Social selling is a research problem, not a content problem
Most LinkedIn programmes fail because they start with what the executive wants to say. The ones that work start with what the prospect has already said.
We have run social-selling programmes from executive profiles in healthcare, document software, software testing and media. The pattern that separates the ones that produced pipeline from the ones that produced likes is not writing quality. It is research.
The failure mode
A founder agrees to post twice a week. An agency writes thoughtful pieces about the industry. Engagement grows among peers and vendors. Six months later, nobody can name a deal that came from it. The programme was broadcasting, and buyers do not respond to broadcasts, however good.
The working model
Start with the target account list, not the content calendar. For each account, know who the buying group is and what they have said publicly: their posts, their comments, the roles they are hiring, what their CEO told an interviewer. Then engage with that. A comment that references a prospect's own argument is worth more than any post.
Content still matters, but its job is different: it is the thing a prospect finds when they check your profile after you commented on theirs. It should answer the question they are already asking, which you know because you did the research.
Where agents fit
This is exactly the kind of work agents are good at: monitoring hundreds of profiles for triggers, summarising what a prospect has said, drafting a comment for approval. The human still decides what to say. The machine makes sure they never miss the moment to say it.
The buyer moved, the channels decayed, and the tooling got smarter. A practical account of what still works and what to stop paying for.
Open-source intelligence started in security and journalism. Applied to accounts, it is the most under-used advantage in B2B sales.