Data Shows Articles with Digits May be Shared More on Facebook Than Those Without

More new Facebook data, continuing this series.

The next Facebook sharing data point I analyzed is the presence of numbers (in digit form, 1 through 9) in titles. In a wide range of marketing arenas digits have been shown to perform very well. They tend to help conversion rates in the form of prices and on social news sites like Digg “Top 10” style posts have always done well.

The difference isn’t huge but according to my data, articles with digits in their titles tend to be shared more on Facebook than stories without digits. I found that most articles in my data set didn’t use numbers in their titles, and you can see the scale of difference in volumes in the gray bars at the bottom of the chart.

For details on my methodology start with this post, then read

Data Shows: Articles Published on the Weekend are Shared on Facebook More

When I started posting my new series of Facebook data points, one of the most requested graphs was the days of the week (and times of day, which is coming soon) that are best to publish on to get lots of Facebook shares. What I found when I looked at days of the week is at first a little unexpected, but upon further thought fairly logical.

While I found less articles posted on the weekends (notice the gray bars at the bottom of the graph which indicate volume of URLs analyzed for each day), those stories that were published on the weekends tended to be shared on Facebook more, on average, than stories that were published during the week. The reasons for this probably include the fact that more than half of companies in the US block Facebook, so people can only use the social network at home, on the weekends. Additionally, the mainstream Facebook audience does not use Facebook for work.

The takeaway? If you want your article to be shared on Facebook by your readers, try posting it over the weekend.

For information on my methodology, start with this page. For this data point I’m using over 5000 stories and “average” is the interquartile mean which is less sensitive to outliers. The 0% line indicates the average number of “shares” stories from each site in my study get, when the line is above 0% it means that stories on that day are shared more than the average, and when it is below, they’re shared less. If you’re curious why it appears most of the stories in the data set are above average, this is because of the difference in the volume of published stories on various days.

Read More

Data Shows that Facebook is Better for Video Marketing Than Twitter

Continuing my series of Facebook sharing data (if you’re curious about my methodology, read the first post), I looked at articles that had the word “video” in their titles.

It turns out that those stories that indicated they contained videos were shared more than the average story on Facebook, while they were actually shared less than the average story on Twitter. This is likely because the Facebook platform makes it easy to embed multimedia content into updates while Twitter does not.

The takeaway here? Facebook may be a better platform for your videos to go viral than Twitter.

And again, if you have any other datapoints you’d like to see, please let me know. I’m really excited about my new Facebook analysis capabilities and I’ve got a ton more stuff planned for you.

Read More

Data Shows: “Twitter”-Centric Stories are Not Heavily Shared on Facebook

A couple of weeks ago, I started collecting a new dataset and I’m really excited about it because it’s the first time I’m collecting data from the mother-of-all social media sites: Facebook.

I’ve begun by capturing links posted to social media sites from 10 extremely popular news outlets. Some of the top blogs, both mainstream and geeky, as well as a handful of the most web-enabled newspapers of record. Then I’m counting the number of times those links are shared on Facebook (in three different ways) and on Twitter (through good old ReTweets). I then find the average number of “shares” for links posted to each site and compare the individual stories to the average in percent form and then combine those numbers to get a percent “effect” as a positive or negative number away from the average.

At this point I’ve got well over a thousand links and counting with full information stored. I’m also getting better at retrieving the data I want faster and more reliably.

I’ve already got a bunch awesome of things to show you, so keep your eyes out for more, but first lets talk about “meta mentions.” A meta mention is when someone on a given site, say Facebook talks about Facebook, or when someone Tweets about Twitter. Typically with ReTweet data I’ve seen that talking about Twitter gets you a lot of ReTweets, and this is to be expected since most people on Twitter are into talking about Twitter. Of course with older technologies like email, people aren’t really “into” email so much as they just use it to get stuff done.

So far my data shows that while articles that use the word “Facebook” in their title get shared more often than the average story on both Facebook and Twitter, stories that mention “Twitter” actually get shared less on Facebook. My assumption here is that Facebook is less of the early adopter crowd that wants to sit around all day and talk about Twitter, while Twitter users are more likely to be social media geeks.

The key takeaway is to know your audience. If you want to go viral on Facebook, don’t talk about Twitter.

And since I’m just starting to get into Facebook data like this, what kind of stuff would you guys like to see?

Read More

Are you a Social Media Snake Oil Salesman or are you a Scientist?

If you’ve read about social media or been to social media conferences, you’ve probably heard tons of advice like “love your customers,” “engage in the conversation,” “be yourself” and “make friends.”

I like to call this kind of stuff “unicorns and rainbows.” Sure, it sounds good and makes you feel all warm and fuzzy, but it’s not actually based on anything other than “truthiness” and guesswork.

It’s the modern day equivalent of the witchdoctor or snake oil salesman. A couple of time-honored adages repeated ad nauseum, coupled with the unquestioning awe of an unaware audience, and pretty soon you’ve got an entire industry made of easy-to-agree with smoke and mirrors.

Problem is, bleeding with leeches and magical tonics don’t actually work. In fact, a lot of the time they do more harm than good.

