Competitive intelligence from advanced Twitter search

We wrote before about getting competitive intelligence out of Twitter. In fact, there is a lot more information available, so let's look at how to extract that intelligence.

First port of call is Twitter's own advanced search function. The options are limited, and need a lot of manpower to read through the results. Searching for the #gonegoogle hashtag is a good source of Google Apps customers. Tweeters within 10 miles of a competitor's headquarters (say near Nike's headquarters at Beaverton, OR) may well work for that competitor. That kind of search is comparable to the needle in a haystack, but the needle does exist, if you are prepared to look for it. Twitter also allows searching for tweets to, from or about a person, which is a good way to search someone's ecosystem. For example, tweets about Microsoft developer evangelist Bruce Kyle.

Beyond Twitter's own search, a number of small third-parties search Twitter profiles and bios, where most of the juicy information exists. Some of these are not aimed at competitive intelligence gathering, but rather are used by people whose goal is to amass followers - bio search helps them target, say, developers or parents. It is possible (which is all that can be hoped for as a starting point) that someone who mentions a company in their profile is an employee, customer, reseller, etc. So, for example, Twellow shows these profiles mentioning ERP vendor Epicor and FollowerWonk shows these profiles for Microsoft resellers. One of the most comprehensive at the moment, although still limited, is Twiangulate, shown here with Accenture profiles.

It is best to use multiple search engines for the same search term. Whereas using Google and not Bing or Yahoo! is often good-enough, the nascent, limited and bootstrap nature of these search engines is such that the results can be very different and complementary. For similar reasons, results often include a lot of noise around the signals: as in this Tweepsearch data for Pfizer profiles. Twitter is a good example of a data source that has a lot of information, but needs careful analysisi to extract it in an efficient way.