of incentives' approach to try and fill that newspapery niche.
Especially with the consideration that traditional newspapers typically don't have niche content like programming articles.
> That’s probably a chicken/egg problem, and I don’t know how I would even begin to untangle it if I were you.
Paywall half of the article so the reader gets an idea on whether or not they want to pay to read more. This would also encourage impulse purchases.
> Most of their job to start would be to build a regex list of article title structures which are overly cliched, so that those articles get de-merited by the recommender algorithm
Categorize articles. There's no need to make cliched articles impossible to find for the cliche-interested reader.
As far as fixing it, the obvious first step is to at least be more generous with the article count... and don't show it at all until someone has seen at least 5 articles on separate days that month. Count a session clicking through a few articles on the same day as at most 2 article views (maybe have a separate limit of something like 5 articles in the same day). Don't be a nag trying to upsell when people haven't even had a chance to get interested yet.
That at least would cut down on the sleazy feeling of the whole site. An easy next step would be hiring someone as an editorial curator. Most of their job to start would be to build a regex list of article title structures which are overly cliched, so that those articles get de-merited by the recommender algorithm
Another way to feed a meaningful recommender: Allow users to tag posts (and other more clever things like deriving tags by looking at, e.g., which sub-reddits link to them). Build user profiles that are similar to how StumbleUpon had a "Stumble DNA" for users.
That's what I would be doing, anyway.