Some aspects I believe are missing from search per se are the words used during the buying cycle. Of course analytics provides various timelines, but as internet and the buying cycle is associated with different IP addresses for various searches now on various devices. Thus, heavy browsing may happen at home, some quick searches during the daytime at work and the adhoc reminders using mobile devices/notebooks.
A personal change in search, adverts etc could be changed to match business and consumer isp adsl packages. Indeed for isps to provide this information to ad searches is invaluable information. For advertisers it could mean different ads to business in the day to home an work isps, with an opt out option if their business model is for sme (ie a remote worker or sole business might use a personal email account for home and work). Isps might use this info to target users from different isps for acquisition for adsl and advertisers could find out which home domains provide the most conversions or type of customer.
For isps and email providers it is both a keen monitiser as well as risk. However from the leverage providing this information to analytic companies who in turn sell this onto the marketeer, this form of targetting will make email campaign analysis of domain openers a cost saving benefit. Wondering why gmail openers are less interested in your offers, well then separate via isp and then make a model for home users from btinternet who use gmail, brand up your campaign and make more sensible partnerships with their own campaigning. It provides another outlet and will help with locating the natural language of the recipient.