Follow Inflation: the circle that digital media needs to square

This piece first appeared in my newsletter, Future Proof. Please subscribe to that to support my work and make me feel good about myself.

What constitutes a lot of followers?

I remember when I first started to believe that social media might play a role in my professional success. This was back in maybe 2010, before I’d started at university, when I was trying my hand as a film blogger. The username I had on Twitter — @theclapperbored — was carefully cultivated to build my brand. I remember the moment when I passed the threshold of 1,000 followers. I felt like Hannibal crossing the Alps, Caesar fording the Rubicon.

More than a decade later, my perspective on that figure is both altered and the same. On the one hand, a thousand followers to an account where I tweet insightful things like “Really enjoyed Captain Phillips — what an actor Tom Hanks is!” feels almost disproportionate. But, at the same time, it’s starkly apparent that 1,000 followers is not a marketable audience any more, in any way.

Perhaps the most common question I get asked, as a podcast industry watcher, is a simple one: how many listens does a podcast have to get before it’s profitable? I then reel off a series of caveats (depends on the hosts, depends on the audience, depends on the geography, the topic, the production…etc) before arriving at my preferred figure: 20,000 listeners per episode. That’s 20,000 listens per episode, before you can begin to make a meaningful ROI on your show.

When I say that to sage media bods they nod in that sage media bod way. Because 20,000 listens is nothing when you’re used to dealing with the millions who tune into top public and commercial radio. When I mention that figure to anxious indie podcasters, they grimace in that anxious indie podcaster way. Because 20,000 listens is a huge amount of traction to require before you can become profitable. How does Joe Nobody, from the Arse End of Nowhere, convince 20,000 people (the capacity of Turf Moor, until recently a Premiership football ground) to listen to their podcast?

And the answer has tended to be a cheap one: build a following on social media. Where podcasting has generally experienced a deflation in listener figures over the past few years (the overall rate of podcast consumption has grown, but no faster than the rate of serious projects have been churned out; the result is a per-show deflation in audience figures) social media has been enjoying a boom time. Look at the chart below, showing total Instagram users over the past five years.

That’s almost 50% growth from 2020 (and bear in mind that for many digital media businesses — like every streaming service — 2020 is used as a high watermark because of the way that covid impacted consumption). And the story is even more pronounced on other services. Take TikTok for example, which is still the ‘new kid on the block’ in terms of a mainstream rival to Meta and Twitter. Here’s what their growth has looked like over the past few years.

TikTok’s growth is particularly interesting because it highlights another issue here. TikTok’s growth peaked during 2020, with an estimated 286m downloads in Q1 and 270m in Q2. That’s compared to growth of just 145m in Q3 last year, its weakest quarter for growth since Q4 of 2020. But even with new user creation slowing, the headline figure — the total TikTok users — is rising. Even when you have a year like 2020, which seems massive growth in the first half and limited growth in the second half, and you think that maybe (just maybe) you’ve reached the entirety of your plausible audience, the app continues to grow.

Some of this is accounted for by onboarding new users at either ends of the demographic spectrum. When TikTok launched it was aimed squarely at Gen Z — over time, it attracted a new base of younger millennials who wanted to keep up with the hot new thing. Then there are the younger people who — duh — get older, and become new digital citizens. This is the narrative that the TikTok Corporation would like to sell: that the pool of TikTok users is becoming ever broader. Access to TikTok from new markets, as the app spreads globally is another major selling point (at present TikTok is inaccessible in very few territories, despite persistent threats from huge jurisdictions like the USA and India).

But the other impact on the growth is the creation of duplicate accounts. Ask any teenager you know any they might confess (knowing teenagers, they probably won’t, actually) to having duplicate accounts. A primary, a spam, an alt — these are all words that mean something to the Gen Z mind. If you create a system where users have 2, 3 maybe even 4 accounts on the same platform, you are necessarily going to enter a period of inflation. From an outside perspective, there appear to be almost 2bn people using TikTok — but what if that’s actually 1bn unique users? How does that impact the influence, the financial potency, of the service?

As you get more and more accounts on all these social media platforms (and even those with fairly stagnant growth are experiencing some growth, which suggests that people create new accounts with greater frequency than they delete old accounts) a secondary question emerges. Not the one that I started this piece with — how many followers is a lot? — but the inverse. What’s a normal amount of accounts to follow?

Instagram has a cap of 7,500 accounts that you can follow at any one time, which is, obviously, insanely high. Twitter has no meaningful cap that I know of. TikTok’s cap is 10,000. This means that a follow is a relatively cheap unit of interaction — you don’t need to worry about following the latest hot young Samoyed to hit the streets of Bushwick, because you’re not going to run out of potential follows. Instead, the question becomes, how many followers becomes unruly?

The old wisdom of social media said that 1,000 accounts was the maximum any sane person would follow. More than that, and you started to get deluged with so much content that, on a linear timeline, you would never see the good stuff. But you’ve spotted the twist here: all these apps have done away with the linear timeline. In part, I suspect, that was to drive this very trend. Now, algorithmic timelines mean that a) you see any high performing content regardless of when it’s posted, and b) you’re going to be introduced to a lot of non-follower material anyway. The latter means that you’re endlessly being introduced to new accounts; the former means that it doesn’t really matter if you follow them.

