The recurring argument in agent marketing is whether social media has made websites unnecessary. The question assumes both things do the same job. They do not, and naming the difference settles most of the debate without anyone having to defend a preference.
Push and pull
Social is a push channel. You create attention among people who were not looking for you. It works on the platform's timeline, which for most content means it peaks within hours and is effectively gone within days. Nothing you posted last March is producing anything today, however good it was.
Your website is a pull channel. It answers questions on demand, for people actively looking, on their timeline rather than yours. A neighbourhood page written two years ago can still be producing enquiries this week, because it gets retrieved by matching a question rather than by being recent.
Neither replaces the other. The specific failure worth avoiding is running only the push channel and wondering why nothing accumulates.
The leak, described precisely
Here is what actually happens to social attention that has nowhere to land.
Someone watches your video in April. They do not need an agent in April. They remember your name, or half of it, and keep scrolling. In August their landlord sells, or their family outgrows the condo, and they decide to move.
They do not scroll back four months to find you. Nobody does. They search your name, or they ask an AI assistant for an agent in their area, and whatever exists at that moment decides whether a year of posting turns into a conversation.
If the answer is a thin profile page or nothing at all, the attention you spent months building cannot be banked. That is the leak, and it is invisible, because you never learn about the searches that found nothing.
Why assistants make this worse
The verification step increasingly happens inside AI tools, which is where social performs least well.
Assistants building an answer about agents lean on public web pages, reviews, and business profiles. Much social content sits behind platform restrictions and is not structured for retrieval, so it contributes little regardless of how well it performed with humans. We covered how that retrieval works in how leads arrive from AI search.
An agent with a large following and no substantive site can therefore be locally famous and simultaneously invisible to the systems now brokering introductions.
What each channel is genuinely good at
Social is good at creating recognition among people who were not searching, showing personality in a way a page cannot, staying present with your existing circle at low cost, and anything time-bound: the listing that closes next week, the reaction to something local.
Your website is good at answering specific questions for years, surviving platform changes, being retrieved by search engines and assistants, carrying proof placed beside claims, and existing on infrastructure you own rather than rent.
The content that suits each is different too. Time-bound and personal belongs in a feed and clutters a website. Durable answers belong on a site and vanish in a feed within a day.
Making them work together
Adapt rather than cross-post. A link to a long page usually underperforms a post that delivers the useful part directly and mentions that a fuller version exists. Social introduces the idea; the site is where it is complete.
Then close the leak. Make sure your name, in any platform bio, resolves to something you control, and that the page it resolves to states plainly who you are and where you work. That is the same identity layer described in the agent bio post, and it is what converts recognition into a contactable outcome.
Measuring social honestly
Engagement is not the measure. Likes and views record attention, and attention can rise for years without producing a client.
The measure is whether anyone contacted you as a result, which analytics will usually fail to attribute, because the path runs through memory and a later search. That is why asking every new client how they ended up talking to you matters here specifically: it is the only instrument that catches the social contribution at all.
Expect the answer to be layered. People rarely say social media. They say they had seen your name around, then looked you up. Both halves count, and only the second one is visible in your data.
Where this sits against how clients actually choose
Per NAR's 2025 data reported by Virginia REALTORS, 43% of buyers found their agent through a referral and 18% used an agent they had worked with before.
Social's honest role in that picture is reinforcement rather than acquisition: staying present with the people who already know you so you remain top of mind when someone near them mentions moving. That is a real job and it is worth doing. It is not the same job as being findable, and treating it as such is what leaves agents with an audience and an empty inbox.
The platform risk nobody plans for
One more asymmetry, and it is the one agents discover at the worst possible moment.
Everything you build on a social platform sits on infrastructure you rent from a company whose priorities are not yours. Reach for a given kind of post can fall by most of its previous level without notice or explanation. Account restrictions happen, sometimes in error, and the appeal process is a form rather than a person. Platforms decline in relevance, and audiences move.
None of that is a reason to avoid social. It is a reason not to let it be the only place your business exists, because the failure mode is total and instant rather than gradual.
The practical hedge is small. Make sure your audience can reach you without the platform: your name resolving to your own domain, an email address you control, and enough information on your own site that somebody who remembers you in three years can still find you if every account you hold today has vanished.
Agents who have been through a lost or restricted account describe the same thing afterwards, which is that the followers were never really theirs. The relationship was real; the connection to it was borrowed. Owning the destination is what converts a borrowed connection into one that survives.
The cheapest insurance is to move a small number of the relationships off the platform deliberately: the people who engage repeatedly, the ones who message you, the ones who have mentioned moving eventually. A real conversation and a saved phone number takes them out of the rented column permanently, and those are the relationships most likely to matter.
The takeaway
Social creates attention that expires. Your website answers questions for years. Run social for recognition and presence with people who already know you, keep the content time-bound and personal, and stop measuring it by engagement. Then make sure every bit of that attention has somewhere durable to land, because the moment that decides whether any of it counts happens months later, in a search box or a chat window, when you are not there to influence it. The posts are the invitation. The site is the address, and an invitation without an address reaches nobody twice.
