A Query That Finds You, Not Your Homonym

A Query That Finds You, Not Your Homonym

Build a Boolean query from a random sample of real mentions, cut the fake volume, and record why every exclusion exists.

brandwatch social-listening boolean-query data-quality brand-monitoring

What it is for

A build order for a listening query: go wide, read a random sample, then write exclusions from what you actually found rather than what you imagined. It gives a Boolean skeleton, tells you to run it through the tool's own validator before saving, and names the two biggest sources of fake volume for consumer brands. It prevents a dashboard that counts coupon spam, job posts and a homonym as your brand's conversation.

The skill

Use before you trust any number coming out of a listening query.

Your brand name is also something else - a common word, a surname, a song, a
place, a competitor's product line. Find that out on purpose, not from a
report that is already wrong.

Start with a wide inclusion, run it, and read a hundred random mentions.
Random, not the top hundred. The top ones are the ones you already expected.

Build exclusions from what you actually saw:

  ("Your Brand" OR "YourBrand" OR "yourbrand.com")
  NOT (lyrics OR "the song" OR "the movie" OR "the film")
  NOT (hiring OR "we are hiring" OR "apply now")
  NOT ("discount code" OR coupon OR "promo code")

Boolean syntax differs between listening tools, and an exclusion that fails to
parse can silently leave the noise in. Paste the query into your tool's own
validator or preview, read the first page of results, and only then save it.

Rules for the query itself:
- Quote every phrase. Unquoted words match separately, and the volume chart
  will look wonderful and mean nothing.
- Keep your own accounts and scheduled posts out of the results, by query or
  by filter, or your share of voice is mostly you talking to yourself.
- Coupon spam and job posts are the two biggest sources of fake volume for a
  consumer brand. Remove them first.
- Do not exclude a whole platform to fix noise. You lose the real conversation
  on it along with the junk.
- Add the misspellings people actually type, and the version of the name
  written without a space.

Keep it readable: one clause per line, with a note beside every exclusion
saying what it removed and when it was added. Otherwise the next person
deletes the line that was holding the whole thing together.

Re-read a random sample every month. Noise changes when the world does.

Failure this prevents: a quarter of reporting built on a number that counted
coupon spam, job ads and a band with your name.

How to install it

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from

Skills app Original