Social signals are not a ranking factor.

Why it cannot work the way it is sold

Two practical problems sit in the way of share counts being a ranking input. The first is that the data is not reliably available: the platforms have restricted and repeatedly changed public access to engagement numbers, so a ranking system depending on them would be building on something it does not control and cannot verify.

The second is that it would be trivially gameable, and cheaply. Engagement can be bought in bulk for very little, which is exactly the property that makes a signal useless. Any metric that can be purchased at scale stops measuring the thing it was supposed to measure the moment it starts deciding rankings, which is the same reason Domain Rating is a reading rather than a lever.

What social genuinely does

Distribution, which is where links start. Journalists, bloggers and industry people find things on social. A piece of research that gets in front of the right twenty people can earn coverage that does carry links, which is the actual mechanism behind most successful digital PR. Social is the distribution layer, not the signal.

Brand and name searches. People who encounter you socially later search your name, and demand for your brand as a query is a genuinely useful thing to grow. It is also far harder to fake than a share count.

Discovery and indexing, incidentally. A link posted publicly can be a route by which a crawler first reaches a new URL, which occasionally matters on a site with a thin link graph. That is discovery, not ranking, and it is the same distinction behind sitemaps not getting pages indexed.

Entity clarity for models. Consistent, accurate profiles are part of how a language model resolves who you are, which is why they belong in the sameAs set of your structured data. That is a real and growing reason to keep them tidy, and it has nothing to do with engagement counts. It is entity work.

What follows from that

Do not buy engagement. You are purchasing a number nobody counts, on platforms increasingly good at detecting it, for a benefit that does not exist. If the goal is rankings, the budget belongs in the thing shares were only ever a proxy for: something genuinely worth covering, and the outreach to put it in front of people who might cover it.

And judge social by what it actually delivers. Reach, referral traffic that converts, conversations with people in your market, and coverage that follows from those. Those are worth having on their own terms. Reporting them as a ranking contribution is where it goes wrong, because it invites optimising the count instead of the content, and the count was never the thing.