The Creator-Influencer Economy pt 1: looking back on the first wave – bloggers

The total addressable market of the Creator-Influencer Economy could double by 2027 to $480 bn, according to Goldman Sachs Research. The forecast made in 2023 projected the Creator Economy to roughly double in five years.

I’ve been mulling the Creator-Influencer Economy, and I’m still rolling this over in my head so file this newsletter in the category of me thinking out loud. As I have seen several times over, a lot of the analysis centres on how screwed legacy media is because they are now competing against an army of individual creators and news influencers across a range of platforms including TikTok, YouTube, Twitch, Substack and more. Yes, there is that, but having been to this rodeo before, I want to think about this in two ways:

  1. The economics and dynamics of the Creator Economy itself.

  2. The business model pivot legacy media would have to make to take advantage of those dynamics and the pros and cons of some of the most obvious approaches.

As I started to think about this present moment of disruption in media, I started to think about previous chapters of digital creators, which created new markets and destroyed old ones. How is the current wave of fragmentation and distribution different from earlier ones? Creators have always been a part of the internet, and casting a look back reminded me of how early pioneers experimented with ideas, including commercial models, that have now become the foundation of the Creator Economy.

The Long History of the Creator Economy - JenniCam and the rise of the bloggers

Digital technology has been challenging and remaking traditional media almost from its inception, and I’ve been covering it and living it in various roles now for more than three decades. Yeah, I’m that old, but this isn’t a shake-my-fist-at-the-sky moment to tell the kids to get off my lawn. (Although, yes, I have those moments.) It is to say, I’ve watched the technology and market evolve through various iterations.

People forget about lifecasting and blogging communities like LiveJournal from the 1990s. Not long after I met my wife in 2006, I remember This American Life rerunning an episode from a decade before, in 1997, talking about experiences they were having on the Internet, which they wondered if they could have anywhere else. It highlighted the story of one of the biggest lifecasters, Jennifer Ringley of JenniCam fame. She started streaming pictures from her dorm room when she was a 19-year-old university student, and at the peak of its popularity, her site received 7m visitors a day. After graduating, she offered a paid service with a higher refresh rate. Voila, a business model that would seem quite familiar in today’s world of camgirls, although I’m hesitant to refer to her as that.

Another chapter in this story was the rise of blogs. In 2024, the US presidential candidates courted influencers who have podcasts, YouTube channels, newsletters and massive TikTok followings. In 2004, with the initial success of Howard Dean, bloggers rose to challenge the establishment media in setting the political agenda. They had a prominent place at the Democrat’s nominating convention that year. Not only did I cover the phenomenon for the BBC, but I ‘blogged’ during the US political conventions and then the final months of the 2004 elections. It was one of the first forays by the BBC into blogging, although the comments had to be taken in via email and then carefully curated. In 2005, I came to the UK to research and write a white paper on how the BBC should respond to revolution in media.

It was a heady time when the lines between traditional media and the audience blurred. I launched a segment on Radio 5 that looked at the news agenda through the lens of blogs and podcasts (and later social media), and then I was on the launch team of the World Service interactive programme World Have Your Say. I often brought bloggers and podcasters, including milbloggers - members of the US military blogging and podcasting from Iraq and Afghanistan - onto the programme.

Much like now with creators, people talked about the democratiisation of media and a force that would “challenge traditional media”. In 2007, Julian Robins wrote a paper as a Reuters Fellow at Oxford University. He wrote:

The blogging revolution comes at a time of turmoil for the traditional media, as digital technology changes the way people consume news, information and entertainment. Audiences and readerships are falling, along with profitability. Many mainstream news organisations have rushed to embrace user-generated content as they seek to fine-tune their business model for the new age.

As I was writing this, I thought that I wouldn’t call the current phase of creators democratisation. It’s a marketplace of attention and commerce. While yes, news and political influencers are part of the mix, it’s also about technology, fashion, food and so much else.

