Imagine you’re a teacher who assigns your students a report on, say, Alaska. One group of students uses only search engines. Because they’re diligent students, they don’t plagiarize Wikipedia articles, but actually visit several websites and write a report based on all the information they perused.
A second group of students uses word of mouth. They speak to librarians, contact people they know in Alaska and follow Twitter users who employ the hashtag #Alaska. Which group of students do you think will produce better reports? All else being equal, let’s assume a better report means one containing original information and insights.
For Alon Sela, a PhD candidate in Industrial Engineering at Tel Aviv University, this question lies at the heart of his doctoral dissertation. Sela and his advisor Professor Irad Ben-Gal, now on sabbatical at Stanford University, have studied how information spreads in society. Using mathematical models, they concluded that when people use a search engine, even one as sophisticated as Google, the variety and breadth of information a society is exposed to is significantly narrower than when the source is other human beings.
This conclusion may either seem blindingly obvious or patently wrong to you, but in a world where many of us spend much of our day on the Internet, it’s an important discussion to have. In an era when artificially intelligent algorithms replace humans in industry after industry, it’s crucial that we examine whether we are inadvertently narrowing humanity’s sources of knowledge instead of increasing them.
“By information I mean knowledge, ideas or new products,” Sela told the Times of Israel. “There are two ways these are passed between people. One is word of mouth, you heard something a friend said and pass it on to another friend. The other way is you hear about a new concept or thing and Google it.”
Sela agrees this division is simplistic, because most of us do both, nor does it account for mass media and advertising.
“We left those out for the sake of simplicity, but in general mass media can be viewed as someone connected to lots of other people.” Similarly, he said, social networks are a digital version of word of mouth.
Sela and Ben-Gal proceeded to ask one group of subjects to ask their friends a set of questions. Some of the questions ithey asked were:
1) Give me a name of a good restaurant in Tel Aviv.
2) What’s the best way to reduce home electricity consumption?
3) What are some recipes for fast (less than 30 minutes) nutritious meals?
4) How does one spend a rainy winter day with small kids?
Researchers asked the second group of subjects to look for the answers on search engines.
“People who used search engines came up with a narrower range of information for most questions than people who got their answers through word of mouth,” they found.
Why Google is not a cornucopia
Sela and Ben-Gal also mathematically modelled the two different ways of disseminating information.
Sela explained to The Times of Israel the mechanism that leads to a narrower range of information from search engines. In the United States, said Sela, close to 65 percent of Internet searches are conducted on Google.
“If it’s not a monopoly it’s almost a monopoly.”
On top of that, 85 percent of people don’t go beyond the first page of search results.
“How does Google decide which results will appear on the first page? It looks at who links to you. If there is a link from CNN or the New York Times to my paper, it will rank higher in the search results. If the link is from a blog that no one has ever heard of, then my chances of ranking high are smaller. This creates a feedback loop where the more famous you are the more famous you become. The rich get richer and the poor get poorer.”
Sela and Ben-Gal have experienced a lot of pushback against this claim.
“Some people said no, not at all. Google customizes the results for everyone so if you do a search and I do one we see totally different things.”
But even with this customization, says Sela, “a hypothetical society that only uses search engines will see less information than those who get information through word of mouth, even digital word of mouth like Facebook or Twitter.”
How our brains accept new ideas
According to Sela, a person typically hears a new idea, say, that global warming is a grave danger or that digital wallets are a great way to carry money. At first they will say, “interesting, maybe.” What induces them to make the idea their own is when they hear the same idea repeated from different sources.
“From an evolutionary point of view this makes sense. If separate people came to the same conclusion independently, then the idea is more likely to have merit.”
The problem with search engines, says Sela, is that more and more people are seeing the same information, and if a piece of information starts out obscure, it has less chance of rising to the top.
That’s why, if you’re a small business or a struggling academic, your odds of spreading your ideas are much higher through social networks, whether those are digital or face-to-face.
“Let’s say you’re a musician or holistic medicine practitioner. You don’t have money for SEO (search engine optimization) but you can still gain followers through word of mouth.”
Is this because Google is centralized and social networks are decentralized?
“Right. The decision of who to be friends with is my decision. Even though Facebook’s servers are centralized and Facebook can push some information, still, the information I’m exposed to depends on the friends I choose.”
In a way what you’re saying is intuitive. It might explain why people prefer living in a city to a suburb. You can walk along and see small stores, as opposed to walking in a mall.
“In a mall you have stores like Zara, Mango and Fox. They’re the same in every mall. But in a small store you might see something new, something the shopkeeper imported from Romania, from Afghanistan.”
