What share of conversation measures
Pick a category — remittance apps, trekking agencies, dental clinics in Kathmandu — and count the public mentions of every brand in it over a period. Your share of conversation is the portion of those mentions that name you. It is measured with social listening tools that scan public posts, comments, forums, reviews and news, matched against a keyword list you define.
Two decisions shape the answer more than the tool does. The first is how you define the category, because a narrow definition flatters you and a wide one buries you. The second is which sources are counted: a platform the tool cannot read, or a private group where your customers actually talk, is simply absent from the total. Set both deliberately and keep them fixed, or the trend line means nothing.
Why share of conversation matters
Most marketing measurement stops at your own property — your site, your ad account, your page. Share of conversation is one of the few numbers that puts you next to the competition in the same frame, which is what a business owner usually wants to know: are we becoming the name people mention, or are we quietly slipping behind?
It also moves before sales do. When a competitor launches, changes price or has a public failure, the talk shifts first and the revenue follows later. Watching the share gives you time to respond while a response still matters. In Nepal, where a great deal of buying advice is passed around in Facebook groups and comment threads rather than published anywhere, it is often the earliest signal available.
Where share of conversation goes wrong
The largest trap is counting mentions without reading them. A share that rises because of a complaint thread, a service outage or a public argument is not a win, and a tool that reports volume alone will present it as one. Pair it with sentiment, and read a sample of the actual posts yourself rather than trusting a classifier on Nepali, mixed-script or heavily abbreviated text.
The second is brand-name ambiguity. A short or common brand name collects mentions that have nothing to do with you, and a name that people spell several ways loses mentions that do. Build the keyword list with the misspellings in, and exclude the obvious false matches, before you believe any figure.
The third is comparing across periods when the tracked sources, keywords or competitor set have changed. That produces movement that is entirely an artefact of the setup.
How to act on it
Start by fixing the measurement: an explicit competitor set, an explicit keyword list including misspellings, an explicit source list, and a period you will not change. Then track direction rather than the absolute level, because the level depends on choices you made.
To move it, give people something specific to mention. Published answers to the questions your category argues about, visible responses to reviews, named case work and staff who post under their own names all produce mentions that carry your name with them. Handle the negative share as carefully as the positive: a fast, plain, public reply is the cheapest reputation work there is, and it belongs inside online reputation management rather than being left to whoever notices first. Read the number alongside share of voice, which covers paid and owned visibility rather than public talk.