The fall of blogging and the rise of a new kind of social media

But blogging faded, although one might argue that newsletters are the new blogs. But blogs and podcasts faded, although podcasts would enjoy a renaissance. The decline of blogs was down to several factors. One issue was discovery, which still bedevils new podcasts. Technorati was the search engine for blogs, long before Google tried to create a subset of its search. Google launched its blog search in 2005, but shut it down as a standalone product in 2011. However, apart from search, some fundamental technology that connected the distributed network of bloggers broke, and the fediverse was a long way off. Note, I am sceptical that most people will embrace distributed networks. Those who do choose to accept a higher degree of friction for what they perceive as freedom, but that’s a whole other discussion.

However, I think what caused the decline of blogs and explains the rise of everything that came after was network effects. Bloggers used a system called trackbacks to see who had linked to them, and it was often via those links that helped bloggers discover each other. Trackbacks became overwhelmed with spam and ceased to function effectively. Comments were another way that bloggers interacted with each other, but these comments were spread across blogs. It was a highly distributed conversation, but the comments and links helped motivate bloggers to keep on writing. I remember what a thrill it was.

Of course, the social media platforms quickly eclipsed blogs. Why?

  • Powerful network effects. It was much easier to find your friends.

  • Simpler engagement mechanisms including lightweight likes and easy commenting.

  • Algorithms that helped you keep in touch with friends and topics you were interested in. (Although we can debate how effective they are now.)

  • Masssive, centralised aggregators for attention.

That last point is key. Grassroots bloggers prided themselves on their distributed nature. It reminds me of the quote from the film Star Wars: The Rise of Skywalker, “That’s not navy, sir. It’s just…people.” However, the attention was spread across thousands of blogs like stars across the sky. Bloggers revelled in the distributed nature of their conversation, but it wasn’t to last and couldn’t have lasted in the face of what came next.

Social media networks captured that attention and aggregated it. They provided much more efficient markets for advertisers, and that helped support a new generation of creators. Attention shifted to these platforms and so did billions of dollars of revenue globally. It drained attention and revenue from traditional media, particularly journalism. But that’s the next chapter.

The AI imperative for media

The biggest news in AI overall was the release of DeepSeek, a disruptive Chinese distilled AI model. It leveraged other models to beat OpenAI and Anthropic by some measures of AI progress, AI expert Karen Hao told the Reuters Institute. Her analysis is worth reading. She said DeepSeek has triggered a long-standing, but dismissed, challenge to the dogma that AI progress needed more resources rather than a more efficient use of resources. It’s an important read. Of course, with this being a Chinese disruptor of America’s perceived dominance of AI, it has political overtones.

While creators are part of the current chapter in media, AI also dominates the innovation agenda for journalism organisations. Brian Morrissey is not impressed. “Unlike past new technologies, publishers have taken a far more cautious approach to using AI beyond the basics or for little skunkworks projects that are unlikely to have much impact. The examples of publishers using AI are usually confined to small-scale frippery like AI-enabled internal search.”

I think he’s being a little harsh. I saw a demo of a tool The Economist is developing to scan academic papers for possible story ideas and came away impressed, and the Nursing Times use of AI to power a question-based way to access the intelligence of its archive is smart, especially how it is connected to their subscription model.

INMA’s Sonali Verna looks at the strategic frameworks that media companies could use to maximise the impact of their AI efforts. At the end of the day, I think AI should be deployed to do what it is uniquely good at and solve pain points for audiences and editorial teams.

The Independent is showing the sophistication in their approach as they grow as a digital English-language news brand. I have been fascinated with their A2K - anonymous to known strategy in the UK. They have 11 different ways to capture first-party data from their audiences. In the US, they are smartly scaling to stake a claim to that market. They have a lot of lessons to teach other media companies on how to survive in this highly competitive, rapidly changing media market.

How ESPN is getting smarter with its Snapchat strategy

A hand holds a cube with the Snapchat logo
Snapchat User, by Blogtrepreneur, from Flickr, Some Rights Reserved

Hello dear readers, in today’s international media newsletter, the story that stood out for me was one looking at the success that ESPN is having with Snapchat. Snapchat, you remember that app, right? The one that Instagram has copy-pasted feature after feature?

As I often say in social media training sessions, Snapchat is mostly pointless for media folks – especially local media folks or media operations outside of the US and UK. The caveat to that is apart from a handful of verticals like fashion, or I guess sports.

As this Digiday+ ($$$) article points out, ESPN is running against the trend. The New York Times has ‘paused’ their channel, but ESPN is finding that their Snap efforts are helping drive viewers and revenue.

