The Silent Switch

How is the silent switch impact­ing your content?

As a PR adviser since 2005 and a digit­al strategist since 2007, I’ve seen the silent switch take over — one seam­less iter­a­tion after another.

In this blog art­icle, I explain how social media algorithms silently have meta­stas­ised into a series of glob­al Pavlovian experiments.

Here we go:

The Silent Switch

All social media algorithms are built dif­fer­ently and are con­stantly being developed. At the same time, social media users’ beha­viours are evolving.

Still, there was a way that social media algorithms used to behave—and there is a way that social media algorithms behave now.

This has been a fun­da­ment­al but silent switch.

How Social Media Algorithms Used To Behave

For more than a dec­ade, social media algorithms would deliv­er organ­ic reach accord­ing to a dis­tri­bu­tion that looked some­thing like this:

This dis­tri­bu­tion of organ­ic reach enabled organ­isa­tions to use social media des­pite not being “media companies.”

How Social Media Algorithms Behave Today

Today, after the silent shift, social media algorithms deliv­er organ­ic reach more like this:

The increased com­pet­i­tion and soph­ist­ic­a­tion among con­tent cre­at­ors par­tially explain this new type of dis­tri­bu­tion. However, going vir­al is still just as pos­sible for anyone.

How does this work?

The Single Content Algorithm

How can a social net­work pre­dict what users will like? 

Content from a trus­ted cre­at­or trus­ted by a large com­munity of fol­low­ers used to be the lead­ing indic­at­or of future per­form­ance. But today, social net­works have found a bet­ter way to pre­dict con­tent success.

The single con­tent algorithm = when social net­works demote con­tent cre­at­or author­ity to pro­mote single con­tent per­form­ance to max­im­ise user engage­ment for ad revenue.

The single con­tent algorithm presents newly pub­lished con­tent to a lim­ited audi­ence sample size:

If the newly pub­lished con­tent tests suc­cess­fully, the social media algorithm pushes that con­tent to a slightly lar­ger stat­ist­ic­al sub­set. And so on.

This iter­at­ive pro­cess means that single pieces of con­tent worthy of going vir­al will go vir­al, a) even if it takes a longer time, and b) regard­less of the con­tent cre­at­or’s num­ber of followers.

Learn more: The Silent Switch

Why Subscriber Counts Matter Less

Like many oth­ers inter­ested in pho­to­graphy, I sub­scribe to Canadian con­tent cre­at­or and pho­to­graph­er Peter McKinnon on YouTube. And I’m not alone; 5,55 mil­lion sub­scribers fol­low McKinnon as I write this.

5,55 mil­lion is a respect­able audi­ence by any measure!

Still, it’s not uncom­mon for McKinnon’s reg­u­lar videos to acquire a few hun­dred thou­sand views. Don’t get me wrong: I’m not say­ing a) that a few hun­dred thou­sand views are noth­ing or b) that we should feel sorry for influencers.

I’m say­ing: A few hun­dred thou­sand views is a lot—but it’s not 5,55 mil­lion.

Instead of basing algorithmic dis­tri­bu­tion on McKinnon’s past per­form­ance (e.g., sub­scriber count, total views earned, etc.), YouTube tests each new video using the single con­tent algorithm.

If McKinnon almost always gets a few hun­dred thou­sand views, his videos typ­ic­ally pass that many iter­at­ive audi­ence tests before hit­ting the “algorithmic wall of bricks” (i.e. IF=failure).

It’s the silent switch.

Now, I appre­ci­ate McKinnon’s con­tent. However, his struggle to earn more views is… his.

In the grander scheme of things, at the level where it affects us all as online media con­sumers, the single con­tent algorithm has one pro­found — and rather ser­i­ous! — implication:

Earning trust over time gets the algorithmic back­seat, while sen­sa­tion­al pieces of con­tent get the roy­al treatment.

Learn more: Why Subscriber Counts Matter Less

Why Social Media Algorithms Are Basic

Earning many sub­scribers, fans, and fol­low­ers used to mat­ter a lot, but since the silent switch, it has mattered less. The old trust-based sys­tem has giv­en way to the single con­tent algorithm, which favours sensationalism.

But, wait!

Why must it be this way? With such pro­found advance­ments in digit­al tech­no­lo­gies (arti­fi­cial intel­li­gence, machine learn­ing, big data ana­lys­is, etc.), should­n’t social media algorithms be more soph­ist­ic­ated?

In an era of prom­in­ent social media issues (fake news, dis­in­form­a­tion, elect­or­al tam­per­ing, etc.), should­n’t social net­works like Facebook, Instagram, TikTok, LinkedIn, and many oth­ers be more inter­ested in favour­ing trust?