Yes, maybe that’s why rents are so high in Tel Aviv, where you feel like there is a chance to discover new things, to be surprised. Is a mall like Google while a city is a social network?
“I agree that malls are monotone. That’s also why, as a researcher, I like to go to conferences. My job is to discover things that are not known. I can go and meet someone who lives three streets away and has something to say. That’s word of mouth. I would never have found them through Google’s algorithm.”
Monetizing their research
Sela says that he and Ben-Gal, together with a third collaborator, Dr. Erez Shmueli who teaches at Tel Aviv University and the MIT Media Lab, have developed an algorithm that uses the power word of mouth to help people and businesses disseminate ideas.
“It’s a recommendation engine that tells you who in your network to approach if you want to sell a product or idea.”
Yes, but aren’t there already programs that scan social media on behalf of marketers, like SocialBro and Tracx?
“Most of those programs tell you who is is talking about what, but they don’t give you feedback on who to turn to. Most social media marketing tools tell you to look for Influencers or thought leaders. But it’s hard to get to those people. Lady Gaga and Benjamin Netanyahu are flooded with requests. Our algorithm will tell you that you need to approach the CEO’s two best friends rather than the CEO.”
So far, Sela and Ben-Gal have received a grant from the Office of Israel’s Chief Scientist for their patent-pending algorithm. They are currently looking for a company with access to their customers’ social network to help them test the algorithm in the real world.
Why is diversity important?
Imagine there’s a shakeout and eventually most of the world relies on BuzzFeed, the Huffington Post and a handful of other news sources for information. Is that such a bad thing?
“A question we asked,” said Sela, “is whether biodiversity is important. I think that’s a great metaphor. In biology, the answer is that if there are changes to our environment, you have a greater variety of DNA that can survive. If only dinosaurs existed and a meteor hit, everything would have been destroyed. But if there are beetles, lizards and birds as well, life can go on.”
As with biodiversity, so too with ideas.
“This is a bit lofty and not part of my empirical research,’ Sela avers before explaining that in the study of “collective intelligence,” also known as the wisdom of crowds, researchers have found time and again that if a group of people make a decision, the average of their decision is often better than the decision of any individual. For instance, in 1907, Charles Darwin’s cousin Francis Galton noticed that when a crowd of people were asked to guess the weight of an ox, the average of their guesses was better than most of the individual guesses, including those of experts.
With one caveat.
“Ideas have to be independent. If people’s ideas are dependent on each other, you get the stupidity of crowds, you get Naziism, fanaticism and stock market crashes.”
What separates the wisdom of crowds from herd behavior, he explained, is that each individual gets her ideas in a way that is independent of the next person and that these ideas are as diverse as possible.
“Take the stock market. Once we had thousands of economists and each said what they thought. Their collective intelligence was the best prediction. Today, three big financial sites are Yahoo Finance, Google Finance and CNN. Even economists pay attention to those sites. So what if all three sites suddenly say the situation is a catastrophe and you need to get out of the market? We will have a panic.”
Yes, but it seems as though the world is headed away from diversity and towards globalization. Some people would say that’s a good thing. Why do we need separate nations and languages? It just causes friction.
“This is a question you need to ask the Creator of the Universe. I am sitting in my sukkah at the moment. Maybe because we had too much hubris? This question is beyond me. But on the economic level, look at Europe, they wanted to be one big body, but the Germans and the Greeks don’t see eye to eye. Or look at the human body. We’re not made of identical cells–we have a brain and hands and a digestive system and each system has strengths and weaknesses. Why aren’t all cells the same? Why aren’t we just all just amoebas swimming together in the sea? Probably that’s not the most successful strategy. Probably differentiation serves a purpose. But this question is above my pay grade.”
From grand theory to app
The Times of Israel asks Sela, why, when he has hit upon an important global problem, the threat to diversity in the age of the Internet, he applying those insights to create a marketing app.
But Sela avers that this is the right career step at present.
Sela said his thesis has encountered some resistance, that not everyone agrees with his conclusions. But he does offer some tips on how, at the individual level, one can strive to keep ideas fresh.
“What helps me is to look for information on other search engines, and also to look on social networks. I like to use Twitter.”
Yes, but if you do a search on Twitter, isn’t that the same as a Google search?
Yes, the first time you and I do a Twitter search, we will get the same results. But then we each find people we want to follow. Some people generate too much noise, so you stop following them. You keep adding and cutting branches. Within a month our networks will be totally different.
Ultimately, said Sela, this is the meaning of word of mouth. “It’s where I get the best information from, from people I find interesting.”