Does search violate Net neutrality?
The principle of Net neutrality is based on the end-to-end design principle: that innovation in a network should be determined by decisions at the ‘edges’. In the case of the Internet, this implies that the decisions of the content and application providers and the end user should govern its growth. The intermediary should be a ‘dumb pipe’ and play no gate-keeping role.
This principle is believed to be responsible for the galloping growth of the Internet. The number of websites has increased from 12 million at the turn of the millennium to close to 1 billion in 2015.
A vital element of the architecture of today’s Internet is the search engine that is at once a gateway into the Internet, allowing us to access sites of interest in milliseconds, as well as a powerful advertising medium bagging close to 35% of the total 2014 Internet ad spend of $135 billion as per a PricewaterhouseCoopers study.
All search engines operate on advanced versions of the ‘page rank’, an algorithm developed by Larry Page and Sergey Brin, co-founders of Google, now enhanced by a complex set of additional innovations including language models (the ability to handle phrases and synonyms), query models (the ability to handle everyday syntax), time models (whether to serve a 15-minute-old page or one that has stood the test of time), and personalization models (since everyone does not like the same thing). Google has captured 65% of market share in the field of search. Its core search team consists of a small group of extremely smart people headed by Amit Singhal.
An Internet user can end up at an Internet site in one of four ways: by directly typing the URL, going through a search engine, via social network referrals or via a miscellaneous category of other pathways that include message boards and email. A 2014 study of 26 top news websites in the US conducted by Pew Internet Research shows that search referrals account for a sizeable chunk of traffic, (although those directly accessing the website end up spending far more time). Overall, it would be fair to say that the search box is a very significant gateway to the Internet.
Estimates of the percentage of traffic coming from paid search referrals vary from 6% to 20%, with the percentage being higher in fiercely-contested industries like travel, which account for a large proportion of the ad spend. Does the institution of paid search, where a company pays an intermediary to get additional eyeballs, raise barriers to entry for firms that do not have the required marketing budgets? Can this be construed as violating the principle of Net neutrality?
Even the institution of organic search merits closer examination. I suggest a simple experiment: go to the Google search box and type in ‘Delhi to Goa flights’ ( or whatever is your preferred choice of trip). Your screen will fill up (in fractions of a second) with returns that either say ‘ad’ or bear the legend ‘sponsored’. The sponsored returns are comparison-shopping services of Google, while the ads reflect vendors who have paid Google for preferred exposure. There will not be a single organic search return in view unless you scroll down (you are right, nobody does).
The European Commission is in the process of investigating charges that Google skews results to favour its comparison-shopping services. After a two-year probe, the US Federal Trade Commission concluded that Google’s practice of favouring its own content in the presentation of search results, sometimes referred to as “search bias”, did not violate US anti-trust laws.
According to Harvard Law School professor Marina Lao, the difficulty of identifying the entity that would determine if a search algorithm was ‘neutral’, the competition Google faces from ‘vertical search engines’ like Amazon, and the possible harm accruing from interfering in the natural process of competition between companies like Amazon, Apple, Facebook and Google makes the US judgement an appropriate one.
In other words, even apart from paid search which is quite evidently non-neutral, the paradigm of neutral search itself has been questioned in the recent judgement.
Finally, the use of personalization rules in serving search returns implies that we get to see what the search engine thinks we would like based on surfing information it has gathered about any people, rather than some ‘objective’ listing. The advantages of personalization come with accompanying concerns on loss of privacy, and the ceding of influence to a bureaucracy of smart people managing the core search engine platforms.
We can either support the end-to-end design principle or the architecture of the Internet as it is evolving today, but not both. We have reached a tipping point where the concept of Net neutrality requires a radical makeover. When a hallowed concept outlives its time, it becomes hostage to vested interests, each attempting to impose their self-serving interpretation. This leads to contradictions and doublespeak. The current debate on Net neutrality is replete with instances of both.
One wonders what Vint Cerf, grand old man of the Internet and fierce advocate of the free Internet, would think of all this from his perch as Chief Internet Evangelist at Google.
Rohit Prasad is a professor at MDI Gurgaon.
Comments are welcome at firstname.lastname@example.org