Brandwatch / Selected practice
Finding the useful conversation.
There is always more conversation than anyone can read. My job is to find the part that matters, understand what is changing, and leave the team with something more useful than a bigger spreadsheet.
Listening in practice
How I listenBrandwatch / an example
A simplified, fictional query
("Brand A" NEAR/5
("scanner" OR "imaging"))
NOT (jobs OR recruitment)Identity + contextNoise excluded
Boolean / query building
First, find the right posts.
I write a search that knows what to include and what to leave out. Brand names and context belong in. Job ads and unrelated namesakes don't.
Portfolio reconstruction of my listening workflow. Fictional names and illustrative examples; no live client data.
Start with a question.
I use Brandwatch for social listening and competitor intelligence. Before I build a dashboard, I want to know what we're trying to understand. Who is getting attention? What is driving it? Which questions keep turning up?
A good search is as much about leaving things out as bringing things in. I build and refine Boolean queries around names, context and exclusions, then read individual results to check that the conversation belongs in the first place.
Give the conversation some shape.
I build categories for themes such as launches, partnerships and customer experience. These answer a different question from sentiment. What people are discussing and how they feel about it shouldn't become the same label.
One competitor-intelligence handoff brought together four competitor queries and five question-led dashboard tabs, with analyst tags for useful posts, spike drivers and noise. The aim was to make it possible to follow a finding back to its source.
Make the comparison fair.
A comparison needs the same dates, markets, languages and sources. Otherwise the chart can look convincing while telling us very little. When something spikes, I go back to the posts before deciding what caused it.
Automated sentiment is a starting point. I check the language, keep the caveats visible, and distinguish an interesting pattern from a conclusion the evidence can actually support.
Leave something useful behind.
My contribution covers the queries, categories, dashboards and interpretation, along with a reusable template and guidance for reading the results. Not just the view of what happened, but a way for the team to keep asking better questions.
The last step is a short brief: what changed, what it might mean, how confident we can be, and what deserves a closer look. A dashboard can show the movement. Someone still has to explain why it matters.