In addition to that, a few other headlines in today’s newsletter:

If you haven’t subscribed you, you can do so on my Nuzzel profile page, and if you spot something, let me know on Twitter. I hope to get comments working again on the ‘ole blog here, but that is a work in progress.

How to get onto Instagram’s Explore tab

Exploring, by Tom Bullock, from Flickr, Some Rights Reserved

Hello new newsletter subscribers! My how your numbers have grown.

Topping today’s international media newsletter is a great summary from TechCrunch on the signals that Instagram uses to put content on the new Explore tab.

At the public media group where I work, we’re seeing some early indications that Insta is helping us reach parts of our community that we want to serve but we currently aren’t connecting with. For instance, we recently ran a series about African-American women who had suffered trauma in their lives and how they received support. Our posts on Instagram took off, while they didn’t get much traction on Facebook, which is the opposite of what we normally see.

TechCrunch says that Explore will match content with topics and accounts similar to what an Insta user already follows. Videos and highly visual stories without much text will also have a higher chance of getting on the Explore tab. It’s a great post to bookmark.

Other topics include:

And that’s a wrap for this week. I’ll see you on Monday. If you don’t already subscribe to the newsletter, you can on my Nuzzel profile page. And please, please send me stories, @kevglobal on Twitter, especially outside of the US. Kevglobal really is global.

TikTok Explained and an Indie Media Collective in Boston

An unsettled cat on TikTok, Giphy.com

How much have you been hearing about TikTok? It has exploded on the scene, and everywhere I turn, I seem to hear about it. What is TikTok? NPR has a really good Q&A about this latest, hottest social media platform, and it is especially popular with really young audiences. My takeaways:

  • Facebook is for Grandma.
  • Snapchat? “Over it.”
  • ” TikTok is kind of a culmination of every viral video app that’s existed in the last five, maybe even 10 years – kind of going back to early YouTube. And it’s comedy. It’s music. It’s sometimes makeup and sometimes monologuing. And it’s a weird hodgepodge of every sort of viral thing that could happen in a short-form video app.” says Brittany Spanos of Rolling Stone magazine.
  • ” But when I showed TikTok to my 20-something producers, they thought that they were too old for it. It feels distinctly for the very young,” Spanos said.

Moving on from TikTok another feature in my newsletter today, I also found a lot of great insights in this overview of Boston Institute for Nonprofit Journalism . Chris Faraone, one of the co-founders, described BINJ like this:

Basically, I wanted to create a miniature ProPublica, a collaborative, free-floating incubator that would be a Make-A-Wish Foundation for all of these news outlets to help them do what they couldn’t afford to do.

Chris Faraon, co-founder, Boston Institute for Nonprofit Journalism

This is a really good, actionable look at the kind of support that this group is providing for local news producers. It is giving small local publishers some room to experiment and also building deeper connections to users. The other thing that really stood out for me what that the institute helps small indies in choosing what they cover and what they don’t cover.

As usual, if you’ve got a story that you think would be good in the newsletter, let me know on Twitter, @kevglobal. If you haven’t subscribed to the newsletter yet, it’s easy to do so here.

10 Journalism Newsletters You Should Subscribe to, Make that 11

Pamphleteer, WikiMedia Commons

I’m closing out this week in a totally meta way in my newsletter: 10 other journalism newsletters that you should check out, well, apart from mine.

But I also want to start something and post the top five stories based on what you have been clicking through to in my newsletter.

  1. Like most media, podcasting is pivoting to paid (with complications)
    From Max Willens, Digiday
  2. 7 reasons a freelance journalist should start a podcast, by my friend
    Suchandrika Chakrabarti, on Muck Rack
  3. Why platforms like Facebook and Apple struggle to boost local news | What’s New in Publishing | Digital Publishing News, a great piece by Simon Owens, on What’s News in Publishing, where I also have been known to write.
  4. The Telegraph’s roadmap to 1m paying subscribers and financial sustainability, by Ian Burrell in The Drum
  5. How publishers are using Snapchat’s curated stories tool for breaking news and more, by Kerry Flynn, in Digiday

I hope that you have a great weekend, and remember, if you have any good stories that I should include in the newsletter, let me know @kevglobal on Twitter.