Instead, social media algorithms are basic. There are a few reas­ons as to why:

So, social media algorithms aren’t as soph­ist­ic­ated as they could be. They’re basic. For now, the single con­tent algorithm reigns supreme.

The single con­tent algorithm = when social net­works demote con­tent cre­at­or author­ity to pro­mote single con­tent per­form­ance to max­im­ise user engage­ment for ad revenue.

Learn more: Why Social Media Algorithms Are Basic

List of Social Media Issues

Social media isn’t just nyan cats and double rainbows. 

With massive change come new social media issues we must deal with.

Here are a few examples of social media issues:

Read also: The List of Social Media Issues

Internet: The Global Pavlovian Experiment

Many people are deeply con­cerned with how the social media algorithms work. They seem to pro­mote polar­isa­tion, fake news, dis­in­form­a­tion, fil­ter bubbles, etc. 

The pro­voca­tion is apparent:

“Social media algorithms are giv­ing us, the people of the world, only what our beha­viours tell them we want. We’ve evolved to care more about sen­sa­tion­al­ism over substance.”

Social media algorithms are slaves under the prin­ciples of clas­sic media logic—and act as amp­li­fi­ers of social media logic.

But if human beings are prone to sen­sa­tion­al­ised con­tent, why does it seem like social media algorithms are mak­ing mat­ters worse?

The answer is simple: It’s a self-rein­for­cing con­di­tion­ing loop.

Pavlovian Conditioning

In clas­sic­al con­di­tion­ing, psy­cho­lo­gist Ivan Pavlov made dogs sal­iv­ate by hear­ing the sound of a tri­angle. Every time he fed the dogs, he played the sound. Soon, he could play the sound to eli­cit the dogs’ phys­ic­al response (sal­iv­a­tion) — without the food. ¹Classical con­di­tion­ing. (2023, January 17). In Wikipedia. https://​en​.wiki​pe​dia​.org/​w​i​k​i​/​C​l​a​s​s​i​c​a​l​_​c​o​n​d​i​t​i​o​n​ing

In oper­ant con­di­tion­ing (also called instru­ment­al con­di­tion­ing), the strength of an exist­ing beha­viour is mod­i­fied by rein­force­ment or pun­ish­ment. ²Operant con­di­tion­ing. (2023, January 13). In Wikipedia. https://​en​.wiki​pe​dia​.org/​w​i​k​i​/​O​p​e​r​a​n​t​_​c​o​n​d​i​t​i​o​n​ing

Both meth­ods modi­fy beha­viour by con­di­tion­ing, and the dis­tinc­tion between these two approaches isn’t clear-cut. While sal­iv­a­tion can be seen as a “reflex,” want­ing treats des­pite not being hungry has ele­ments of being a vol­un­tary beha­viour. And modi­fy­ing beha­viour to seek rein­force­ment act­ively or escape pun­ish­ment can be more or less unconscious.

Tech giants like Meta, Alphabet, and TikTok have asked us to allow them to con­duct Pavlovian exper­i­ments glob­ally, and we have gran­ted them that permission.

It’s no coin­cid­ence that their terms of ser­vice agree­ments are lengthy and hard-to-read dis­claim­ers designed for scrolling past. At the same time, their algorithms pro­mote con­tent that can only be described as dopam­ine-indu­cing clickbaits.

And as a res­ult, we’re flooded by a vast array of hard-to-tackle social media issues. The phe­nomen­on is not nov­el. Neil Postman warned us that we might amuse ourselves to death in 1985!

It’s a Pavlovian loop: Human beha­viours amp­li­fy social media algorithms, and social media algorithms amp­li­fy human behaviours.

Love it, hate it. If you’re a PR pro­fes­sion­al, it doesn’t mat­ter.
If the play­book is chan­ging, so must we.

Learn more: Internet: The Global Pavlovian Experiment (to be published)

Enter: Social Media Logic

Media logic is a set of the­or­ies describ­ing how the medi­um affects the media. Typically, the format (as the medi­um dic­tates) influ­ences the medi­ated message.

“Media logic is defined as a form of com­mu­nic­a­tion, and the pro­cess through which media trans­mit and com­mu­nic­ate inform­a­tion. The logic and guidelines become taken for gran­ted, often insti­tu­tion­al­ized, and inform social inter­ac­tion. A basic prin­ciple is that media, inform­a­tion tech­no­lo­gies, and com­mu­nic­a­tion formats can affect events and social activ­it­ies.“
Source: The International Encyclopedia of Political Communication ³Altheide, D. L. (2016). Media Logic. The International Encyclopedia of Political Communication, 1 – 6. https://​doi​.org/​1​0​.​1​0​0​2​/​9​7​8​1​1​1​8​5​4​1​5​5​5​.​w​b​i​e​p​c​088

As fam­ously stip­u­lated by Marshall McLuhan, “The medi­um is the mes­sage.” What are the typ­ic­al effects of media logic on medi­ated messages?

