New York – Recently, Google released a product named Google Squared which shows an interesting way forward for the future of search technologies. This product is seen as a competitor to Wolfram Alpha, a computational engine which was launched in 2009, but it also has competitive implications for research and data mining companies that currently sell professional web data extraction services for business and financial clients.
Essentially, Google Squared is a tool that attempts to extract structured data from the web and presents the data in a spreadsheet format. Each search query returns a table of search results. This table is organized into rows for each entity, and sortable columns representing common attributes that are associated with the topic of the search.
For example, a Google Squared search for “US states” provides us a table listing each state, with columns containing information for state attributes like state capital, state motto, state map, etc.
Google Squared is still very much a beta product, more a technology toy than a research tool. It is clear that it was not built with financial or economic research in mind. For instance, the “US states” search does not present the kind of economic data that an analyst might be interested in – GDP, household income, population, etc.
In contrast, a Wolfram Alpha search for “US states” leads us to a far more analytically useful data set, presented in a clean and well organized fashion.
Similarly, an attempt to compare stocks (like: “AAPL IBM MSFT”) on Squared leads to a useless page presenting no results whatsoever, whereas Wolfram Alpha automatically builds a very informative tearsheet comparing the three companies’ financial attributes.
Although Google Squared is far from ready for prime time, we presume that Google could throw development resources at the tool to make it competitive for serious research. Google does have a tendency to abandon projects in mid-development, so it is not clear whether Squared will ever develop into a real research tool. Wolfram Alpha also had a rocky start, with users complaining about irrelevant results and a difficult interface; however, more than a year after release, it seems to have been built into a highly polished and powerful research tool, aimed primarily at mathematicians and other academics, but also useful for anyone doing research on financial topics. Wolfram Alpha is also available as a mobile app for iPad users – the idea of instant, mobile access to powerful research tools is very attractive. From our perspective, the main weakness with Wolfram Alpha is its occasionally quirky interface (one has to spend time “learning” how to build queries on Wolfram Alpha); another weakness is that it only makes data downloads available in Mathematica or PDF format – some additional data export options would be very useful
The concept behind Google Squared and Wolfram Alpha is not new. Other firms, like Connotate, Kapow, Mozenda, Firstrain, and Denodo, build professional tools that enable the structured collection of web data. These companies seem to provide a far higher level of professional service, design options, and end user support than Google is likely to provide – professional services has never been one of Google’s strong points. However, Wolfram Alpha and Google Squared have the potential to be very competitive for data collection at a cut-rate (or free) price. These sites have the potential to put some serious research tools into the hands of small businesses and investors – tearsheets that would have taken non-trivial amounts of time to put together can now be built within seconds by Wolfram Alpha. At the same time, the data extraction companies serving the higher end of the market will be looking to differentiate themselves on the customizability of their product, the level of support they provide, and the sophistication of the data collection they can carry out.
1 Comment
The key issue for business people is whether the system (search and analysis) they are using understands their world or not. John Battelle has written about the web is moving from intentions to insights and technologies continue to develop, like Google Squared, to create structure out of web data. But to get useful structure you have to make an assumption about the end domain. Just put “Cisco” or “networking” into Google Squared and you’ll see the problem.
The win for end users – especially business and research users – is when the system knows the end domain and can create business structure (and so useful insights) from the web. This is discussed here: http://www.huffingtonpost.com/penny-herscher/the-latest-web-search-dev_b_636854.html