Already master in the acquisition and management of mostly unstructured data¹, whether it be web pages or word processing documents, Google is now taking on structured data as well.
Google Base, currently only available in English, allows users to insert data items which become searchable through a form at base.google.com.
16 types of data to insert – or specify your own
Items can be posted using templates for predefined Google Base item types (16 currently: blogs, coupons, course schedules, events and activities, jobs, news and articles, people profiles, products, recipes, reference articles, reviews, services, vehicles, wanted ads, rentals and, of all things, comic books). Alternatively, Google Base allows the definition of new item types and their associated data elements.
While the manual data entry interface is a bit cumbersome, an unneeded predefined data element (field) can be removed; user-defined elements can be added, for a total of 30. Predefined data elements often contain a pick-list of mandatory or optional values. The current list of data elements and values can be found examining the Google Base schema.
Bulk data upload is also possible using TSV, RSS 1.0, RSS 2.0 or Atom 0.3 formats.
Retrieving data: searching the hay stack
Enter a search word or phrase in the classic Google-minimalist search box at base.google.com to see Google’s highest ranking “Posted Items”. While this approach has served unstructured web data well, it would be nice to see the possibility to limit item types included in the search results. An advanced search page would be useful as well as documentation of power-user syntax.
Returned results may be sorted by relevance or by posting date (ascending / descending). A series of suggested filters, based on the popularity of labels assigned to the results items, allows for further search refinement.
From the Google Base home page it is possible to limit a search to a particular Google Base item type by selecting from a category under the heading “See items posted by others”. A search box appears along with a page of results for this item type.
Beyond the ability to define searches by “item type” or refine results by “labels”, one immediate difference compared to the standard Google search will be apparent to dial-up Internet users: results heavy with photos take awhile to appear.
User interface betrays Google Base’s Beta status.
The strong impression one has using Google Base is that the system, while promising, has been released in a partially finished state.
Inserting data online is a rather smooth process, although it would be nice to have a visual indicator noting the presence of a drop down list of suggested values for predefined elements. A traditional down arrow only appears when the value is restricted to the list.
Searching unfortunately does not yet take full advantage of the filtering permitted by the use of fixed fields when defining an item. If you start searching Google Base knowing you are only interested in vehicle listings, shouldn’t you be able to specify this item type as part of your query, all in one form?
A counter argument can be made that information is often found where one least expects it, even in structured databases. Google has been the master of providing relevant results from disparate, unstructured data. An advanced search form with an associated power user syntax would allow for a hybrid mix of structured and unstructured data query without compromising the proven success of Google’s relevancy algorithms.
On the content provider side, it would be nice to see statistics on how many times how many times an item was viewed for Google Base items containing a detail page.
Cross listing in the family
Google Base documentation notes that items from Google Base may appear elsewhere in the Google universe, “depending on their relevance”. Besides Google Search, the Google universe candidates for Google Base items include Froogle (now Google Products)and Google Local. As with the rest of Google, what appears where, and in what time frame, is unknown and ever changing.
Déjà Vu? Froogle (Google Products), since 2002.
In reality, Google Base appears to be an extension of the online product directory service Froogle. Froogle (now Google Products) is almost 3, having launched December 11, 2002. It is interesting to note that Froogle’s help pages are currently available in the Google Base Help Center. Besides extending the data schema, the other major advance with the release of Base is the possibility to manually enter items online. By removing the need to master batch data preparation and loading techniques, Google Base becomes more accessible to smaller, less technical content providers.
Who’s not rejoicing?
While content providers are merrily experimenting with Base, a Google patent application indicates an eye on the traditional classified ad market. While some of the user interface issues may hinder quick, mass adoption, Google Base has significant potential to disrupt a major source of revenue for the incumbent players of information intermediation — from the print world of daily newspapers in North America and second hand papers everywhere to the online world of Craig’s list, E-Bay and Homestore. Having more than proved it’s technical prowess, the extent of the disruption Google causes will depend significantly the business model adopted: a free Google Base, supported by AdWords? Or Google Base as a pay to play system? Or as a hybrid where just some categories (item types) are free? Beta in the Google universe can last a long time – yet now as a public company, Google’s every move is carefully watched.
Update 2009-06-08: The new Google Squared table representation of search results is certainly one solution to representing Google Base data.
Search Engine Optimization
While intended to promote “items” rather than sites, Search Engine Optimization (SEO) practitioners are busy considering Google Base’s ramifications on website promotion. Adult sites have already populated people profiles with “items” which probably won’t pass the “report a bad item” test.
¹ While html markup does structure a document to a limited extent, the structure is poorly enforced (titles and headings are optional). We also don’t really know what type of document it is. Is the html document an offer of services, a personal profile or a vehicle listing? This is the gap Google Base aims to close.
Similar Posts:
If you haven't already, you might subscribe to my feed by Email, RSS feed and/or follow me on Twitter, which is updated on a more frequent – and more meaningless – basis. Finally, if you're a Sphinn user, Sphinn love is welcome :-). Thanks for visiting!
Share
0 responses so far ↓
There are no comments yet...Kick things off by filling out the form below.
Leave a Comment
Warning: Comments are welcome insofar as they add something to the discussion. Anonymous and/or polemical comments without a rational justification of the author's position risk being mercilessly deleted at the sole discretion of the administrator. Yes, life is hard :-).