Classic Media Logic Effects

Classic media logic is hypo­thes­ised to influ­ence the news media in the fol­low­ing ways: ⁴Nord, L., & Strömbäck, J. (2002, January). Tio dagar som skakade världen. En stud­ie av medi­ernas beskrivningar av ter­ror­at­tack­erna mot USA och kri­get i Afghanistan hösten 2001. … Continue read­ing

The effects of the above media logic can also be recog­nised in social media. Still, social net­work algorithms seem to add even more effects:

Social Media Logic Effects

“Social media logic, rooted in pro­gram­mab­il­ity, pop­ular­ity, con­nectiv­ity, and datafic­a­tion, is increas­ingly entangled with mass media logic, impact­ing vari­ous areas of pub­lic life.”
Source: Writing Technologies eJournal ⁵Dijck, J., & Poell, T. (2013). Understanding Social Media Logic. Writing Technologies eJournal. https://​doi​.org/​1​0​.​1​7​6​4​5​/​M​A​C​.​V​1​I​1​.70

Based on the sug­ges­ted addi­tions for social plat­forms, we can add four extra dimen­sions to the clas­sic media logic effects model:

Social media logic seems entangled with clas­sic media logic. While more com­plex, social net­works seem to amp­li­fy the effects of clas­sic media logic.

Learn more: Social Media Logic

Those Who Crush the Algorithm

Content that attracts instant­an­eous momentum and makes a big splash through­out each iter­at­ive test­ing series will out­per­form oth­er con­tent types. 

As for YouTube, this is per­haps most clearly demon­strated by the record-break­ing MrBeast:

The fact that MrBeast has mil­lions of fol­low­ers on YouTube is not fun­da­ment­al for his con­tinu­ous suc­cess — his extra­vag­ant, over-the-top videos would go vir­al regard­less of which account pos­ted them. ⁶ I’ve con­duc­ted sev­er­al in-depth inter­views with MrBeast, and it’s clear that he’s a bril­liant con­tent cre­at­or who pub­lished videos on YouTube for a long time before his pub­lic break­through.

Here’s the point: Algorithmic dis­tri­bu­tion doesn’t mean con­tent cre­at­ors can’t reach a vast audi­ence. Innovative cre­at­ors, like MrBeast, will find ways to nego­ti­ate the algorithms and go vir­al repeatedly.

While we can appre­ci­ate MrBeast’s videos as pure dopam­ine-infused enter­tain­ment for the masses, we might be miss­ing out on con­tent cre­ated by indi­vidu­als and organ­isa­tions that could’ve reached their already-earned sub­scribers, fans, and followers.

Social Media Must Be Managed

Today’s ongo­ing amp­li­fic­a­tion of sen­sa­tion­al­ism, i.e. the silent switch, has pro­found implic­a­tions for the pub­lic rela­tions function:

Many organ­isa­tions have spent ser­i­ous resources build­ing trust and attract­ing sub­scribers, fans, and fol­low­ers on social media. But the value of hav­ing sub­scribers, fans, and fol­low­ers is quickly eroding.

As PR pro­fes­sion­als, we must be rational. 

We could com­pare the situ­ation to our long-stand­ing rela­tion­ship with tra­di­tion­al news media. Traditional news media is power­ful but imper­fect, so PR pro­fes­sion­als must man­age it. The same goes for social media.

It’s not our place as pro­fes­sion­al com­mu­nic­at­ors to tell journ­al­ists how to do their jobs, just as it isn’t our place to tell social net­works or con­tent cre­at­ors how to do theirs.

If any­thing, we should be more ser­i­ous about the import­ance of social media man­age­ment.

Because social media must be managed.

As PR pro­fes­sion­als, our job is to help organ­isa­tions com­mu­nic­ate effi­ciently dur­ing these times of con­stant change.

Today, social media man­age­ment is crit­ic­al for organ­isa­tions aim­ing to engage effect­ively with their pub­lics, man­age their brand, and nav­ig­ate the com­plex land­scape of online com­mu­nic­a­tion. This involves vari­ous activ­it­ies, from con­tent cre­ation to stra­tegic com­mu­nic­a­tion and repu­ta­tion management.

Learn more: Social Media Must Be Managed

What should you study next?