Feeds are full of fluent, interchangeable posts. The reps who stand out are the ones drawing on what no model has: their customers' words, their company's own experience, and the live conversation in their market.
Why every feed started sounding the same
Scroll a B2B feed for five minutes and you notice the sameness before you notice any single post. Polished hooks, tidy lists, confident advice that could have come from anyone in the category. Much of it effectively did: a growing share is drafted by the same handful of language models, trained on broadly the same public internet. When the inputs are similar, the outputs converge.
That has a practical consequence for social selling. If your reps post what a model would write about your category, they are posting what your competitors' reps are posting too. Buyers learn quickly to scroll past it. The posts that stop the thumb are the ones that could only have come from a specific person, at a specific company, who has talked to real buyers recently.
So the question shifts from "how do we post more?" to "what do we have that a model doesn't?" The way we frame it at Click Shift, the answer comes from three voices.
Social selling is selling, so it has to sound like a seller
Social selling is not a content calendar with the sales team's faces on it. It is reps, founders and subject experts showing up where buyers research, being useful before a deal exists, and turning that attention into conversations. Gartner's research on the B2B buying journey found buyers spend only around 17% of their buying time meeting potential suppliers, and that complex purchases typically involve a buying group of six to ten people. Most of the persuading happens when you are not in the room, often through what those people read and pass to each other.
It also works on timing. Research by Professor John Dawes of the Ehrenberg-Bass Institute, published with the LinkedIn B2B Institute, estimates that only around 5% of business buyers are in-market at any given time. The other 95% will buy later, from whoever they already remember. A rep who has been visibly, specifically useful for six months is the person they remember. A rep who reposted generated advice is not.
We cover the wider strategy in our B2B social selling strategy guide. This article is about the raw material: what your people should actually say.
The three voices
The way we frame it, differentiated content draws on at least one of three sources of context that no model has on its own. This is our working model, not a piece of research, but it holds up well in practice.
1. The voice of the customer
What buyers actually say: how they describe the problem before they have learned your vocabulary, what they objected to, what nearly stopped them, what finally tipped them. It lives in sales calls, support tickets, onboarding questions, reviews (yours and your competitors') and win-loss conversations.
For social selling this voice is gold, because buyers recognise their own words instantly. A post that opens with the exact phrase a prospect used on a discovery call reads as "this person gets it". A post that opens with your positioning statement reads as marketing. If you want a structured way to collect this material, voice-of-customer research is the discipline built for it.
2. The voice of the company
Your own data, opinions and scars: the things only your team could publish because only your team lived them. Results with the numbers attached. What went wrong and what you changed. Opinions someone is willing to put their name to. What you have learned to stop doing.
A quick test: if a competitor could publish the post word for word with their logo on it, it is not in your company's voice. It might be accurate. It is not differentiating you from anything.
3. The voice of the market
The live conversation around your category: what is being argued about in communities, what analysts and competitors are saying, what regulation is changing, what your buyers are worried about this month. This voice supplies timing. A rep who responds thoughtfully to a conversation already under way gets read, because the attention is already there. Content produced from a calendar alone tends to answer last month's question.
Where AI fits, and where it doesn't
None of this is an argument against AI. It is an argument about where the advantage now sits. Fluent sentences are no longer scarce; context is. Models are genuinely useful for research, summarising call transcripts, spotting recurring themes in reviews, monitoring market conversations and drafting from material you give them.
What they cannot do is supply the voices. They were not on your calls, did not make your mistakes and do not know which thread your best prospect commented on yesterday. And on social, where the whole point is a human relationship, the final words should come from, and be approved by, the person whose name is on the post. Let the machine do the miles; keep a human on the last one.
Enablement: making it easy for reps to sound like themselves
Most social selling programmes stall not because reps lack opinions but because they lack material and time. Telling a busy account executive to "post more" produces either silence or generic posts. Enablement closes the gap:
- Give them the customer voice. Share verbatim phrases, objections and questions from calls, organised by theme, so reps start from what buyers actually said.
- Give them the company voice. Results, teardowns and points of view they can put in their own words, not pre-written posts to copy and paste.
- Give them the market voice. A short weekly note on what buyers are talking about, so reps know which conversations are worth joining.
- Give them assets for real conversations. The teardown to send after discovery, the story that answers this month's most common objection.
- Connect it to pipeline. Treat replies, messages and engagement from target accounts as demand signals and route them like any other lead. Our guide to ICP targeting and intent signals covers how to tell which signals matter.
The same material pays off twice. Customer phrases that make a good post also make sharper ad copy, better keywords and a clearer picture of who feels the problem most acutely. That is why we treat the three voices as market research that happens to produce content, not as a content tactic.
Try this this week
- Pull five recent sales call notes or recordings. Copy out, word for word, every phrase a buyer used to describe their problem. Do not paraphrase.
- Write down one result, one mistake and one opinion from your team this month.
- Note two conversations your buyers were having in public this week, and whether anyone at your company joined them.
- Ask one rep to write one post that opens with a buyer's exact phrase and answers it with one of the company-voice items. They write it, they approve it, it goes out under their name.
- Compare the replies and conversations it starts with that rep's last generic post.
One post proves nothing statistically. It will show you quickly whether your team has the raw material, or whether you need to go and collect it first.
The honest limits
- It is slow. Social selling works mostly on the 95% who are not buying yet. Expect months before it shows up in pipeline, and do not judge it on this quarter's closed deals.
- The 95-5 split is an estimate. It rests on an assumed purchase cycle; your market may cycle faster or slower.
- A human voice is not a guarantee. A genuine but dull post still gets scrolled past. The voices give you material; your people still have to say something worth hearing.
- It does not replace capture. Buyers who are in-market today still need to find you in search and get a fast response. Social selling complements that work.
- Not every rep will do it. Start with the few who want to, and let their results do the persuading.
Going further
The full working version, including how we structure a voice harvest, brief reps from it and judge whether it is turning into pipeline, is part of the methodology we share with clients. For a quick read on where you stand, the social selling audit takes a few minutes and covers customer voice, company voice, market voice and enablement. If you would rather have help running it, see our social selling services, or get in touch with Dave to talk it through.
Questions
What are the three voices in B2B social selling?
The way we frame it at Click Shift, they are the three sources of context no language model has on its own: the voice of the customer (buyers' own words for the problem), the voice of the company (your own results, mistakes and opinions) and the voice of the market (the live conversation in your category). Content that draws on at least one of them is hard for competitors to copy.
Can we use AI to write our social selling posts?
Use it for the groundwork: research, summarising calls, finding themes in reviews and monitoring market conversations. But the final words on a rep's post should be written or rewritten and approved by that rep. On social the relationship is the point, and generated posts tend to sound like everyone else's.
How long does social selling take to show results?
Usually months rather than weeks. Most of the value comes from being remembered by buyers who are not in-market yet, so early signs are conversations, replies and engagement from target accounts. Pipeline follows later. Judge it on that timeline, not on this quarter's closed deals.
Sources
- Gartner: The B2B Buying Journey
- LinkedIn B2B Institute: How B2B Brands Grow (with the Ehrenberg-Bass Institute, including John Dawes on the 95-5 rule)
More from the method
- Sales speed metrics: how fast you respond decides what you win
- Demand strategy: stages of awareness, the 95-5 rule and market stage
- Where B2B Ad Budgets Leak, and How to Find the Waste
- Sales Pipeline Automation Inside the Stack You Already Own
- Guardrailed AI Automation: Bounded Autonomy for Marketing Ops
- Measure marketing in pipeline dollars, not leads
- Audit before you automate: find what you are about to multiply