Agency multi-brand listening without shared inboxes

Agencies do not fail listening because they lack keywords. They fail because yesterday's client filter is still on, or because two brands share an inbox and someone screenshots the wrong thread into a status deck.

What "multi-brand" has to mean

Separation is the product. Each client needs its own monitors, its own mention history, and a shareable view that cannot leak another retainer. A single global search box across every client is a feature for the agency and a liability for the client.

We model that as projects: one workspace login, hard walls between brands, public dashboard links that only show what you chose to publish. The Team plan exists for that shape of work — not because agencies like paying more, but because collecting five brands costs five times the provider spend.

Reporting without theatre

Clients do not need another PDF of sentiment pie charts. They need proof you saw the angry thread before it became a ticket, and a short list of what you did. Urgency and category on each mention make that list honest; vanity reach charts do not.

If your current tool forces every client into one stream with colour-coded tags, ask what happens when a junior exports the CSV. If the answer involves "be careful which rows," the architecture is wrong.

The buying question

Ask whether API keys and webhooks are scoped per brand. Ask whether a shared link can be branded as the agency without exposing the vendor chrome of a competitor's client. Those are dull questions. They are also the ones that decide whether you can run listening as a service instead of as a risky spreadsheet.