By Usman Latif
Published: August 16, 2012
Search is a multi-billion dollar business which has seen dramatic innovations over the years. The innovations have not touched the search interface though. Google, the most popular search engine, essentially uses the same user-interface it started out with in 1997.
In the technology world where things get outmoded by the day, this is an amazing record for longevity. This record is all the more surprising as the classic web-search interface is not exactly a model of usable design.
To perform a search, the user starts by typing the url of a search engine website, she then waits for the website to load, types her query, waits for search results to appear, and finally visits the website(s) of her liking. The first two steps in this chain of steps are the most time consuming and they are also unnecessary. The other steps can be optimized as well, so it is not clear why normal people put up with such an inefficient and outmoded interface.
By now, you must be screaming in your head that nobody goes through the whole sequence of steps as outlined above. Some people use the address-bar, others have their favorite search engine set as their homepage, and still others use a search-bar to reduce the number of steps they have to go through.
This assertion is true, but it only validates the claim that the classic web-search interface is too much work for a normal person. It forces people into using search shortcuts in the form of toolbars and search homepages to bypass the full chain of steps it forces on the user.
Things wouldn't be so bad if the search shortcuts that people are so fond of did not have a downside. The search shortcuts do work and this is why people use them but unfortunately they create a much bigger problem than the one they solve.
The search shortcuts generally eliminate two steps: the typing of the search website address and the waiting for the search website to load. They eliminate these steps by directly forwarding the user's queries to a default search engine.
The key here is the default search-engine: the search shortcuts can't function without a default search-engine. However, once you have a default search-engine, the game is rigged against every other search-engine. A user querying a non-default search-engine has to bear the full cost of the traditional search user-interface. That cost as we have seen previously is nontrivial. Consequently, users tend to stick to just the default search-engine for all their search needs.
This has dire consequences for the competitiveness of the search business. It creates potential for anticompetitive manipulation of search markets: a search engine provider can dominate a search market by having its search engine set as the default search engine on the web browsers used to generate the search traffic.
Users can mitigate the potential for such abuse by changing the default search engine on their browsers to something different but the average user is not knowledgeable about her search engine options. She is usually not even aware of the existence of any search engines besides the one that came set as the default search engine on her browser.
Things wouldn't be so bad if the risks of anticompetitive collusion were only theoretical and there was no actual manipulation of search markets but this is not the case. There is clear evidence that the competitiveness of the search business is being hurt by anticompetitive agreements between browser vendors and search engine providers.
The first thing to understand is that any agreement where a search engine provider is paying a browser vendor to have its search engine set as the default search engine in the browser address-bar, search bar, or set as the browser homepage is anticompetitive. Such agreements, euphemistically referred to as traffic acquisition deals, exclude all but the most cash rich search vendors. A cash-strapped search startup can never hope to cut such a deal and it gets hurt when established search players make such deals.
Take the example of the search royalty agreement between Mozilla and Google announced in December 2011. According to a report by AllThingD covering the announcement:
The search giant [Google] will pay just under $300 million per year to be the default choice in Mozilla’s Firefox browser, a huge jump from its previous arrangement, due to competing interest from both Yahoo and Microsoft.
The deal effectively shuts out search startups and marginal search vendors from vying for search traffic from Firefox users. Google has a similar deal with Opera as well, reportedly to the tune of $50 million a year.
It makes perfect sense for Mozilla and Opera to cut such deals as such deals constitute their primary source of revenue but the rationale for Google is suspect. Google has its own browser and the Google brand is one of the best known brands in the world. Consequently, such deals only make small marginal contributions to its revenue. The real motivation for cutting such deals can only be shutting out competition from the search business.
Such deals are only the tip of the iceberg. The real anticompetitive maneuvering takes place because of vertical integration in the search and browser markets. For instance, Microsoft and Google own two of the most popular browsers and two of the most popular search engines as well. Owning both products allows them to set the default search engine on their browsers to the search engines they own.
Such behavior can be very damaging to the competitiveness of the search business. A recent report by the Guardian newspaper highlighted complaints by Ilya Segalovich co-founder and chief technology officer of the Russian search vendor Yandex that Google is shutting out its rivals. The report says:
Segalovich suggested Google was guilty of foul play with its Chrome browser, which he said made it difficult for users to choose rival search engines, including Yahoo, Bing and Yandex, over its own market-leading product. He said internet users were effectively coerced into using the Californian internet giant's products over those offered by rivals.
