Tracking, the fundamental action to follow a reputation on the web. Reputation of a brand, an individual, a problem, etc. Useful for all agencies that pride themselves on selling “on-line reputation management“, “online watch”, etc.
Briefly, in methodological terms, it will be:
– to request as many sources as possible (for example with Blogbar , even if it does not offer an exhaustive view of the web)
– deduplicate the results
– categorize the results (according to their theme and the influence of their sources)
– extract relevant verbatim
– memorize the alerts and the bookmark (the requested sources, especially manually added sources),
– generate pie charts and other relevant indicators (map / chart aspect) for the customer; This is a particularly complex point because each problem requires its own pie charts, beyond the pie charts “type” (number of citations on a theme vs. number of global quotes).
– to manage problems of languages and studied corpora differentiated for international watches
Companies like Digimind or Mediaveille offer solutions, often quite expensive but allowing a process automation which allows the consultants to devote themselves to the exploitation of the data: to detect signals of crisis, to follow the inflection of a speech on a subject particular and – ideally – discover new words without having to re-feed the bookmark.
Other actors come to be placed on the real-time web-based web site and the cartography, like TrendyBuzz , which I have not yet seen at work on the aspects of deduplication, categorization, extraction, memorization