Playing the algorithm game

A post by media analyst Thomas Baekdal on LinkedIn triggered an intriguing discussion about LinkedIn, its algorithm, and how to work it to get traffic back to your site. Clearly, I was all over it. This is not only one of my main areas of teaching and research (I've been training journalists on social media use for nearly two decades…), but also LinkedIn is now the single biggest social referrer to this very blog.

Here’s what Thomas asked:

I'm confused. Why do so many people post something here and then say “link in the comments”. Why the heck don't they just put the link in the actual post? Am I missing something?!?

Well, yes and no. This is an example of platform drift: an idea from one platform drifting to another, and then another, borne on the currents of received wisdom. And it started, as so many things did, on Twitter.

In the immediate pre-Musk period of Twitter, it became increasingly common practice to put the link you wanted to share in a reply to your original post. The theory was that by doing so, you’d avoid the first post getting the lower weighting of link posts in the Twitter algorithm. If you put good enough content in the first post, the second tweet would travel with it as a thread, and you’d end up with more click-throughs than if you'd posted a native link. It worked, for a while, until Elon reshaped the algorithm into something bizarre and inexplicable that mainly serviced his desires.

Link in comment: From Twitter to Facebook

But, by then, the concept had already drifted into Facebook, where people were also seeing the link downranking effect, and were seeking ways around it. The whole concept is something of a roll of the dice. You’ll get a smaller clickthrough rate from a comment than a post. But if the main post goes further, a smaller percentage of a bigger number can be more than you would have had in the first place.

Again, I’ve seen responsibly convincing proof that it works on Facebook.

Or, at least, that it works for now.

Remember, every time that you’ve found a way to “cheat” the algorithm, and it’s become a widely used approach, you run the risk of the platform just changing their algorithm to compensate. There have been multiple approaches that have come and gone in the last decade, as Facebook eventually catches on and quashes the “growth hack”.

And now, Link in LinkedIn comments

The drift continues, and as LinkedIn started deprioritising links both visually (by shrinking the share card in the feed) and algorithmically, people have started using this technique there. Is it just a cargo-cult like attempt to work the algorithm, or is there something more at play?

Jacob, at Journalism UK (still not used to that), explored this in the site's daily newsletter. He ran some of his own experiments which seemed to confirm the link deprioritisation effect. Then he started reading:

But then I went back to Thomas' post and one of his commentators pulled out an answer from the horse's mouth. A top LinkedIn exec says the problem isn't the links, but that posts are usually boring. Don't skimp out on the insights and advice, and you'll give yourself a fighting chance.

Hmmm.

I am sceptical because generally one thing you can guarantee about top platform execs is that they aren’t ever going to give you a clear answer about their algorithm. Why?

1. The more they tell us, the more we can manipulate it
2. It’s competitive information, and so sensitive
3. The algorithm is being constantly tweaked, so it’s likely even they don’t know exactly what it does at any point.

LinkedIn “insights“ from top execs

But let’s look at the quote. From the LinkedIn post Jacob mentioned:

𝗟𝗶𝗻𝗸𝘀 𝗮𝗿𝗲𝗻’𝘁 𝘁𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺. 𝗕𝗼𝗿𝗶𝗻𝗴 𝗽𝗼𝘀𝘁𝘀 𝗮𝗿𝗲.
Most people drop a link and basically write: “𝘊𝘩𝘦𝘤𝘬 𝘪𝘵 𝘰𝘶𝘵!” No value. Just promotion.
But LinkedIn’s algorithm rewards 𝘂𝘀𝗲𝗳𝘂𝗹 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 — so it ignores your empty promo.

Again, hmm.

Let’s note two things:

1. This isn’t a direct quote
2. She doesn’t actually say that links aren’t down-ranked in the feed. She’s talking around this issue, not attacking it head on. (And I’ve listened to the interview — she clearly avoids saying that.)

However, the core point is accurate. Just dumping “read my stuff” with a link is pretty much always a failed technique. Winning attention in a social feed is normally a combination of three things:

1. A compelling visual to stop people scrolling
2. Some copy that brings value to readers even if they don’t click through
3. That copy written is such a way as it stimulates discussion – because comments are a strong signal in most platform algorithms.

So, these things can both be true:

• LinkedIn gives links lower weight in the algorithm
• Great copy or images can help alleviate that.

And that brings us to something important: by the time executives from a platform are being questioned about a hack, it's probably already dead. And, indeed, the smart conversation has already moved on from this.

What’s actually working on LinkedIn

As soon as you start looking at publishers who are doing decent numbers on LinkedIn, it becomes clear that the “link in comments” discussion lags the reality on the (platform) ground. Because publishers who are really good at LinkedIn are doing something different.

Look at this post from the Financial Times:

This is a post with:

• Some copy
• A link
• A specially designed share graphic

The thinking here is either that LinkedIn is most likely to classify it as a photo post, and give it that weighting in the algorithm, or that by putting something visually compelling in the feed, you’re generating enough engagement that the post spreads. (I did drop an email to a former student who is head of social media at the FT for some insight, be she has rather inconveniently gone on holiday…)

Here’s a rule of thumb: if you want to see what techniques are actually working on social, look at what’s working for publishers with decent-sized audience teams. They’re the ones doing the constant, iterative testing to keep shifting their approach to deal with algorithm updates.

That’s going to be much more productive that listening to the many competing ideas that are circulating as “received wisdom”.

If you want to know more, the link's in the comments…