Before anyone types a search, four things have already happened.
1. They noticed something was wrong.
2. They gave it a name.
3. They decided it was the kind of problem you can buy your way out of.
4. And they picked a category to buy it in.
That is slow, deliberate work, and all of it finished before your website was involved. Google's search box only collected the result of the whole slow thinking process.
Social delivers a person who has done none of it. Scrolling is what the fast part of the mind does when there is nothing to decide. Nobody opens a feed holding a formulated question.
Contentsquare measured 99 billion sessions across 6,500 sites. Paid search converted at 2.8%. Organic social at 0.7%. (They publish no paid social figure, so that is the closest pair available, and it is not quite like for like.)
The comfortable reading is that social sends worse people. It is comfortable because it makes the fix somebody else's job. Better targeting, better creative.
The visitor is fine. The four things simply have not happened yet, and your page is now where they have to happen.
Then there is the third arrival. AI-referred traffic converted at 1.3%, and it was the only channel that grew. Same mechanism as search, done by a machine on the visitor's behalf. The assistant noticed the problem, named it, sorted the category, produced a shortlist. Your page arrives afterwards.
So there are three ways a person reaches you. On search, they did the thinking. Through an assistant, something did it for them. On social, nobody has done it yet.
Worth noticing where the thinking never happened: at Google. Google only decided which of those people landed on you instead of on a competitor, which is what the industry spent twenty years fighting over. Nobody was fighting over whether the visitor had thought it through. That came free with the act of typing.
So when you read a conversion rate, part of it is the page. Most of it arrived with the visitor.
And every point of traffic that moves from search to social relocates a piece of the buying decision onto your website. Not a worse visitor. The same decision, moved.
Your first screen answers a question. It has never had to raise one. For search and AI traffic that is right, because raising a question they have already answered wastes the second you have. For social traffic it is fatal. The answer lands on nothing.
Which is fixable, and there is evidence. Unbounce measured 41,000 landing pages. Paid social converted at 12%, ahead of paid search. Those pages state the problem before they state the product, which is the one thing a website almost never does. Send social traffic somewhere built for it and it stops looking like a weak channel.
Now read that number again. On a landing page, a conversion is a form. On a website, it is usually a sale. Search hands you a customer and social hands you an email address, and nobody publishes what happens in between.
Which is where B2B parts company with everyone else. Of the platforms Unbounce measured, LinkedIn converted four to five times worse than Facebook and Instagram, and LinkedIn is where your buyers are. That figure is still only the lead. In B2B the four things are harder to finish, because the problem is usually institutional and nobody has put a name on it yet. Social can start that. Search is still where it ends.
So write the missing line. One sentence at the top, naming the problem in the words your buyer would use, before anything about what you sell. A search visitor reads past it at no cost. A social visitor cannot start without it.
And if your ads and your search traffic point at the same URL, that line will fight one of them. Give them separate pages. Same offer, different opening.
Then open the page your ads point at now. Read the first sentence and work out what the reader has to already believe for it to make sense. Write that belief above it.
They used to bring that sentence with them.
Featured image: Ben Grosser: The Endless Doomscroller (2024)