Influencing by Search

What does make you frustrated the most, when you look at a set of inconclusive results from a search engine? For most people it is the irrelevancy of the result; but sometimes it is the unimportance and trivialness of the result. Let’s look at this problem from another angle.  How does the result influence us? “Influential search,” is about searching for information that is most likely to make us take an action. If we are looking for a gadget and see one that is recommended by Steve Jobs -and is not an Apple product; no conflict of interest- we probably find it very appealing. A restaurant that has been given 5 stars by 90% of its customers would be another example of good search results, when we are searching for a place to take our family to dinner. In his book, Influence, Science and Practice, Robert Cialdini defines six basic categories of influence: scarcity, authority, social proof, liking, reciprocity, and consistency. Interestingly, some of these methods of influence are used by search engines, or search functionality of other online services, to provide a more convincing result to the users.

Since Google is the synonym for search in our era, let’s look at how Google’s search works. Of course, Google has a complicated secret sauce, but the essential element is finding authoritative sources. For example, for a piece of news, Reuters, CNN or BBC are authoritative sources, since most other news pages quoting them and referring to them. In fact, in many search engine niches, researchers are scrambling to find what source are the most dominant in a specific subject. A challenging problem considering the formidable number of subjects, and fluid and dynamic nature of authority in the world wide web, where everyone, at every second, can challenge the status quo authorities.

Another hot topic in online media is customization. The best restaurant for you might be a bad choice for someone in Mumbai who is looking for a place to eat: it is to far, to pricey, and lacks foods spicy enough for his appetite.  Simple customization of the search results, based on where do you run the search from and what has been your search history, makes both you and our fellow Indian friend happy customers. This is consistency, another influence factor, at work. Most search engines and online services (such as Amazon and Netflix) use cookies, or, if you need to sign in to use their services, your account history, to guess what choices are more consistent with your behavior and rank those higher in the results. Again, many armies of researchers, and lawyers, try to make this influence factor more accurate, and less threatening, to the users.

If you consider yourself a nerd, you have heard of collaborative filtering, a method that became famous after Amazon used it by telling its users what did the “customers who bought this item also bought.” This worked magic for many users when they stumbled upon items that they didn’t directly search for but were perfect choices for them. It is a major leap from the conventional search, where the users have to explicitly state what they are looking for. It is a form of discovery. And it is based on yet another method of influence: similarity and social proof. From stock market to following of celebrities, evidence is overwhelming that we are a herd species. We follow what majority follows, especially among the people who are more similar to us. And collaborative filtering is a way to find similar people to us by comparing their buying habits, or other habits. Again because there are too many people, following too many things, recognizing them and their trends is not an easy task except in very specific settings, like what Amazon and Netflix do in very well defined, limited, domains.

So far, we have covered three methods of influence that are consistent with the recent advances in search technology. When you are looking for something, and often times when you are not looking for something, you are pleased when presented with information that is from authoritative sources, or from people you can relate to. But how about the other three? Are their search technologies that use scarcity, or liking, or reciprocity?

Let’s start with liking. Assume that you are looking for a book to read, and you adore ,as weird as it my sounds, Steve Jobs. How about finding Steve Job’s favorite reading list? On the Internet, there is not an easy way to find out how does your favorite people behave, e.g., what type of food they eat, what type of sport they play, how do they deal with their personal issues, etc. One exception to this is those advertisements that use celebrities to persuade us to buy articles of interests to them. However, that is rarely genuine, and is limited only to a small group of celebrities. How about your other favorite people: friends, relatives, co-workers, the goal keeper in your local soccer team, and the cashier in your regular Starbucks? The most likely way to achieve this is to use the social networks that are formed based on some measures of “liking.” Facebook is a good example, since most of the people connected to you are friends and family; people whom you are likely to like. LinkedIn is not a good example, since in LinkedIn closeness is not necessarily a measure of liking but of “professional proximity.” LinkedIn is probably a better tool for measuring similarity, especially in work related attributes, than liking. In Facebook, however, people tend to express themselves more emotionally than professionally. Incidentally, or maybe not so, there is a “like” button on Facebook, where you can express your liking of other events or items. Facebook is situated to influence people by automating the process of discovering their favorite people and groups and offering them with the information that is consistent with those favorites’ behavior. Of course, there are many challenges, not least of it the privacy concerns that has been raised since the time that Facebook became a household name.

Scarcity and reciprocity are even more difficult to use by online media to influence masses. Reciprocity, in general, is an action oriented phenomena. We tend to reciprocate favors, and favors are usually actions. How can it be used when a person is searching for information? How can we make certain sources of information more interesting to a user by making him feel obligated to reciprocate an action? Though this method of influence hasn’t been fully exploited for search, there are rudimentary uses of reciprocity as a method of ranking search results in sites like Greenmaven. Greenmaven rank the more eco-friendly sites higher in its result, thereby reciprocating their good deeds by bringing them more business from people who are more environmentally conscious. Wherever there is more symmetry between searcher and searched population, for example in eBay or Linkedin, reciprocity could be utilized to help shaping people’s decision on who and what to promote. Simple tools like recommendations and star rankings are used to help people reciprocate others favors. We should see more sophisticated uses of reciprocity in social networking and auction sites in the years to come.

As for scarcity, it seems to be the hidden jewel not yet discovered by the online media. Scarcity -based search is about finding the one of the kind, the small restaurant hidden in the remote alley that only locals know about; the hairstyle that only one in a million wears; the new dance tricks that only a handful master. Scarcity is ephemeral in the Internet. If something is rare and interesting, it will, in a very short time, make a large following. The rare hairstyle becomes a heap fashion; the small restaurant becomes a chain; and the dance trick finds its way to “Dancing with the Stars.” Therefore, finding the rare things that have other elements of influence is the holy grail of search and discovery.  This could explain the recent popularity of real time search. The faster you can identify a potentially hot, but currently unknown, item, the more you can benefit from its later success