If you ask anyone what social media is, you will most likely get a list of apps as an answer: Facebook, Instagram, TikTok, WhatsApp. For everyday purposes that answer works, but it falls short as soon as you want to understand why these platforms show you what they show you, why some brands grow there while others post into the void, and why people spend on average more than two hours a day on them, according to DataReportal's reports. So let's take it step by step, from the definition to how they are changing the way we search for information.
What is social media exactly?
A social network is a digital platform where people create a profile, connect with other people, and share content that everyone else can see, comment on, and redistribute. The definition sounds simple, but it has two parts worth separating, because they explain almost everything that comes next.
The first part is the network itself: the map of connections between profiles. Your friends, your followers, the accounts you follow. Engineers call it the social graph, and it is what sets these platforms apart from a news site or a blog: content travels through people connected to each other, not from a central broadcaster.
The second part is the media: the stream of content flowing through that network. Photos, videos, text, live streams. That is why the term social media covers both, and some purists insist that "social network" means the structure of connections while "social media" means the whole platform. For practical purposes, most of us use the terms interchangeably, and that is fine.
The phenomenon is also not as new as it seems. SixDegrees, considered the first modern social network, launched in 1997, before Google was even a company. Facebook arrived in 2004 and today hovers around 3 billion active users. In other words, a significant share of connected humanity is on the same system of profiles and content.
How does it work under the hood?
Here comes the part most worth understanding, because it is the one that almost no definitional article explains well: the mechanism that decides what you see.
Early social networks were chronological. Your feed showed whatever your contacts posted, sorted by time, and that was it. That died years ago. Today every major platform uses recommendation algorithms that rank (and filter) content based on the probability that you will keep watching.
My favorite analogy is the neighborhood newsstand owner: at first he offers you a bit of everything, and as he figures out which magazines you spend the most time flipping through, he starts leaving those within view and tucking the rest away. The algorithm does the same, but measuring signals every second: how long you pause on a video, what you comment on, what you share, what you close immediately. That is where the metric everyone in this industry obsesses over comes from, engagement, because it is literally the fuel the algorithm uses to decide who gets shown what.
And in recent years there was a deeper shift that explains the rise of TikTok: the move from the social graph to the interest graph. In the classic model, your feed depended on who you followed. In the new model, it depends on what you like to watch, no matter who posts it. That is why on TikTok an account with 200 followers can reach a million people if the content retains attention, and why Instagram keeps inserting more content from accounts you do not follow. The network of people still exists, but attention is now distributed by interest, not by friendship.
What types of social networks exist?
You can classify them in several ways, but the most useful one is by the purpose people have when they open each one.
There are the mass content platforms, like Facebook, Instagram, and TikTok, where everything fits and attention is distributed by algorithm. The professional ones, with LinkedIn at the top, where the context is work and the same joke that lands on TikTok can cost you dearly. The video platforms like YouTube and Twitch, which work as a social network and a search engine at the same time. The messaging apps like WhatsApp and Telegram, which are technically a different category, but concentrate so much content forwarding and so many groups that in practice they operate as networks. And the community and interest platforms, like Reddit, Discord, or Strava, where people group by topic rather than by knowing each other.
For a brand, this classification matters more than it seems: each type has its own codes, and the classic mistake is publishing exactly the same thing everywhere, as if they were the same wall painted a different color.
What are they for, and what problems do they bring?
For people, social media serves fairly concrete purposes: keeping relationships alive at a distance, staying informed (with all the caveats that deserves), learning from specific niches and, increasingly, working, because entire industries now recruit and sell through these platforms.
For a business, they are a distribution channel with an entry cost far below any traditional medium: you can build a brand, serve customers, test messages, and sell directly. I wrote about how that day-to-day work is managed in my guide on the community manager role, and about how it fits into a larger strategy in the one on content marketing.
The problems are concrete too. For people: these platforms are designed to optimize retention, and the same signals that feed the algorithmic newsstand owner can end up serving content that is ever more extreme or ever more addictive. For businesses there is a structural risk I find myself explaining often: the audience you build on a social network is rented land. The platform sets the rules for how attention is distributed, and when they change (and they change every so often), your reach can drop by half without you doing anything differently. Any business that lives on social media alone is one algorithm change away from a very bad month.
What does social media have to do with Google and AI?
This is the layer that has changed the most in recent years, and the reason someone in SEO like me ends up writing about social media.
First, social networks became search engines. A Google executive acknowledged back in 2022 that around 40% of young people, when looking for a place to have lunch, start on TikTok or Instagram instead of Google. Searching for a restaurant, a dermatologist, or a travel idea directly in a social network's search bar is now an established habit, and the brands that optimize their profiles and descriptions for those searches show up; the rest simply do not exist in that channel.
Second, what happens on social feeds your brand profile everywhere. Google and AI assistants do not just evaluate your website: they build a profile of your company with everything they can find, and mentions, reviews, and conversations on social media are part of that material. When a language model decides which brands to recommend, familiarity carries enormous weight: I wrote about that based on a recent study about AI's favorite brands, and the short conclusion is that being a known, mentioned brand, on social media too, improves your chances of appearing in the answers.
In other words, what you do on social media also counts toward the ecosystem of signals that search engines and AI assistants use to decide whether to recommend you, even if the monthly report still files it under likes and reach.
How do you use it well if you run a business?
The full recipe, in six steps you can execute without a budget.
First, choose one or two networks, not five. The criterion is where your audience is, not where the trend is: if you sell B2B services, LinkedIn probably yields more than TikTok. If you are not sure who your audience is, that is a prior step, and defining a buyer persona is what it is for.
Second, define what your brand is there for: building the brand, generating inquiries, serving customers. Each objective is measured differently, and without an objective any number looks good.
Third, adopt a sustainable frequency. Three posts a week that you can keep up for a year beat twenty posts the first week and silence afterwards. The algorithm rewards consistency and so does the audience.
Fourth, converse. Reply to comments and messages, participate where your industry hangs out. It is a social network, and accounts that only broadcast end up talking to themselves.
Fifth, move your audience toward land you own: your website, your email list, your customer base. That is how you make sure an algorithm change does not wipe out all the work.
And sixth, measure something beyond followers: reach, interactions, clicks to your site, inquiries generated. Accumulated followers go up almost always, so they tell you very little.
This week's takeaway
Open the social network where your brand (or you) publishes the most and look for one specific number in the analytics: what percentage of your reach comes from people who do not follow you. If the platform shows it, that number tells you which game you are playing: if it is high, the interest graph is distributing you and your content competes on retention against the entire catalog; if it is low, you depend on your community and the work is about conversation and consistency. They are two different games, with different decisions, and knowing which one you are playing is the first piece of data that changes how you publish. The exercise takes ten minutes and costs nothing.