Google Sets Challenged by Niches like “Home Health Monitoring”
Google Sets is an interesting new tool from Google Labs that finds associations between words or phrases and related terms that we might not have thought about. Who could be better able to offer such a tool than Google: More than any other organization on the planet, Google knows what we are thinking. When something is on our minds, chances are that a significant sample of us are typing it into Google. The resulting data puts Google in a unique position not only to identify associations between words, but also to derive insights into human intentions and behavior.
For obvious examples, type “Porsche” and “Mercedes” into Google Sets, and the site returns Ferrari, BMW, Audi and Lamborghini, among others. Type in “brain fitness” and Google Sets returns brain health, mind fitness, mental health, stress, cognitive neuroscience, brain training, and more.
For something less obvious, I challenged Google with the name of my old company, Health Hero Network. Google Sets accurately found AMGA, the acronym for the American Medical Group Association, the company’s partner on an important but not very well known Medicare chronic care improvement demonstration project. Google Sets also identified two of the company’s competitors, Viterion Telehealthcare and AMD Telemedicine, but missed the larger rivals Philips Telehealth and Honeywell HomMed as well as Intel Digital Health, the latest entrant into the home health monitoring market. Google Sets also didn’t pick up the fact that Health Hero Network is now part of Robert Bosch GmbH.
Digging deeper, I tested the phrase “home health monitoring”, the field of capturing health information from patients at home and sharing it with caregivers. This niche appeared to have stumped Google Sets. The website returned a number of terms related to nursing, but veered off into some bizarrely unrelated phrases like “buy sexy Halloween shoes” and “wholesale shoes deal”. Maybe Google has discovered a connection between home health monitoring and sexy Halloween shoes that is not yet obvious to the general population. Maybe this blog post is only going to contribute to the weird association in the Google search index. If enough people press the Digg button above to send this link to the web’s largest social bookmarking site, the behavioral anthropologists studying Google search data could see a stream of even more confusing data!
Not only is Google in a unique position to know what the world’s online population is thinking, but Google also serves as a collective memory for our thoughts and interests over time. In past blogs and presentations I have used Google Trends to illustrate how our interests change over time, illustrating the rise of sharing on social networks like Facebook, using web analytics to foretell the fates of presidential campaign contenders, and to lament the relative lack of interest in important topics like global warming compared to things like football and beer, as shown above. See my presentations on life sciences innovation, new chronic care models, and telehealth technology for more examples.
Google Labs is currently running an experiment that could extend our collective memory decades into the past with an new search filter called Timeline. For awhile now, Google has been indexing libraries and archives of books, news, and patents from the past. By looking for the date of origin in footnotes, endnotes and any other notable information in its database, Google may soon be able to give us a clearer sense for how our interests have evolved over the past century, not just the past year.
Take note of Google Trends, Google Sets and Google Experimental Search using Timeline from Google Labs to get more insights on just what Google means when the company describes its mission to “organize the world’s information and make it universally accessible and useful.”




