There’s been a heated online debate about the ways consumers are influenced to buy things and adopt new social trends. Some people say a small group of Influencers drive most consumer decisions. Others argue that ideas spread through society from random starting points, without a hierarchy. The evidence shows that both groups are wrong in important ways, especially when the web is taken into consideration. This has huge implications for how companies market online.
For years, many marketing thinkers have told us that the best way to influence consumers is to work through Influencers — a small group of early adopters in society who try out ideas and then recommend them to all of their friends. The idea goes back several decades, to sociological studies on things like the spread of new seed types among farmers in the Midwest, and the adoption of hygiene practices among villagers in the developing world.
The researchers noticed that some people in those social settings were opinion leaders who had high social status, and whose decisions swayed others around them. If the opinion leaders adopted new practices or products, they were quickly copied by others in the society. On the other hand, some other people who tried new ideas were social misfits who had little credibility with others. When they embraced an idea or practice, it did not spread.
The researchers concluded that it’s important to target the opinion leaders if you wanted to get people to try something new. Everett Rogers wrote up the idea in 1962, in the book Diffusion of Innovations. He and others popularized the bell-shaped adoption curve and the idea of Early Adopters, work that has been revisited repeatedly since then, in books including Geoffrey Moore’s Crossing the Chasm (link), Malcolm Gladwell’s The Tipping Point (link), and Berry & Keller’s The Influentials (link).
“One American in 10 tells the other 9 how to vote, where to eat, and what to buy.”
–Ed Keller & Jon Berry, The Influentials
When you apply the Influencer idea to marketing, you get a very hierarchical view of the world — spend your money courting those early adopters, and they’ll carry your ideas to everyone else:
The counterattack: Welcome to the swarm
Recently the influencer idea has come under heavy attack from some prominent people online. One of the first was Duncan Watts, a researcher at Yahoo who ran a series of experimental simulations on the spread of ideas. Watts’ conclusions were reported in a widely-read article in Fast Company earlier this year (link; you can read one of his papers here).
In the article, Watts said that the whole idea of “influencers” is an after the fact rationalization of an almost random adoption process. Ideas, he said, can be advocated first by anyone. If society is ready for the ideas, they will spread. If society is not ready for them, the ideas won’t spread. The so-called Influencers are just people who happened to be the first to stumble onto an idea that was ripe for adoption in society.
“Whatever influence these individuals exert on the collective outcome is an accidental consequence of their randomly assigned position in the queue.”
–Duncan Watts, Yahoo
Watts’ ideas, as portrayed by Fast Company, paint a picture of society as a swarm of randomly-interacting bits:
If you want to spread an idea, the best thing you can do is seed it broadly. Don’t bother trying to identify Influencers; just seed as many people as you can, and hope that your idea will take hold and spread.
The shotgun vs. the rifle
For a marketer, the most troubling aspect of these two ideas is that they give 100% opposite advice on marketing strategies. The Influencer advocates say to focus on intense communication to early adopters, using targeted media. It’s like firing a sniper rifle. The swarm advocates say to use shotgun-blast mass marketing techniques to reach as many people as you can at low cost. Each camp argues passionately that using the other’s techniques is a waste of money.
There’s a third perspective: They’re both wrong.
For my team and I, the Influencer and Swarm theories in their pure forms both fail the “sniff test” of whether they make sense in the real world. And they don’t match what we’ve found in our own market research. Here’s why:
Of course some people have more influence
Consider, please, the case of Paris Hilton, postmodern media starlet, a woman famous mostly for being famous (link). No matter what you think of Paris’s taste in boyfriends (or much of anything else), you have to admit that she’s a trend-setter. If she wears a certain style of dress, or drinks a certain vodka, it’s extremely likely that a lot of people will imitate her. She’s not just a random influence, because her effect on fashion trends is repeatable. That’s why companies will gladly pay her to wear their clothing or hang out in their club.