Segalovich does not complain about the issue of traffic acquisition deals as Yandex itself engages in many such deals to the detriment of smaller search vendors. Evidently, in the search business, there are no "ethical" players: a search vendor will try to have an advantage over its competitors by whatever means it can muster as long as they are "legal".
Currently, every US search engine other than Google is either struggling or on life-support. Other search markets have different dominant search players, but the trend is the same. Russia has Yandex leading with greater than 60 percent market share, China has Baidu leading with 56 percent market share, South Korea has Naver with over 70 percent market share, and Japan has Yahoo in front with 51 percent market share. These markets do have other search-engines but they are heavily propped-up by deep-pocketed players. Overall most of the world's search markets are highly uncompetitive.
With things as they are, effective competition can not emerge in the world's search markets by itself. The dominant players can pay through the nose for traffic acquisition deals and keep themselves in control. This is likely the reason why most search markets have a dominant search player in the first place. Addressing this issue is clearly important for the health of the search industry.
Some economists would argue that the search business is a case of "market failure". A market failure occurs when a free market fails to provide as many goods and services as it ought to be doing. In the case of search, we have a lot fewer search vendors than there ought to be.
A market failure is typical justification for government intervention. It seems the search business could also use some form of government enacted regulation to even the playing field for the smaller players.
Unfortunately, government regulation is rarely a good fix for market failures. Market failures are a controversial concept to start with. Making the case that a particular market has failed is hard. Worse, government regulators tend to be clueless about the problems they are trying to solve. They tend to time their regulation efforts poorly, enact useless regulation, yield to political pressures, and at times manage to offer legal protection to the parties most responsible for anticompetitive behavior.
Expecting the government to address the issue in a timely and helpful manner is wishful thinking at best. Moreover, the search business is global and regulating it in an effective manner would require sensible regulatory action from very many governments. Surely, this is not possible.
Fortunately for the search market, its "market failure" can be fixed without government intervention. The key to the resolution of the market failure is the default search-engine. As previously discussed, to get around the inefficiencies of the traditional search interface one search engine needs to be made the default search engine. Once this is done the default search engine gets to handle the lion's share of the user's search queries. This in turn opens the way for anticompetitive traffic acquisition deals.
If we can design a search interface that is efficient and accords equal status to any search engines the user may want to use, we can trivially address the lack of competitiveness of search markets.
Fortunately for us, such an interface has already been designed. The following screenshots show one such search interface design.
The interface consist of a toolbar with multiple search buttons. A user can click a search button or press its shortcut key to open a search menu. Search results appear in the search menu as the user types.
The new design does not require the user to type the address of a search website and it does not require the user to wait for the search website to load. It eliminates the first two steps required by the traditional search interface, and is at least as efficient as any search interface designed to date.
There is no privileged search button and no need for one.
Search history is centrally managed and can be shared across search engines. Even better, with a query entered, a user can migrate the query to another search engine simply by clicking another search button. This feature is super useful when the user fails to find relevant results with her first choice search engine. This feature encourages the user to experiment with niche search engines as she can trivially repeat the query with a regular search engine whenever the need arises for that.
The design allows the user to make a selection in a webpage and turn into a search query by clicking a search button. With this feature a user can lookup the definition of a word with three clicks: a double click to select a word and a search button click for the search.
The design even integrates bookmarks and user history in the same toolbar. Bookmarks and user history are simply treated as search buttons.
The design treats even the navigation toolbar as another search button. This demonstrates the conceptual elegance of the design.
In fact, the new design integrates all the major browser toolbars into one universal toolbar. The universal toolbar does everything that to date required multiple toolbars and requires no more space than a modern browser's navigation toolbar. The new design has many other interesting features that are documented in another article on this website. The design is not even complete yet and has considerable potential to deliver a much improved search experience over the large usability gains it already demonstrates.
Compared with this new design, the traditional search interface is a hulking dinosaur. It is inefficient, feature deficient, and anticompetitive. Browser vendors need not implement the new design as is but there is no longer any excuse for offering search shortcuts based off the traditional search interface as the primary search interface to the user. The world will be a better place the sooner it lets go of the traditional search interface and all the inefficiencies it entails.