Data on this is hard to gain but anecdotally and based on the logic of my own brain, I feel like the average number of accounts followed (on Instagram, TikTok, Twitter, wherever) is increasing. People follow new accounts with far greater alacrity than they move to cull old accounts.

At the margins, this is probably pushing up the true average number of social media followers by maybe 10% year on year. Which doesn’t really matter. If you have 300 followers one year and 330 the next year, you’re not suddenly going to get signed to promote Snapple. But at the other end of the social media spectrum, where the accounts built for commercial purposes live, it really does matter. Take my old film blog — would 1,000 followers be a figure worth celebrating today? Fat chance.

I’m going to go on Instagram right now (you can’t see me but you’ll have to take it on trust). I’m going to open up Reels and tell you the accounts, and followings, of the first five reels that appear. Reel #1: kbfilms.92, content creator, 13k followers. Reel #2: edjonesuk, actor, 103k followers. Reel #3: ohwhataknight_, hiker, 72.6k followers. Reel #4: spawnranch, memes, 185k followers. Reel #5: this_humble_homestead, strange Christian trad-wife, 47.4k followers.

Yes Nick, you’re all screaming. We’re perfectly aware that algorithmic, doom-scrolling social media outlets are a swamp of nonsense, repurposed, stolen or derivative content. We know that people are desperate to monetise their dogs or their houses or their husbands. But the weird thing is, you can quite easily gain a vast audience doing so. Some of these meme accounts have into the millions of followers. Take @kirawontmiss on Twitter, an account that the algorithm feeds me about 20 times a day. This anonymous account — with an avatar that looks like Jimmy from The Fairly Odd Parents, but presumably isn’t — has 1.7m followers. One point seven million. For reference, the Financial Times has 5.9m, CNBC has 5.2m, The Guardian News has 3.9m and the New Statesman, the Spectator, Prospect etc all have way under a million. Foreign Policy, that august magazine of global affairs, has 1.2m, just half a million fewer than kirawontmiss.

I know people who have social media accounts that serve as terrific showcases for their work. They gain really good followings, often well into six figures. And they still find it impossible to monetise that platform. A million followers on Instagram is no longer a sensational figure, the sort of figure that would get you invited to elite parties and have brands biting your hand off just to get a sniff at that audience. A million followers is the sort of following you can get with a poorly trained Dalmatian, a boxset of The Office DVDs and some screengrabbing software, or a Brazilian butt lift. A million followers isn’t cool. You know what’s cool?

I no longer know what’s cool. But what’s clear is that anyone building a brand, particularly a media brand, in the current environment needs to understand that they are operating in a landscape of grotesque follower inflation. This has two down-steam impacts. Firstly, it is much harder to impress advertisers with social media followings today than it was five years ago. Secondly, the attrition rate of followers, when trying to migrate from one platform to another, is huge. That falloff is even starker when you’re trying to monetise them. Ask anyone who has tried to set up some paywalled content around their subscriber base — if you can convince 1% of an unpaying audience to become a paying one, you’re doing brilliantly. (The exception is OnlyFans, of course).

This whole piece has been written without mentioning bot-farms, the banks of computers secreted away in the Balkans or India, churning out new users for nefarious purposes. The assumed average of fake followers on Instagram is said to be anywhere from 5% to 30% (a uselessly huge range). But anyone who’s used Instagram or Twitter in the past year will be instantly familiar with the concept of “porn bots”, AI-trained photographs called things like “SELENA XXX” or “Haileeeee4uuuu” who slide into your DMs and, at best, try to sign you up to an OnlyFans, and, at worst, try to extort you.

The impact of follower inflation is something we will see in the next year or two, as major advertisers wise up to the levels of crap they’re paying for. We will also likely see a further stratification where the required audience for commercialisation reaches some insane figure (maybe 2m, enough to fill Turf Moor a hundred times). We will see some platforms, like Elon Musk’s X (which I’ve called Twitter maybe a dozen times in this piece), attempting to reward grassroots creators through monetisation programmes that only reward more and more spam. And we will see even greater difficulty in attempting to convert social media users to non-social media users. Social media platforms have already made clear their intention to, algorithmically, trap people on the app by de-preferencing external links. But what happens when people follow so many different accounts that they are simply never engaging with any singular brand for long enough? How do you convince someone to read, let alone pay, when their touchstones only last for a second at most?

The reality is that it is in the interests of both social media and the broader media to work out a better way to manage follower inflation. This doesn’t just mean killing off the Russian bot factories. It means, firstly, encouraging people into a more rational number of follows, in order to bolster engagement with the people they would naturally reach out to (whether that’s grandma in Cincinnati or Loewe). Secondly it means more aggressively killing off defunct or duplicate accounts, even if that means that you see a net reduction in total follows (a hard sell for any company to make in the arms race for new users). Because if social media doesn’t get its house in order, it will end up with a product that is so thinly spread, so questionably verified, that the value to advertisers will diminish radically. The value of data too — that crude oil of the internet era — will also be challenged in a world of replication and redundancy.

So, what does constitute a lot of followers now?

The truth is, that question is harder than ever to answer. But the word ‘follower’ feels wrong. You can’t follow 10,000 accounts or even 7,500 accounts. That’s not ‘following’, any more than I ‘follow’ the cars that come speeding up my road at 60mph and disappear into the London sunset before I can take down their license plate.