There’s no way you can convince us that Paris is not an influencer, no matter how powerful the computer simulations you run on the servers at Yahoo. That’s good news for the Influencer camp.
However, there’s a catch.
“There’s no mystical energy field controls my destiny.” –Han Solo
Classical Influencer theory says the same 10% of the population drives all consumer decisions. The influencers are supposed to be social leaders who are looked up to and imitated by everyone else in all fields. That may or may not work in a traditional village in Africa, but it breaks down into absurdity in a large media-driven society with its blizzard of different connections. Paris may well be an influencer in fashion, but she has zero influence over the market for, say, home gardening tools. In fact, if Paris endorsed a lawn aerator she would probably drive customers away from it.
In a modern, connected society, people can and do look to different thought leaders for different subjects. The world isn’t a hierarchy, it’s a wildly interconnected network in which an influencer for one subject is a follower for another. That makes it very hard to identify the influencers for a particular decision or market. So maybe the Swarm folks are right after all, and you should break out the shotgun.
But then factor in the Web, and the picture changes again.
Gravity and radiation: A new form of influence
The old models of influence were all built around word of mouth communication: One person individually shares an idea with someone else. Even Watts’ critique of the influencer idea assumed a word-of-mouth process in which an influencer could reach up to a few dozen people. The Web has fundamentally changed that situation in two ways:
— One person can now potentially communicate to thousands of other people, and
— Influence is separated from conventional social status. (As the old joke puts it, on the Internet no one can tell you’re a dog. Or that you’re Paris Hilton.)
So let’s modify the picture of society we drew earlier. Rather than being a completely scattered swarm, or a rigid hierarchy, the web lets people cluster according to their interests. One person can belong to many different clusters. They did this before the web, of course, via things like clubs and special interest magazines, but the process now is much faster, supports narrower verticals, reaches worldwide (depending on your language), and enables a lot more peer to peer communication (instead of writing a letter to the editor, who may or may not publish it a month later, you post a comment to everyone else). So online society looks a bit like a clump.
In that clumped-up online world, people can choose how much influence they’ll put into each group. In some places you may be a listener and in others you may be a talker. Post more comments and your voice will be heard more loudly. Obviously there are still social limits on how much you can influence (if you express yourself poorly or you annoy people, they will tune you out), but the process is much more open than it was in the past, and each individual has control over how much they contribute. They don’t have to go through an editor or other gateway to get their voice heard.
The people who contribute the most to a discussion start to work like lighthouses, radiating ideas to others in the community. Those online voices have a big impact on purchases of products. When we surveyed people about their product selection process, they said online interviews and comments from real users are second only to a personal recommendation from a friend as a purchase influencer.
As Watts points out in his papers, it’s risky to trust what people tell you in a survey after the fact. Maybe they say something influenced them, but it really didn’t. But there’s also a lot of other evidence that topical review sites like Yelp (link) and TripAdvisor (link) have a real impact on business. A study by researchers at Texas A&M detailed TripAdvisor’s influence (link) and a survey by sponsored by Comscore reported that good online reviews affect both the likelihood to purchase a product, and the price people are willing to pay for it (link).
The other thing that the web changes about those lighthouses is how brightly they can shine. A member of a social club 40 years ago might have been able to influence 20 or 30 people once a month, when the club met. Someone posting comments or reviews online can reach thousands of people with a single post, and they can do it over and over again. In our surveys, the most frequent contributors to discussion sites posted about 300 times more comments and reviews than the average user. The combination of broader reach and higher frequency enormously multiplies the influence of the people who are very active online.
They’re influential, but they’re not Influencers
When we first looked at these numbers, in the web user research we reported on several weeks ago (link), we thought our findings confirmed the Influencer thesis. But as we looked at the numbers more deeply, we realized that in fact the most frequent contributors (MFCs) are very different from traditional Influencers. In the traditional view, Influencers are people with high social status, the pillars of the community. The Influentials describes them as middle-aged, educated, upper-middle class, married, conservative, computer-using professionals.
