Social Media Algorithms: What They Reward and What They Ignore

A social media algorithm is a set of rules a platform uses to decide which content gets shown to which people, and in what order. Every major platform runs one, and every one of them is optimised around the same core objective: keeping users on the platform for as long as possible.

That single fact explains more about how to succeed on social media than any tactical playbook. The algorithm is not your enemy, and it is not your friend. It is a system designed to serve the platform’s business model. Your content either fits that system or it does not.

Why Most Marketers Misread the Algorithm

Early in my career, I spent a lot of time optimising for what I could measure most easily. Clicks, impressions, reach. The numbers that appeared in dashboards and made it into client reports. It took me longer than I would like to admit to realise that optimising for easy measurement and optimising for actual outcomes are two very different things. The same trap exists with algorithm thinking.

Most marketers treat algorithms as either magic or mystery. Either they follow a set of received rules, post at the right time, use the right number of hashtags, hit the right word count, and expect results. Or they throw their hands up and say the algorithm is unpredictable, so why bother trying to understand it. Both responses miss the point.

Algorithms are not random. They are trained systems with clear objectives. The platforms publish documentation about how they work, even if that documentation is incomplete. Platform engineers give interviews. Independent researchers run tests. There is more signal available than most marketing teams act on, because reading it carefully requires more effort than reading a listicle about the best time to post on Instagram.

If you want a broader grounding in how social media fits into a commercial marketing strategy, the Social Growth and Content hub covers the full landscape, from platform selection to content systems to paid amplification.

How Each Major Platform Algorithm Actually Works

The mechanics differ by platform, but the underlying logic is consistent. Each algorithm is trying to predict: will this person engage with this content, or will they scroll past it? The prediction is based on historical behaviour, content signals, and the relationship between the user and the account posting.

Instagram

Instagram uses separate ranking systems for the Feed, Explore, Reels, and Stories. They do not all work the same way. Feed and Stories weight relationship signals heavily: how often you interact with someone, whether you have searched for them, whether you have commented on their posts before. Reels and Explore lean more toward interest signals: what topics you engage with, regardless of whether you follow the creator.

Instagram has been explicit that it does not penalise accounts for posting frequency, but it does distribute reach based on engagement rate. A post that earns strong early engagement, within the first 30 to 60 minutes, gets pushed to a wider audience. A post that earns weak early engagement gets limited distribution. This is not a conspiracy. It is a rational system for a platform managing an enormous volume of content.

LinkedIn

LinkedIn’s algorithm is arguably the most misunderstood in B2B marketing. The platform weights dwell time heavily, meaning content that people stop scrolling to read performs better than content that generates a quick like and nothing else. It also weights comments from people outside your immediate network, because that signals the content is reaching beyond your existing connections.

LinkedIn has been publicly vocal about penalising certain behaviours: engagement pods, posts that bait reactions, and content that drives users off-platform too quickly. The platform wants people to stay on LinkedIn, so posts that include external links in the body copy have historically received less distribution than posts where the link is in the first comment. Whether that specific mechanic persists is worth testing, but the underlying logic will not change.

Facebook

Facebook’s organic reach for brand pages has been declining for the better part of a decade. This is not an accident. The platform shifted its feed algorithm to prioritise content from friends and family over content from pages, and that decision was commercial: it pushed brands toward paid media. Understanding this is more useful than trying to game your way back to 2012 reach numbers.

Where Facebook’s algorithm still rewards organic content is in groups and in content that generates meaningful conversation. “Meaningful” is defined by the platform as comments that are substantive rather than single-word reactions. Posts that generate back-and-forth discussion between users, not just between the brand and its followers, tend to get broader distribution.

YouTube

YouTube is different from the others in one important way: it is a search engine as much as it is a social platform. The algorithm has two main surfaces: search results and the recommendation system. Search rewards relevance to the query, channel authority, and watch time. The recommendation system rewards click-through rate from the thumbnail and title, followed by watch time and completion rate.

The implication for marketers is that YouTube content needs to be planned with both surfaces in mind. A video optimised purely for search might have a dull thumbnail. A video optimised purely for recommendations might be impossible to find through search. The best-performing channels on YouTube tend to do both deliberately.

What Engagement Signals Actually Matter

Not all engagement is equal, and treating it as such is one of the more expensive mistakes in social media management. Likes are the lowest-value signal on most platforms. They require minimal effort from the user, they do not indicate that someone read or watched your content carefully, and they are easy to generate through low-quality tactics.

