Social listening

Social listening is a filter, not a feed

Social listening is the practice of monitoring public conversations for the specific words, people, and questions that matter to your business, so you can act on them while they are still live.

Almost every listening setup fails the same way. Someone builds an alert for the brand name, the alert fires, and nobody does anything with it. Listening only pays when the output is a short list of conversations a human is about to reply to, which means the hard part is not collection. It is the filter.

Listening for your brand name is the least valuable kind

If somebody says your name you will usually hear about it anyway, and by then the conversation is about you rather than about the problem you solve. The valuable listening is for the problem: the question a buyer asks two months before they ever type your category name. Those conversations are where a useful answer still changes the outcome, and they are invisible unless you go looking.

Volume is the enemy, not the goal

A term broad enough to fire a hundred times a day gets muted within a week. A term narrow enough to fire three times a day gets read. When you are choosing between coverage and signal, choose signal, because the only listening that works is listening a human actually reads to the end.

X and LinkedIn reward different queries

On X the useful unit is the phrase somebody types in frustration or in public thinking out loud, so search reads like speech. On LinkedIn the useful unit is closer to a role plus a situation, because the platform is built around who someone is rather than what they just said. A query list copied from one platform to the other will underperform on both.

Listening without a response path is a hobby

Decide before you start what happens when a term fires. If the answer is that a specific person will reply within a day, the setup is worth building. If the answer is that it goes into a report, you are building a report, and you should be honest about that rather than calling it a growth motion.

How to do it

  1. Write down the problem, not the product

    List the sentences a buyer says before they know your category exists. These become your seed terms. A term that only your existing customers would use is a term that finds people you already have.

  2. Split terms by intent

    Separate terms that signal someone shopping right now from terms that signal someone merely interested. They deserve different response speeds and different people answering them.

  3. Add the people, not just the words

    Some accounts are worth watching whatever they post: the analysts, the loud customers, the competitors. Track those as a watchlist alongside the keyword searches.

  4. Prune weekly for the first month

    Every term that produced nothing worth answering in a week goes. Every term that produced too much gets narrower. Most listening setups are ruined by never deleting anything.

  5. Give every hit an owner and a deadline

    A conversation is worth joining for a day, sometimes two. Assign who replies and by when, or the filter you built will quietly become an archive.

Common mistakes

  • Tracking the brand name and nothing else, which finds you the conversations you are already in.
  • Setting up twenty terms at once so nothing gets pruned and everything gets muted.
  • Treating a listening tool as a reporting tool, then wondering why it produced no pipeline.
  • Replying to a thread that is four days old, where the conversation has already ended.

How Quillen handles social listening

Quillen runs keyword and creator searches across X and LinkedIn plus watchlists of specific accounts, and scores what it finds against your ideal customer profile before it reaches you. X discovery reads go through twitterapi.io, a third party aggregator, and LinkedIn reads go through a third party provider, both of which we name rather than describe as official platform APIs. The output is a queue of candidate conversations with a draft reply already written in your voice, which is the response path most listening setups never build. Discovery is a paid capability and each search costs credits, so a term that fires constantly costs more than a term that fires usefully.

Common questions

What is the difference between social listening and social monitoring?
Monitoring is watching what is said about you. Listening, done properly, is watching for the problem you solve whether or not your name comes up. Monitoring protects a reputation. Listening finds demand.
How many keywords should I track?
Fewer than you want to. Start with three to five terms you can read every hit for, and only add a sixth once you have deleted one. Coverage is easy and worthless; a list you read to the end is hard and valuable.
How fast do I need to reply?
Within a day for most threads on X and LinkedIn. Both platforms show a thread to fewer people as it ages, and a reply on a dead thread is read by the author and almost nobody else.