Golly, we’ve rediscovered Ward Cleaver.
Several online studies, including our own at Rubicon, have painted a very different picture of the MFCs. The Pew Internet folks found that they tend to be early adopters of the Internet who have graduated from reading content to creating it (link). A survey by Universal McCann characterized them as risk-taking, gregarious novelty-seekers (link). And in our own research, we found that they are:
— Young (half are under age 22)
— Technically sophisticated
— 60% male
— 40% students
— 31% nonwhite (more than the US population as a whole)
— 2:1 Democrats vs. Republicans
In other words, they are not the average person, and they are not traditional Influencers. They are a different type of people, with their own dynamics.
For marketers, they are a new frontier.
What to do about it
Here’s how the influencing process appears to work online:
— Many web users rely heavily on online reviews and comments by users when making purchase decisions.
— A small subset of users write most of those online reviews and comments.
— Therefore, those most frequent contributors have a disproportionate effect on product purchases.
If you’re in marketing, what do you do about this?
The first thing you need to do is pause and think about your company’s particular situation. The biggest flaw of both the swarm and influencer hypotheses is that they try to reduce a very complex influencing process to a few simple aphorisms. The real world has a lot more grays in it. The role and power of online reviews appears to vary from market to market, and in most cases the right approach for you will be a mix of traditional marketing and web activity. You need to tailor that mix to your particular situation.
Don’t let the absence of a simple answer depress you — it means companies will have an ongoing need for smart marketing people who can work the full range of marketing tools both online and offline. If the answer were simple, they could replace you with a chimp.
Here are some of the questions you should ask on order to assess the role of online reviews and the MFCs in your market:
1. What’s the pattern of influence in your market? Studies consistently show that the impact of online reviews varies widely from market to market. For activities like finding hotels, buying consumer electronics, and picking restaurants, users say online reviews are very influential. For tasks like picking a doctor or choosing who to vote for, users say online reviews have much less influence. The impact of online reviews can vary even within a category. For example, TripAdvisor users said they are likely to use online reviews to pick a hotel, but not necessarily to decide which attractions to visit once they get to a destination.
Several factors seem to drive the importance of online reviews:
Is the product hard for a customer to sample in advance? If it is, online reviews become more important. So they’re very important for restaurants and hotels, but less important for music (which people can sample directly by listening to the radio, online previews, or friends’ music players).
Is the product an impulse buy? This is a function of both price and durability. If you buy a bad carton of orange juice, you’ll just throw it out and buy a different brand. If you buy a bad stereo system, you’ll have to either return it or live with your bad decision for years. Guess which purchase is more likely to be researched online.
Is there a good place to go for reviews? In categories where there are well-established review sites pointed to by search engines, online reviews are more likely to be used. To see how this operates, look up the name of a restaurant on a search engine and see how prominent that restaurant’s Yelp listing is. On the other hand, in a category where there isn’t a well-defined forum for reviews, customers can’t find them even if they want to. This situation is true for many services categories, such as medicine and auto repair.
The influence of online reviews on product purchase:
If there isn’t a good source for reviews in your product category, you might want to think about sponsoring one. A vendor-owned review site rarely has credibility with users (do not waste your time trying to set one up), but independent community sites are often hungry for sponsorship, and a budget that feels small to a company may have a huge impact on the viability of a community site. Being a sponsor doesn’t let you control the content on a site, but it gives you an opening to build relationships and create good will among the MFCs who will write most of the content.
2. Are you focused on the right sort of online communication? Much of the literature about online marketing focuses on how to manage bloggers. The evidence shows that for purchase decision-making, weblogs are often very low on the totem pole. Most web users do not read weblogs frequently (if at all), and when someone is doing a web search for a purchase, weblogs typically are not ranked highly in the results. The review sites are where the recommendations happen.