Can the ‘Wisdom of the Crowd’ fix disinformation?

eam members assemble a puzzle during the problem solving phase of the "Whacky Relay" at the base track 2 May.
SCHRIEVER AIR FORCE BASE, Colo. — Team members assemble a puzzle during the problem solving phase of the “Whacky Relay” at the base track 2 May. (U.S. Air Force Photo/Dennis Rogers)

Thank goodness it’s Good Friday. In today’s newsletter, I highlight a project from my friend Claire Wardle that she hopes will allow researchers a new tool to fight disinformation. She is currently the director of Civic, the Coalition to Integrate Values Into the Information Commons , and she has proposed a project that she called:

“a Wikipedia of Trust,” a back-end contributor model where regular people could volunteer to flag, decipher, and catalog fake memes and bot activity, and add crucial cultural context to images and information that might be a zombie rumor.

“A Wild Plan to Crowdsource the Fight Against Misinformation”, Wired, by Emily Dreyfuss

It is the information equivalent of capturing virulent memes for study. The challenge as Claire highlights is that disinformation is moving from being out in the wild on the open web – or what remains of it – and public social networks to places hidden by algorithms and inside messaging platforms where they can spread in ways resistant to observation and rebuttal.

I also highlight CNN’s use of a new Snapchat service, called Curated Stories to cover breaking news. The tool has helped lure CNN back into Snap’s Discover. Will this be enough to stem Snap’s slide? I doubt it because what might be useful for media won’t necessarily address the fundamental user issues that Snap has.

Apart from those two big stories, we also have:

The ad tech bubble may be about to burst. US newspapers are struggling for cash in the rush consolidation. Sift launches ‘news therapy’ app. New media investment fund for LatAm. UK Telegraph aims for 1m subs.

Have a great weekend, and I’ll see you next week.

If you spot a good story about the business of media, especially digital, feel free to send it to me @kevglobal on Twitter. If you don’t get my international media newsletter in your inbox, you can get a taste of it and subscribe here

Zuckerberg’s Magical Garden of Horrors Has Another Bad Day

The statue of a dinosaur eating a smaller dinosaur.A metaphor for yet another bad day at Facebook.

A dino has dinner. A metaphor for yet another bad day at Facebook. Photo by Mike Bird from Pexels

Pick your adjective or metaphor when it comes to Facebook’s current run of horrible, awful, no good press. And this isn’t just an optics or PR thing. Facebook is embattled because it has:

  1. Screwed up, repeatedly.
  2. Can’t or won’t, or a mix of both, seem to fix its problems.
  3. And is in a footrace with Trump’s West Wing in terms of a petty, backstabbing leakfest.

In the newsletter today, we highlight a couple of stories yesterday that make for yet another shitty day in Zucklandia. Wired dropped a story with 65 sources talking about “15 months of fresh hell” at the anti-social behaviour network. And NBC reported a leak of thousands of documents that show that Facebook used data to reward its friends and punish or at least kneecap potential competitors. The second story is the kind of stuff that would instantly see Facebook in anti-trust court if the US actually enforced anti-trust laws anymore. Oh Europe, where are you when the world needs you?

But that’s not all today, we also have:

Scandi publishing giant Schibsted joins complaint against Apple’s app dominance. Vox Media acquires publisher with history of turning journalism into movie deals. Publishers turn to ‘expert networks’. UX lessons still be learned from print.

If you spot a good story about the business of media, especially digital, feel free to send it to me @kevglobal on Twitter. If you don’t get my international media newsletter in your inbox, you can get a taste of it and subscribe here

 

Are Facebook ads good value for money?

I’ve never used Facebook to advertise anything to do with Ada Lovelace Day, but I thought I’d give it a go with a post about Ada Lovelace Day Live, just to see what happened. I assumed that FB would be quite effective at delivering my post to a large, relevant audience, but that’s not what happened.

When I set up the ad, FB said I’d reach 2,700 – 7,200 people, but in fact it only reached 1,588. The idea that FB somehow can’t find 2,700 people in the UK, over 7 days, who match my audience profile (ie, graduate or higher, in all the STEM-related fields they have) is absurd. Indeed, FB itself says that the potential size of my audience is 28 million, but it couldn’t find more than 1,588 people. Sure. Right. I totally believe that. *cough*

Of those 1,588 people reached, 25 “reacted” to it (ie used the like button), two commented (one of those comments is a guy being an asshole), five people shared it, and six people clicked the link.

