Keyword tracking
Most keyword lists are too broad to be read
Keyword tracking is running standing searches for specific phrases so that new posts containing them reach you continuously, rather than being found by scrolling.
A keyword list is a budget. Every term you add spends attention that some other term needed, and the total attention available is one person reading for ten minutes a day. Treat the list as capped and it works. Treat it as free and it turns into noise within a fortnight.
A good term describes a moment, not a topic
Topic terms match everything written about a subject forever. Moment terms match a person in a situation: asking for a recommendation, complaining about a tool, announcing a project, hiring for a role. A moment gives you a reason to be in the thread that is not just having an opinion about the topic.
Narrow a noisy term by adding a verb
Nouns are broad and verbs are narrow. The topic word plus a verb like looking for, switching from, moved to, or replaced, cuts a term that fires two hundred times a day down to one that fires five times a day and is worth every one of them.
Negative terms are half the work
The fastest improvement to most lists is not a better keyword, it is excluding the recurring context that ruins the good one. Job posts, giveaway threads, and the one meme that reuses your phrase account for most of the noise in a typical term.
A slow term is not a dead term
Some of the best terms fire twice a week. Judge a term on how many of its hits you actually replied to, not on how many hits it produced. A term with two hits and two replies is beating a term with fifty hits and one.
How to do it
Collect real sentences first
Go and read twenty posts from people you would want as customers, and copy the exact phrasing they used. Do not invent keywords at a whiteboard, because buyers do not use your internal vocabulary.
Turn each phrase into a moment query
Add the verb that signals the moment. Aim for a query that would be a slightly odd thing to write unless you were actually in the situation you care about.
Run it for a week before judging it
Track two numbers per term: hits, and hits you replied to. A single day is not enough signal on a term that fires a few times a week.
Cut, narrow, or keep
Zero replies in a week means cut. Too many hits to read means narrow with a verb or a negative. Anything else, keep and leave alone.
Recheck the list every quarter
Vocabulary moves. The phrase everyone used for your category last year is not the phrase they use now, and a list that is never revisited slowly stops matching how people talk.
Common mistakes
- Tracking a category noun on its own, which matches every marketer writing about the category and almost no buyers.
- Judging a term by hit count instead of by how many hits earned a reply.
- Never adding negatives, so one recurring meme or job post format dominates the results.
- Copying an X query straight into LinkedIn, where posts are longer and phrased more formally.
How Quillen handles keyword tracking
Quillen takes keywords per campaign and runs them across X and LinkedIn, then scores each result against your ideal customer profile so a post that merely contains the word ranks below a post showing intent. Each discovery search costs a flat number of credits, which makes the budget analogy literal: a term that fires constantly is a term you are paying for constantly. Deduplication is enforced across the whole workspace, so the same post cannot be surfaced and replied to twice, and the same person is capped at three replies per rolling fourteen days.
Common questions
- How many keywords is too many?
- The limit is whatever you stop reading. For most solo operators that is around five terms; for a small team with someone owning the queue, ten to fifteen. Past that, hits go unread and the entire setup stops producing replies.
- Should I track my competitors as keywords?
- Yes, but as one term, not as a set. Competitor mentions are useful because someone comparing tools is already in market, but they are a small slice of the conversations worth joining and they crowd out problem terms if you let them.
- Why does my keyword return so much irrelevant content?
- Almost always because it is a noun with no situation attached. Add a verb that only someone in the moment you care about would write, and add negatives for the two or three recurring contexts you keep seeing.