From influentials to small connected groups

Much marketing activity in the last ten years has been focused on finding and seeding messages with “influentials” – people who have a high number of connections, and are thought capable of setting off cascades of a message. Based on a lot of research since The Tipping Point was published over ten years ago, we’re now seeing that this idea was largely a myth. We’re learning that most people aren’t influenced by people with these attributes. We’re learning that many of our decisions are made subconsciously (even when we think we made a conscious decision) and that the people who do have influence over our behavior are usually the people who are emotionally closest to us. We’re learning that because of over-confidence in their knowledge, many experts are worse at making predictions than random chance.

I feel that we’re at the beginning of a cycle in business where we move away from this idea of “influentials”, and instead focus marketing activity on small connected groups of close friends. I think this is what marketers are starting to think about, and will be the prominent theme for this decade. Let me briefly explain three driving factors behind this shift.

  • The first driving factor is that our online world is catching up with our offline world. Just as we are surrounded by people throughout our daily life, the web is being rebuilt around people. There is overwhelming evidence of this change all around us. Many businesses are reorienting themselves around people. As a result, people are increasingly using the web to get the information they need from each other, rather than from businesses directly. People always sourced information from each other offline, but up until now, online information retrieval tending to be from a business to a person.
  • The second driving factor is an acknowledgment in our business models of the fact that people live in networks. For many years we have often considered people as isolated independent actors. Most of our consumer behavior models are structured this way. People moving down a decision funnel, making objective choices along the way. Based on lots of recent research in the fields of psychology and neuroscience, we now know that this isn’t true. Peoples’ networks influence almost every aspect of their lives. What they do. Where they go. What brands they prefer. What products they buy.
  • The third driving factor is that for the first time in humanity, we can accurately map and measure human to human interaction. We now have multiple networks that digitally connect millions of people, and support interaction between these people. We can measure who is connected to who, who talks to who, and who shares ideas with who. This allows us to understand how messages spread.

When we look at these networks, very clear patterns emerge. People have multiple independent groups of friends, they tend to have less than five groups, and these groups usually have less than 10 members.

When we look at individual relationships between people, we also see clear patterns. When it comes to trusting others, people tend to have a trusted “inner circle” of less than five people. People turn to their closest friends for advice, even when they know that their friend has no knowledge of the topic.

We see that people have a limit on the number of people they can keep up with. This is a physiological limit of the brain and isn’t going to change any time soon. We can only keep 150 other people in our heads at any one time. We may have many more “friends” on Facebook than 150, but we have no idea what is going on in their lives.

The most interesting thing about the patterns found in social networks, is that although people move in small circles and trust a handful of people, they are all linked together. I may only have four groups of friends, with five people in each group – 20 people – but all of those people also have different groups of friends. People I don’t know. And their friends have different groups of friends. Through our friends of friends of friends, we’re connected to tens of thousands of people. All of us. All closely connected to tens of thousands of people.

We have ~4 independent groups of friends (left). Focusing in on our ten closest friends, the people we trust the most (right)…

Removing the detail (left) and spreading people out (right) shows you one person and their ten closest friends…

This friend of theirs (left) has four independent groups of friends. This other friend also has four independent groups of friends (right).

These two friends (left) also have four groups. In fact, all ten closest friends have four groups (right). This is hundreds of trusted connections. This is the opportunity. Many people, at an exponential scale, connected by their closest, trusted friends.

The most critical point to understand about this shift, is that the people connecting groups are not special. They are just you and I. Each one of you is likely a bridge between at least four groups.  Knowing that groups are independent, each one of you is the only person on earth connecting about 30-40 other people together. If we want ideas to spread, if we want people to evangelize our brand and for their messages to spread, we need to focus on everyday people, and understand how their groups of friends are connected.

This is where marketers will start to focus their attention in the next year or two. This is our opportunity. Many connected groups of friends.

- Find out more in this slidedeck.
- I’m writing a book about this. It’s called ‘Grouped’. Check it out here.
- Here is an incomplete set of references. I’ll post a new set over the next few weeks. In the meantime, start by reading this book and this book.


  1. Relevant: ‘The wisdom of crowd effect is valuable for society. While the wisdom of crowd effect is a statistical phenomenon, social influence can have an impact on individual decision-making…’

  2. with a little more data this could have been the post of the year. (ops i think my previous comment was spammed by mistake :(

  3. Insightful as ever.

    I was thinking of this shift toward marketing to groups when noticing how many beer commercials have changed towards making a beer “the drink for the guys” as opposed to the drink for him or her. It’s no longer about convincing one person to buy the product, it’s about creating the group demand.

    Guinness used to be the perfect “man at a bar savouring every drop” advert, but now adverts show a tray full of 6 or 7 pints, headed towards a gang of friends.

    It’s hard to do this with products that are sold to “unique individuals”, even Apple have shifted. Where once we heard “the misfits, the rebels, the round pegs in square holes” we now see ads like Facetime where it’s all about staying connected with your group.

    Suffice to say the book is pre-ordered.

  4. I think there may be a need for developing measures of ‘closeness’ (social affinity score as a ‘measure of social tendency’) to differentiate ‘the public influentials’ from ‘influential friends’. Influential friends are reachable within 2-3 hops from your own social perimeter.