These are not particularly impressive figures to me.

Now, I know that I only spent £10, but I run ALD on a shoestring, and that £10 was a test to see if it would be worth spending more. Frankly, I can’t say that I’m confident that it was even £10 well spent.

I had assumed that FB would be a cost-effective way to reach lots of people, but at £1.67 per click, I don’t think that’s the case at all. Frankly, it feels more like a rate-limited con than a useful service.

Kevin is more sanguine than I — he thinks 1,588 is good for a low-follower page (we have 124 likes on our page), and he has more experience than I do with the way that FB works. However, the point is that the whole reason for paying for an ad is because our page has few followers, and because FB has destroyed organic reach in order to force us to pay to reach more of the people we previously would have reached anyway.

But they’ve done a shit job of up-selling, because I would have invested £100 in ads, and would have expanded my ad horizons to include merchandise and similar if FB had delivered on this test. They didn’t deliver, so they’ve lost a potential advertiser and, sadly, I’ll have to just battle on and try to grow my organic reach.

This is a huge shame. The promise of social networks was that it would level the playing field, and that the smallest organisation or the least famous person still had the opportunity to reach hundreds of thousands, if not millions of people. There is, of course, now a huge issue with noise which didn’t exist at the dawning of the social media age, but that’s not the problem with Facebook.

The problem is that they have deliberately locked small folks out of building reach organically in order to drive ad revenue, but are not providing good value for money when people with limited resources pay a small amount for ads. Had they delivered even the lower end of their estimated reach range, I might have considered investing more. Had they delivered 7,200 people, then I certainly would have, even though I still think that’s an artificially low number given that they said my audience is 28 million.

What rankles most is that not only is there no good reason for limiting ad reach this severely, but also that it hurts the very people that social media was suppose to help: those of us stuck in the long tail without the resources to spend loads on advertising.

How Tor failed Social Media 101

There are some companies that appear to be native to the web, not just on the web but of the web. Often these companies were early adopters, building websites whilst others called the web a ‘fad’, starting blogs before most people knew what they were, and using social media in a way that makes them appear to have a sound and thorough understanding of the medium. Tor is one such company, but sadly, it has recently become clear that Tor does not actually understand social media and, in particular, has not developed or adhered to a crisis communications policy. 

The short story is that some month ago, two overlapping groups of people calling themselves the Sad Puppies and the Rabid Puppies campaigned to game the Hugo Awards, in which both nominations and awards are via a popular vote. Whatever one thinks of the people in and supporters of these groups, it is fair to say that they are engaging in what one might call grievance politics. Certainly there’s also an awful lot of identity politics involved, so temperatures on all sides are running high. If your’e not familiar with the backstory, a quick google will provide you with a wide variety of opinions on the Puppies, their politics and their activities.

What I am specifically concerned about, and why I’m disappointed in Tor, is their reaction to a complaint from one of the Puppies about a comment made by a Tor employee, Irene Gallo, on her personal Facebook page. Rather than taking a considered approach, Tor threw their employee under a bus, and appear to have broken every rule in the crisis comms rulebook. It’s sad to see that a company that in many other respects really gets the web, fails to understand how to manage the fallout from an online furore. 

Note: I have no insider knowledge of what went down at Tor, I only have their public statement to go on, but that in itself tells me a lot about what probably did and didn’t happen. 

1. Consider the situation 

The first mistake Tor appear to have made is that they did not fully understand and consider the situation. There are several aspects to this situation that raise red flags and call for especially careful handling of the response: 

Any one of those issues would flag a complaint as requiring careful thought, but all of them together add up to a warning to tread incredibly carefully indeed. I don’t think Tor did that. The wording of Tom Doherty’s blog post in response to the complaint is clumsy and ill-considered, and shows no signs of having been properly thought through. 

2. Take enough time, but not too much or too little

When the shit hits the social media fan, it is important to respond in a timely manner, but it’s even more important to avoid a kneejerk reaction. If an issue needs further inquiry before a full response is issued, then it’s acceptable to publicly acknowledge the complaint and say that it’s being looked into.