Saves, shares, and comments carry substantially more weight, because they require more from the user. When someone saves a post, they are signalling that they found it worth returning to. When someone shares it, they are putting their own reputation behind it. When someone leaves a substantive comment, they are investing time. Algorithms are designed to detect and reward these signals because they correlate with content quality in a way that passive impressions do not.

When I was running iProspect and we were building out the content side of the business, one of the discipline shifts we had to make was getting clients to stop reporting on reach and start reporting on engagement quality. Reach tells you how many people saw something. It does not tell you whether it did anything. A post that reached 100,000 people and generated 12 saves performed worse, from an algorithmic standpoint, than a post that reached 8,000 people and generated 400 saves.

The other signal worth understanding is completion rate, particularly for video. Platforms track how far through a video the average viewer gets. A video that people start and abandon after five seconds is a negative signal. A video that most viewers watch to the end is a strong positive signal, regardless of whether they hit the like button. This is why hooks matter so much in video content: not as a creative gimmick, but because the algorithm is watching whether your opening earns the next 30 seconds.

The Consistency Problem Most Brands Get Wrong

Consistency is one of the most cited pieces of advice in social media marketing, and also one of the least followed. The reason is usually structural rather than intentional. A brand launches a content push, posts frequently for six to eight weeks, runs out of content or budget, goes quiet for three weeks, then tries to restart from where it left off. The algorithm does not work that way.

When an account goes quiet, the platform stops learning about its audience. When it starts posting again, the platform has to re-establish which users are likely to engage. This re-learning period can feel like a penalty, even though it is just the algorithm doing what it is designed to do. The practical consequence is that inconsistent posting does not just miss the audience during the gap. It also suppresses reach after the gap.

I have seen this pattern across dozens of client accounts. A brand that posts three times a week, every week, for twelve months will almost always outperform a brand that posts daily for two months and then disappears. The compounding effect of consistent algorithmic trust is real, even if it is hard to quantify in a single reporting period. A structured content calendar is not a nice-to-have. It is the operational backbone of any serious social strategy.

The related mistake is treating consistency as purely about frequency. Consistency also means consistency of topic, tone, and format. Platforms build audience models based on what your content is about. If you post about financial services for six weeks and then start posting about food, the platform’s model of your audience becomes confused, and distribution suffers. This does not mean you can never vary your content. It means that variation should be intentional and gradual, not random.

Organic Reach Is Not Free, and It Never Was

There is a persistent belief in marketing that organic social is a free channel. It is not. The cost is time, and time has a value. Creating content, managing publishing schedules, responding to comments, analysing performance, and iterating on what works all require real resource. When I see brands treating organic social as a zero-cost activity, I know they are not accounting for the labour involved.

More importantly, organic reach on most platforms has been in structural decline for years. This is not a temporary dip. It is a deliberate commercial decision by platforms that have mature advertising businesses. The ceiling on what organic social can deliver for most brand accounts is lower than it was five years ago, and it will likely be lower still in five years’ time.

This does not mean organic social is not worth doing. It means the return expectation needs to be calibrated accurately. Organic social builds community, creates content that paid media can amplify, and establishes credibility with audiences who encounter your brand through other channels. These are real and valuable outcomes. They are just not the same as direct acquisition, and conflating the two creates bad planning decisions. Understanding what social media ROI actually measures is a prerequisite for setting sensible objectives.

The brands that perform best on organic social tend to be the ones that have stopped trying to compete with paid media on its own terms. They are not trying to drive direct conversions from every post. They are building an audience, earning trust, and creating content that paid campaigns can then put in front of new people at scale.

How AI Is Changing Algorithm Behaviour

Every major platform is now using machine learning at scale to power its recommendation systems. This is not new, but the sophistication has increased substantially in the last two to three years. The practical effect is that algorithms have become better at predicting what individual users want to see, which means the gap between content that earns attention and content that does not has widened.

For marketers, this cuts both ways. Better recommendation systems mean that genuinely good content can find its audience more efficiently than it could five years ago. A piece of content that resonates with a specific niche can reach that niche without requiring a large existing following, because the algorithm is better at matching content to interest. This is why you occasionally see accounts with a few thousand followers produce a post that reaches hundreds of thousands of people.

The flip side is that mediocre content is also more efficiently filtered out. When the algorithm has a rich model of what each user engages with, content that does not fit that model gets suppressed faster. The middle ground, content that is neither terrible nor excellent, is where most brand accounts live, and it is getting harder to sustain reach from that position.