    “In the linear cascade model when a user in a social network adopts a new behavior there is a probability that each neighbor in the network will adopt it. [...] Likewise, we wouldn’t want to notify people who are very far apart on the social network because a user is more likely to adopt a behavior if more than one of his friends has also adopted it.”

    Also see: “Maximizing the Spread of Influence through a Social Network” by Kempe, et al. Link is in this article:

    ‘how influence propagates in a social network if it were represented as a directed graph’ and ‘modeling how information flows/diffuses in a directed graph of social entities as nodes’ (Kronecker multiplications). (See comment from “Sagar Mehta” in the url given above) It especially interests me how to track/trace the flow between any two given social entities in the graph when one of them is to be reached from the other. (in graph theory they call it Shortest path problem – The new startup is modeled on this aspect.

    When graph theory is applied to social network modeling, we can find that the shortest path may not be the most efficient. A longer path may have better affinity scores. Its like:You know and relate with person B more than person A.

  5. Thanks for the comments folks.

    Simo, I’ll check my spam filter, sorry about that.

    Des, you’re always great at citing examples from your own observations! The beer commercials trend is very interesting, not sure if it is coincidence or correlation.

  6. Mohan, I think your observations are spot on. We will definitely see measures of closeness develop over the next couple of years. As you point out, it’s not a simple linear model, in fact it’s incredibly complicated to figure out who trusts who in any given context.

  7. [...] From influentials to small connected groups [...]

  8. [...] che si concretizzano, sotto il profilo della relazione on line con le persone, nell’evoluzione dagli influencers a piccoli gruppi di persone connesse tra loro con interessi comuni e legami [...]

  9. [...] From influentials to small connected groups ,本文实际上是Paul对自己之前的社会化网络研究报告 How Your Customers’ Social Circles Influence,What They Buy, What They Do and Where They Go的概要,也非常值得一看,不过非常长,有精力的同学翻译一下吧。 [...]

  10. Paul – I couldn’t agree more. What’s interesting about these small groups is the amount of conversation that takes place within them, the amount of user time spent there.

    People have natural aversion to communication in large groups. Same with sorting through the noise. As soon as a group gets smaller and more private, people a) talk more, and b) listen more… The result is a lot better influence. If the viral loop between one group and the next is narrowed in time, you have a more powerful, viral spread, than a broadcast to a million people, because the level of influence is so high, and the multiplier can get you to that size with only a few degrees of separation…

    Let’s connect. I run LoKast – an instant social network for the physical world ( We are attacking this exact equation by focusing on physical groups – groups at events, nightclubs, classrooms, living rooms, offices, etc. etc.


  11. [...] From Paul Adams, Facebook/ex-Google: I feel that we’re at the beginning of a cycle in business where we move away from this idea of “influentials”, and instead focus marketing activity on small connected groups of close friends. I think this is what marketers are starting to think about, and will be the prominent theme for this decade. [...]

  12. [...] hat Paul Adams einen sehr guten Blogbeitrag über Social Circles verfasst. Im November wird sein Buch „Grouped: How small groups of friends are the key to [...]

  13. Hi Paul,

    Very insightful post!

    There is something I want to comment on and what you have been talking about in former presentations as well and that is the degree of spreading. You state that one person with 10 strong ties could in three steps influence 10thousand other persons because each person could influence 10 other persons (10 * 10 * 10 * 10 = 10k). Although I know you are telling this conceptually, I think you left something out that is quite important and makes this statement untrue. Between those that are influenceable could be quite an overlap. Especially because they are sensitive to be influenced means that they in certain ways have the same interest and thus most likely could be in the same circle. The overlap of persons between two different circles but connected by one person is quite significant. Let’s say that Peter is telling his best friend Mike at the tennisclub to buy a Playstation because he is interested in games, and in another circle (his work) he tells David to buy a Playstation because he is also interested in games. Now I don’t want to say that Mike and David are in the same circle and will be influencing eachother, but they could be in the third or fourth degree; it is so likely because these people will be hanging around this ‘game/Playstation’ social circle. In later stages, when someone recommends a Playstation, this person might say ‘yeh I know it’s great, I already heard’, which significantly diminishes the amount of influenceable people. Any thoughts on that?

  14. [...] out some references and other resources he’s found useful in his writing.  One post, “Small Connected Groups” made me think about IdealXChange, how we are organized, and how we can better organize, [...]

  15. [...] that was liberally dosed with the company’s “Social by design” mantra. He spent some time examining the social ties that bind us, and what that meant for online applications. “Social” was no longer a bolt-on, he argued, but [...]

  16. [...] dosed with the company’s “Social by design” mantra. He spent some time examining the social ties that bind us, and what that meant for online applications. “Social” was no longer a bolt-on, he [...]

  17. [...] outsized value. There’s been a lively debate around this (see, for example, Paul Adams excellent discussion and his comprehensive collection of relevant research links). It seems to me that most of the [...]

  18. [...] poco profesional, como el análisis de influyentes. En este sentido, será interesante leer este artículo sobre influencia y pequeños [...]