It may even be that no response is required – not every complaint is deserving of employer intervention. If an employee has a disagreement with a member of the public on her own Facebook page, it is possible that her apology on said Facebook page is sufficient, and that her employer need not step in at all. One can debate whether that was the case here or not, but it is an option that should have been considered, along with all others.  

Doherty’s response reads very much like a kneejerk reaction. it is, to all intents and purposes, a public disciplining of Gallo, which is entirely inappropriate no matter what Gallo did. If you address a complaint, you do not use it as an opportunity to shame your staff. Doherty should have taken more time to think about exactly what was going on and how his post would be read by the broader Tor community. 

3. Remember there are three sides to every argument

Any public response to a public complaint is made more complex by the fact that there are three parties involved: You, them, and the audience. In his rush to appease Gallo’s critics, Doherty appears to have forgotten that he might also anger people who agree or sympathise with Gallo, or who do not believe that the complaint against her has merit, or who, after reading his post, believe that the complaint has merit but that his response was inappropriate, etc. 

In chastising Gallo online, Doherty has alienated a lot of people, and that in and of itself is a massive failure for Tor that Doherty himself should be disciplined for. You simply do not rush in with a response that inflames the situation, especially when it’s obvious from the beginning that tempers are running high and offence is being easily taken. Indeed, the taking of offence is a key weapon in grievance politics, and Doherty should have both realised there was a major risk that his response as written might make the situation worse rather than better. 

4. Talk to your employee, work with them on both your response and theirs

Whether or not you agree with Gallo’s initial comment, it is clear that there was insufficient conversation between Gallo, Doherty and others at Tor about how best to deal with the situation. Gallo’s apology has been deemed a ‘fauxpology’ by some, and I can see how they would reach that conclusion. The key line is “I apologize to anyone hurt by my comments”, which might have been more appropriate worded as “I apologise for saying something offensive”. If you’re going to apologise, swallow your pride and do it properly and graciously, even if you feel you shouldn’t have to, and be very careful to avoid any wording which can be interpreted as shifting the blame on to those offended.

But equally, Doherty, does not appear to have discussed his response with either Gallo or anyone else who might have pointed out that it reads very poorly. When you respond to a complaint, you do not need to defend the complainant, you simply need to address the substance of the complaint, where it is valid, and explain if necessary any parts of the complaint you have concluded are not valid. Doherty did not do that. 

So what should Tor have done? 

There are two things that Tor should have done, and that all companies should do right now, if they haven’t already: 

1. Draft an employee social media policy

Work with their staff to draw up a social media policy, governing appropriate behaviour online. This policy should not have the effect of chilling speech, so it absolutely has to respect the fact that employees need to have their private spaces online. But it should discuss how to protect those spaces, and how to think about what can be said publicly and how to think through the potential fall-out of controversial statements.

A social media policy should also tie in to standard disciplinary procedures, so that staff are clear on what would constitute a serious transgression that would invoke that procedure, and how it will play out. Social media is not special or different, so should always follow standard HR procedures. Staff should never, ever, be chastised in public, and that this happened is a failure of senior Tor management that needs to be addressed.

2. Draft a crisis communications policy and procedure

When something goes awry, it’s essential that people across the company know what to do, who to talk to, and how to minimise the impact. Doherty has not done a single thing that I would recommend a company do, and instead of soothing ruffled feathers, he has inflamed the situation and alienated core customers. 

A crisis comms policy should again be drafted with staff, discuss the kinds of issues that can crop up, particularly the different between external crises, where an event outside of the company’s control causes a problem, and internal crises such as this one, where staff members says something without giving it enough thought.

There should be a chain of command, so that everyone knows how to escalate a problem, and there should always be two pairs of eyes on the response, particularly in small businesses where it’s easy to feel personally attacked and thus to overreact when things go wrong. There should be guidelines on how to properly respond, what to say, what not to say, how to formulate a reply that addresses the facts and not the emotion of a complaint, and when not to respond at all. 

I find it unlikely that Tor has such a procedure, given what’s just happened, and that again is a failure of senior management that they need to address, urgently. 