AI tools are also changing how content is produced, which creates an interesting tension. If everyone uses the same AI tools to generate captions, ideas, and formats, the output starts to converge. Algorithms trained to reward originality and genuine engagement will, over time, penalise content that looks and sounds like everything else. Using AI in a social media strategy is not inherently a problem, but using it as a substitute for a genuine point of view is.

What Brands Should Actually Optimise For

There is a version of algorithm optimisation that is essentially a game of whack-a-mole. Every time a platform updates its ranking signals, you adjust your tactics. Post length changes, hashtag strategy changes, optimal posting times change. Teams spend significant energy chasing these updates, and the results are often marginal.

The more durable approach is to optimise for the things algorithms are trying to measure: genuine attention, real engagement, and content that earns a response. These things do not change when the algorithm updates, because they are what the algorithm is trying to surface. A platform might change how it weights saves versus shares, but it will never stop rewarding content that people genuinely want to see.

When I judged the Effie Awards, the work that stood out was not the work that had the cleverest media placement or the most sophisticated targeting. It was the work that had something real to say, and said it in a way that earned attention. The same principle applies to organic social. The brands that build durable reach are the ones that give their audience a reason to come back, not the ones that have the most optimised posting schedule.

Practically, this means investing in understanding your audience before investing in content volume. What do they care about? What problems do they have that you can address? What perspective can you offer that they cannot get from someone else? These questions sound obvious, but most brand social accounts are built around what the brand wants to say, not what the audience wants to hear. The algorithm is, in a roundabout way, enforcing audience centricity. Brands that have not made that shift find it increasingly difficult to sustain organic reach.

Earlier in my career, I overvalued lower-funnel performance signals. The numbers looked clean, the attribution looked tight, and it was easy to build a case for continued investment. What took longer to see was how much of that performance was capturing intent that already existed, rather than creating new demand. The same bias shows up in social media measurement. Brands optimise for the metrics that are easiest to attribute, and miss the longer-term value of building an audience that trusts them. Building a social media presence that compounds over time requires a different measurement frame than campaign-by-campaign performance reporting.

Platform-Specific Tactics That Still Work

With the strategic principles established, there are some tactical observations worth making. These are not hacks. They are observations about how platforms currently behave, grounded in the underlying logic of what algorithms reward.

On Instagram, Reels continue to receive preferential distribution over static posts. This is not because video is inherently better content. It is because Reels is the surface Instagram is competing with TikTok on, and the platform is incentivising creators to produce content there. That commercial incentive will persist for as long as the competitive dynamic does.

On LinkedIn, posts that generate comments from people outside your network perform better than posts that generate comments only from existing connections. The implication is that asking a genuine question, rather than making a statement and waiting for agreement, tends to generate more distributed reach. Not because the algorithm rewards questions specifically, but because genuine questions attract responses from a broader range of people.

On YouTube, the title and thumbnail are the most important variables for Shorts and the recommendation surface. Not because they are gimmicks, but because they are the only information a viewer has when deciding whether to click. A video with a weak thumbnail will underperform a weaker video with a strong thumbnail, in most cases. This is uncomfortable for people who believe quality should speak for itself, but it is how the system works.

Across all platforms, the accounts that perform best tend to have a clear and consistent perspective. Not a brand voice document with approved adjectives, but an actual point of view on something their audience cares about. This is harder to manufacture than a content calendar, but it is also harder to replicate. For smaller brands especially, a genuine perspective is often the only sustainable competitive advantage in organic social.

When to Stop Chasing the Algorithm

There is a point at which algorithm optimisation becomes a distraction from the actual work of building an audience. If your team is spending more time reading about algorithm updates than creating content worth reading, something has gone wrong.

The platforms that matter most have published enough about how their systems work to inform a sensible strategy. Beyond that, the marginal return on algorithm research diminishes quickly. The variables that matter most, content quality, posting consistency, genuine audience understanding, are not secret. They are just harder to execute than tweaking hashtag counts.

I have worked with brands that had genuinely excellent products and a story worth telling, but whose social media presence was invisible because the content was built around what the brand wanted to communicate rather than what the audience wanted to engage with. No amount of algorithm optimisation fixes that problem. The algorithm is not suppressing you because you used the wrong posting time. It is suppressing you because your content is not earning attention.

That is a harder conversation to have with a client or a leadership team than “we need to post at 9am instead of 11am.” But it is the honest one, and it is the one that leads to work that actually improves.

If you are working through a broader social media strategy and want to understand how organic, paid, and content all fit together, the Social Growth and Content hub covers each of these areas in depth, with the same commercially grounded approach.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions