Social Selling Signals: How to Spot Buying Intent in Conversations

Every day, a subset of your potential customers post the thing a salesperson would kill for: "Can anyone recommend a tool for X?" The window to respond usefully is shorter than almost anyone realizes, and the behaviour of most companies — including mine, for a while — doesn't match what the research says works.

The Harvard Business Review study that keeps getting quoted on this, and which has been replicated multiple times since, analyzed response times across 2,241 U.S. companies and found that firms who contacted a lead within an hour were roughly seven times more likely to have a meaningful conversation with a decision-maker than firms that waited even one hour longer. The version I see quoted on LinkedIn — "leads reached within 5 minutes are 21x more likely to convert" — comes from subsequent research by the Lead Response Management study, which found similar sharp decay curves.

The point the numbers all agree on: intent signals are perishable. If you're checking once a day, you're responding to yesterday's decisions. That is the whole game.

The Four Tiers Worth Monitoring

Not every signal is worth the same response. I sort them into four tiers, and the response protocol is different for each.

Tier 1: Direct Intent

Phrases where someone is explicitly asking. "Can anyone recommend a tool for X?" "What's the best [category] for [specific use case]?" "We're looking to switch from [competitor]." "Has anyone used [your product] vs [competitor]?"

The response window is one to four hours, not twenty-four. Someone posting a recommendation request on Monday morning has usually decided by Tuesday. The useful monitoring frequency for this tier is real-time or close to it — daily digests don't work here.

Tier 2: Problem Intent

Phrases where someone has named a problem but hasn't yet connected it to a solution. "Struggling with X." "We spend too much time doing X manually." "How do other companies handle X?"

The window is longer — 24 to 48 hours — because the person is still in problem-definition mode, not solution-comparison mode. The response here isn't "buy my thing," it's to be useful in a way that positions you favourably when they do start comparing.

Tier 3: Situational Intent

Changes in circumstance that typically precede buying. "Just joined [new company] as [role your product serves]." "We just raised [funding round]." "Hiring for [role that uses your product]."

New leaders make tool decisions in their first 90 days. Freshly-funded companies increase spending velocity. These signals have a longer window (one to two weeks) but represent larger buying windows when they do convert.

Tier 4: Sentiment Signals

Engagement with industry content. Following competitors. Commenting on productivity posts. Low individual value, high volume. Don't sell to these. Engage with them as relationship-building.

Why Speed Matters the Way It Does

The speed-to-lead numbers are famous. What's less-discussed is that they don't just measure responsiveness — they measure whether the person is still in the right mental state to buy.

When someone posts "can anyone recommend," they're in evaluation mode. That mode lasts hours, not days. By Tuesday evening they've usually collected three-to-five options, sorted them, and moved on. Your reply at 48 hours arrives after the conversation has closed inside their head, even if the post is still live on the platform. This is why the decay isn't linear — the drop-off is steep at the start and then flat. The person is either still deciding or they're not.

The practical consequence: the difference between responding in one hour and responding in four hours is significant. The difference between responding in four hours and responding in 48 hours is almost everything.

The Response Framework

Spotting signals is useless without a response approach that doesn't read as sales-bot output. I use what I started calling HELP internally, mostly because I needed a mnemonic, not because it's clever.

Honest positioning. Disclose your affiliation in the first sentence. "I'm the founder of X" or "Full disclosure, I work at X." Transparency is the minimum bar for engaging in someone else's recommendation thread. Hidden promotion gets your account flagged and your brand damaged.

Educate first. Lead with useful information, not a pitch. Answer the question. Share relevant context. The response should be valuable even if they never buy from you. This is the part most "social selling" playbooks skip and it's the difference between a reply people thank you for and one that gets reported.

Listen to their specific situation. Before recommending, ask one follow-up about their use case. "What's your team size?" or "Are you publishing to LinkedIn specifically or multiple platforms?" Genuine interest outperforms generic pitch every time.

Propose next steps gently. If appropriate, suggest something with low commitment. "Happy to walk you through how we approach this if it's useful" beats "Book a demo."

Platform-Specific Notes

The rules above work everywhere, but each platform has a tilt.

Reddit has the strongest anti-promotional culture of any mainstream platform. Mentioning your own product without disclosure is an instant brand-damage event. With disclosure and genuine helpfulness, it's tolerated. The community norm on most subreddits is: contribute ten times before you promote once. Read the sidebar rules of any subreddit before replying.

LinkedIn is the friendliest to product mentions if the response is substantive. B2B buyers expect vendors to be in the conversation. The failure mode here is templated replies — if your response could apply to any similar thread, it doesn't count.

Threads is still forming its own norms. With 400M+ monthly actives as of August 2025, the culture is conversational and more forgiving than Reddit, more casual than LinkedIn. Matching the register — reply with the same informality people are posting with — is what works.

X rewards speed above almost everything else, which pairs badly with long disclosures. The move is: reply fast with substance, and put your disclosure in the first line compressed ("Founder of X here — …").

The System I Actually Run

For FeedSquad, I run a narrow version of this for my own category. Real-time alerts on five or six specific phrases. Morning scan of saved Reddit searches. Threads keyword monitoring through the day. One rule: any Tier 1 signal gets a response within the hour or not at all, because a four-hour reply reads as desperate more than it reads as helpful.

The weekly routine is less glamorous than the tool vendors suggest. Log what I replied to. Note which conversations became actual customer calls. Drop phrases from my keyword list that generate noise, add ones that generate signal. Over six months, the list gets much sharper and the response rate improves without any increase in time spent monitoring.

What Not to Do

Don't copy-paste responses. Everyone notices.

Don't reply twice to the same thread if the first reply got no response. Following up in a public thread is a fast route to reputation damage.

Don't pitch in situational-intent threads. "I saw you joined [company] as [role] — here's how our product helps" is an outreach that lands as creepy. The move is connection request, short note introducing yourself, and patience. The sales conversation happens later, if it happens.

Don't ignore negative signals. If someone posts something unflattering about your product, respond constructively. Ignoring criticism is worse than addressing it — other prospects are watching.

Don't monitor without acting. Collecting signals in a sheet nobody opens is cosplay. Every signal needs a clear owner and a response expectation, even if you're the only person in the company.

---

FeedSquad's Pulse agent watches X for buying-intent phrases in your category in real time, so you see them inside the hour instead of the next morning.

Sources:

• Harvard Business Review — The Short Life of Online Sales Leads
• LeanData — Speed to Lead: The B2B Guide to Faster Response
• TechCrunch — Threads now has more than 400 million monthly active users