The role of weblogs seems to be more about forming general ideas about companies and product categories. We haven’t found a lot of good market research on this, but weblogs seem to work as demonstration gardens for ideas. An idea that looks good there can be picked up and magnified by other media. New communication tools like Twitter have sped up this idea transfer process among web power users.
So courting the weblogs may be much more useful for creating a new category, advocating an idea, or heading off bad PR, than it is for driving demand and purchase recommendations.
3. Carefully court the MFCs. Once you understand the influence structure for your market, it’s time to start building relationships. Your goal isn’t to directly control what happens in online communities and weblogs; that is impossible, and heavy-handed efforts at control usually drive customers away rather than attracting them. Your goal is to make yourself a member of the online community, to build up good will so that people will treat you fairly, and give you the benefit of the doubt when you make a mistake. You can’t control what people say about you, but you can do a lot of influence the emotion behind those comments.
If online reviews are important in your market, start by getting to know your MFCs. Although the average MFC is young, male, and technically sophisticated, there’s evidence that the MFC profile varies dramatically from category to category. For example, the Texas A&M study on TripAdvisor found that heavy users of its reviews tend to be women aged 35 to 64. The people who write reviews on Trip Advisor are a little younger and less skewed toward females, but still they are quite different from the average MFC profile across the web.
Once you know who your MFCs are, it’s important to monitor what they’re doing, and reach out to them when possible. The reaching out part can be tricky, because most reviewers value the credibility they get by being independent from any vendor. Review-writers tend to view themselves as volunteer community helpers who assist others and reward companies that have good products. They don’t review for money, so it’s not wise to try to directly pay them. But it is possible to contact them asking for advice, and to politely request a new review when they’ve had a bad experience with you in the past.
Some sites even allow vendors to post direct responses to negative reviews. If you do this, be very careful to stay factual and humble; you’re a guest in their community, and they will judge you by both what you say and how you say it. Often it’s better to just apologize than to try to discredit an angry reviewer. It’s very hard for people to stay angry at someone who answers emotion with courtesy.
4. Work with the community’s personality and structure. In addition to understanding the MFCs themselves, it’s also important to understand the overall behavior of the community where reviews are being posted. The stronger an online community, the more likely it is to have developed its own personality and code of behavior for contributors. For example, Wikipedia has a formal, almost legalistic set of guidelines for writing and reviewing articles. Yelp doesn’t have as many written rules, but its most active reviewers tend to be young and very focused on hip new restaurants and interesting clubs. In contrast, Trip Advisor reviewers tend to give high ratings to places that have good service and friendly proprietors, even if the facilities are a bit run down.
You should adapt your community outreach to the personality of a review community. For example, some restaurants try to court the Yelp community by offering party nights for Yelp members. That has a chance of working because Yelp’s MFCs are young and socially active. But the same tactic would be unlikely to work on TripAdvisor, which has less of a social aspect, and whose members are much less concentrated in any particular region. Special offers for community members, and sponsoring the sites, can be ways to show that you care without giving people the feeling that you’re trying to control the online discussion.
What not to do with MFCs: Product design. MFCs are not normal users. The demographics and behavior of people who write a lot of reviews and comments are often radically different from the people who read and rely on those reviews. So although it’s important to take good care of the MFCs, designing products with them in mind is very risky, because they are not a good proxy for the typical customer. You may get great reviews and a quick burst of sales, but if the users aren’t satisfied your sales will eventually drop off. We’ve seen this happen frequently with technology products; a high tech device that pleases the average online reviewer is often too complex and expensive for the average user. When designing a product, it’s much better to use traditional market research methods to get the reactions of average users, rather than just asking for feedback online.
This also means it’s important to explain to online communities exactly who you’re targeting with a new product or service. If you tell them it’s designed for a user who’s different from them, they’ll usually take that into account in their reviews. This is one more reason why it’s important to have a good ongoing dialog with the community sites in your category.
(Note: This post was co-authored with my team at Rubicon and represents all of our points of view; we all of us had a ton of diverse experiences in online community and what it means to companies.)