Google Advisor and the Growth of Semantic Tools

Google introduced its “Knowledge Graph” last week with many heralding a new era of semantic search technology.  As I wrote last month, semantic search isn’t new and has been a work-in-progress since 2003.  More importantly, semantic search’s implications for the internet, data and e-commerce go way beyond this new feature.

Take for example, Google Advisor.  This Google feature is designed to help users “find financial offers from multiple providers, compare them side by side, and apply online”.  Google Advisor, similar to Google Flights, offers customers a comparison service all within the confines of Google. These comparison services offer a new referral revenue stream in addition to Google’s current PPC advertisement model. Okay, so it’s interesting that Google might be pivoting towards an ad revenue model lower down the purchase funnel, but what does that have to do with semantic search?

Well, let’s look at the note that appears next to the Google Advisor toolbar:

“Based on your search query, we think you are trying to find a credit card. Clicking in this box will show you credit card providers that can fulfill your request.”  Google attempts to guess the intent of your search and then offers you a customized tool for that purpose- can’t get much more semantic than that.  Currently, Google Advisor returns for queries as varied as “best credit cards”, “compare credit cards”, “airline miles credit cards”, and “credit cards for bad credit”.  You can be sure that Google will monitor the CTR of these and other terms to further hone their semantic intelligence around credit cards to determine when it should present Google Advisor.

The Demise of Third-Party Aggregators?

The most obvious loser in the case of Google Advisor is third-party aggregator and comparison sites that make money from credit card referrals.  Depending on the popularity of Google Advisor, sites like creditcards.com, bankrate.com and nerdwallet.com that traditionally dominate search engine results may quickly be outfoxed and outgunned.  If Google Advisor conquers comparison shopping while official credit card websites (Chase, Bank of America, etc.) have detailed product information, these third-party sites will need to find value in the tight space in-between.

This is well-trodden ground for Google who entered the travel comparison industry with Google Flight Search last year in an attempt to displace Expedia and other online travel agencies.  How successful has Google been so far in this market?  Experian Hitwise found that from the September 2011 beta-launch to March 2012, Google Flight Search was the 57th most visited travel agency website, with over 508,000 visits.  Kayak.com (5th place) recorded 21.5 million during this interval and Expedia (1st place) received 46 million visits.  Google hasn’t yet destroyed its flight competition but visits to GFS have grown 300% since September.  Google Advisor will most likely follow this trajectory, aiming for a measured, long-term strategy in the credit card industry.

What Data Fuels Google’s Semantic Search Tools?

Garbage in, garbage out- so the saying goes.  Semantic mapping of industries and topics is not an easy task and so far Google has used 3 different methods for gathering this data:

1. Buy It: Google has grown its database of linked entities for years through acquisitions.  In 2010, Google bought Metaweb and gained new semantic search technologies as well as the company’s enormous open-source database, Freebase.  When Google wanted to create Google Flights it simply bought ITA and all of its structured flight databases for $700 million. In a blog post last week, Google guru Amit Singhal claimed that as a result of these purchases Google’s Knowledge Graph contains more than “500 million objects, as well as more than 3.5 billion facts about and relationships between these different objects.”
2. Borrow It:  Google acknowledges that it uses open source databases like Wikipedia (which is developing more structured metadata schemas) and the CIA Factbook to fill its data reserves.  Google could also attempt to scrape regular webpages for information but since comparison tools would need to be 100% accurate and up-to-date, this strategy would likely be too risky.
3. Leverage: Google deftly uses its dominance in search to create better data for itself.  Google has started displaying rich-snippets in search results if sites use semantic mark-up language like schema.org or RDFa in the HTML of their pages.  These rich-snippets improve click-thru rates for websites and provide a free, scalable solution for Google to collect structured data.  Google also leans on its massive paid ad network to create structured data for itself; Google Advsior doesn’t list every credit card because it only has comparison information from credit card companies that use Google Comparison Ads.

5 Predictions for Google Semantic Search:

1. AdWords is Google’s bread and butter but as the search engine offers better semantic answers it may see the majority of its revenue come from referral based fees.
2. Google semantic tools will fight the app-model adopted by smartphones/tablets and create one singular Google experience that can answer questions as well (or better) than single-purpose apps.
3. Antitrust lawsuits, like those caused by Google Flight Search, will continue to hound the company as its attempts to simplify the search process will invariably wipeout existing websites/industries.
4. Google will adopt a pay-to-play system (see recent Google Shopping policy change) for companies and products to be listed in all Google comparison tools.
5. As Google collects both open-source and company-specific data, there will be serious questions about data ownership and Google’s use of this information across its products.

 Conclusion:

Google Semantic Tools are a big deal.   I believe Google will successfully create scalable data collection techniques and use its position in the search industry to popularize these products.  Google will cut through the overcrowded web and use semantic technology to connect users to products and information with an efficiency we have never seen before.