Note about comments: I am travelling at the moment, and because all comments are moderated there may be a delay in approval of your comment, should you choose to leave one. Abuse, rudeness or any incivility will simply result in the comment being deleted. Repeat offenders will be banned. I am not interested in a discussion of the Puppies or their politics, so those comments will not be published, along with any other off-topic comments. For the sake of clarity, on-topic comments are those about crisis comms and social media. 

Can comments withstand Google-scale communities?

Not long after I joined the Guardian as blogs editor in 2006, I was at an online publishers event in London. Forefront in my mind was how to build engagement at the scale that we would quickly find with Guardian blogs, a particularly important question given that this was several months after the launch of Comment is Free, which was already suffering from serious teething pains socially. I asked Tim O’Reilly how to scale community, and he said:

You have to stay small as you grow big.

Being at a conference, Tim didn’t elaborate on his Zen k?an, but I still ponder it today, especially as I read this: How Ta-Nehisi Coates built the best comment section on the internet—and why it can’t last.

Flipping back to 2006, I had recently done a lot of work in my previous job at the BBC regarding social media – blogging as social media was then – I wondered how well the social aspects of blogs would hold up when we threw the scale of audience at it that the BBC could generate. I foresaw some of the issues that BBC bloggers would have in terms of trying to deal with volume of comments that they would instantly face.

In traditional blogs, the core audience is a community of choice, a group of people that shared an interest in a particular topic. That community of commenters grew to have social connections, bonds that helped foster civility. At the BBC, our bloggers could quickly generate an audience, but that didn’t mean we could create anything like the thoughtful communities of choice that we saw with smaller blogs. Our commenters didn’t know each other and would often be drawn to hotly contested topics because of strong, often diametrically opposed opinions.

To make that kind of conversation work, you need people with a very different skill set than your traditional columnist, who often come at issues with a ‘what do I want to be outraged about today’ type of mindset. What you need in order to make comments sections really work is exactly the kind of mindset that Coates’ brings to his work, but that doesn’t mean that you don’t still face the challenge of staying small as you grow big.

The issue is that the internet rewards scale and some social strategies simply fall apart at Google-scale. Once The Golden Horde – the name given to Coates community of commenters – grows beyond a community of choice and starts to attract people wanting to vent about some of the most sensitive issues that we face today, the comments often become unmanageable.

Back when I was blogs editor at The Guardian, I was often asked how I measured success, and I said it was not by the number of comments. Any fool can generate comments. You simply pick the hot topic of the day, push people’s buttons until they bleed and then survey the wonderful wreckage of human outrage. That is why in 2015 the internet sometimes feel like one gigantic generator of human fury.

Outrage was the standard editorial strategy in the 20th Century, when the only return channel was analogue and scarcity of letters to the editor set a high enough bar to keep the Angry from Milton Keynes interplay to a manageable and highly controlled level. That editorial strategy does not work in the 21st Century with a completely open return channel for which the technical and social tools have not kept pace.

In this century, the big challenge is generating thoughtful conversation. I have tried it a number of times, to the point where commenters knew that I was building a digital media Skinner box: Rewarding productive contributions more than punishing transgressions and guiding the comments towards conversation rather than confrontation. It is hard work and it does not scale particularly well. I was blessed with the autonomy and time to do it, but it takes time and a lot of effort.

At the end of the of the day, maybe staying small as you grow big is like Google where you break up a team when it gets too large. Maybe the only way to make this work is to keep the communities and conversations small enough so that people actually develop the kind of social bonds that allow them to disagree with civility. But that runs counter to the gigantic scales that make a business sustainable on the internet.

So, what happens when the business imperative to scale comments runs counter to social strategies to manage them? Things fall apart. It’s that simple. They become unmanageable and eventually people burn out and the communities of choice leave, in search of some other space where things work on a more human scale. I’ve seen it happen time and again. Some people seem like serial early adopters, but what they are really doing is seeking a space in which they can find community.

As an editor, this raises all kinds of questions. But the one thing I would say is that while we all want the biggest audience possible, we have to accept that strong, highly-engaged community strategies operate on a different scale. Push them beyond their social bounds and they will simply fall apart.

Update: After thinking about, small communities of choice also tend to be rather homogenous. This begs the question on how we develop technologies and social techniques that allow for diversity of thought and civil dissent.