And then along comes real science–real medicine and real data about what works and what doesn’t. Curing disease moved out of the dark ages and started making progress; now it’s time for social media to move past the unicorns and rainbows.

The great thing about the web is that nearly every interaction can be measured and observed in aggregates of tens and hundreds of millions. We can gather more qualitative and quantitative data about human behavior than at any other time in history. Yet the future of marketing, the very industry that is trying to push communications, business and public relations forward, is built on advice that comes from nothing more meaningful than soft-focus fantasies.

To the snake-oil salespeople, social media success isn’t something repeatable. It’s not the outcome of a process; it is black magic, guessing and praying.

Those of us who are a part of this social media thing now will be the forefathers of the next generation of marketing. We’re going to be the ones who decide how it plays out. Of course, there aren’t any formal degrees in this yet, and most of us don’t wear labs coats. But we need to decide if we are going to leave the future of social media to magical tonics, or if we are going to use science and data to discover what really works to motivate people.

To the scientists, social media success is something you can iterate on, plan for and learn from. Things that work can be analyzed to produce repeatable, dependable results.

The next time you see or read or hear someone giving superstitious, feel-good social media advice, question them. Ask what data it is based on, what science? Ask them to prove what they’re saying.

Most importantly ask yourself: are you a snake-oil salesman or are you a scientist?

And if you look around hard enough, I’m sure you’ll see that social media science is catching on. Its encouraging to see that there are at least 7 people on Twitter with “social media scientist” in their bios, lets blow that number up.

Read More

How to Use “Us vs Them” Stories to create Social Media Evangelists

Us vs them is one of the oldest, and most powerful marketing ideas. Apple is a quintessential example: from their beginnings they’ve portrayed themselves as the small guy against the big powerful bully. In 1983 it was IBM and more recently its been Microsoft. The company turns customers into evangelists who are more than happy to spread the word about the good fight, but how exactly does it work?

In a 1983 article titled “On Viral Sentences and Self-Replicating Structures” the author, Douglas R. Hofstadter recounts a letter he was sent in response to a previous peice. This letter describes self replicating (viral) sentences, beginning with the rudimentary:

It is your duty to convince others that this is true.

The letter writer notes the obvious: unless the listener believes the statement above, it won’t spread. (Tweet this and see what happens.) He then moves on to a more sophisticated structure where the above sentance occurs at the end of a set of beliefs:

If the listener accepts statements S1 through S99 they will act on S100.This is how many religions work, the belief system is the bait and attached to it is an evangelism hook.

The letter then explores a more subtle variation based on a simple structure:

The villain is wronging the victim.

If the listener believes this statement, and believes that the victim deserves to be saved and if the villain is bigger or more powerful than them they will realize that the only way to effectively challenge the villain is to recruit more people to help. The evangelism hook is implicit, subtle and powerful.

When I looked at urban legends I found a similar phenomenon that occurs with striking regularity online called the Goliath effect. Simply put, people love to communicate about abuses of power against the underdog. Microsoft and the RIAA are favorite Goliaths of the web.

If we want to design a viral idea based on this structure, we have 3 blanks to fill: “villain,” “victim,” and “wronging.” For example let’s look at the title of one of my most popular blog posts ever:

Twitter plans to Mangle ReTweets.”

The easiest way to make someone believe that he victim is worth saving is to make them identify with the victim, in this case anyone who ReTweets. ReTweeters were the target of this sentence, and given their contagious behavior, they make a wonderful audience.

The villain of this sentence is, of course, Twitter. They may not be a huge company, but they’re larger and more powerful in this area than any individual user. The only way someone could hope to #saveretweets would be to recruit all of their followers in the fight.

In the post I spent considerable time asserting and proving the “wronging” part. I explained why the proposed changes were going bad and needed to be stopped or at least challenged, this is the “bait” part in the figure above.

Read More

The Five Elements of Viral Calls to Action

Most marketers know that to get someone to do what you want, you have to ask them to do it, you have to have a call-to-action (CTA) that persuades them to buy your product. With social media marketing, the action we’re aiming for is to get our readers to share our content with their friends and networks, so our CTAs must entice them to do just that. Here’s the 5 most important concepts to think about when you’re constructing your viral calls to action.

Subtly

People like to think that everything they do comes from some logical, un-manipulateable part of their own brain. Doing what you’re told doesn’t feel as good as doing what you want to do, and nobody really wants to believe that what they “want to do” can be easily directed. This is especially true when it comes to sharing ideas and content with friends. Who would share something just because they’ve been told to by some marketer they don’t know?

The point here is that if you want to persuade your readers to spread your content, its not a great idea to whack them over the head with painfully obvious commands (although it will work in special instances, like when you’ve giving something away). You shouldn’t tell them what to do, you should make them want to do it in such a way that it feels like the idea was their own.

Motivation

There are a number of reasons why people spread ideas and content, and when you’re constructing your viral call to action, you should be leveraging one or more of these specifically.

If you get a group of marketers in a room together and ask them how to “make something go viral” one of the first things someone suggests is to give something away for free. And it works. People love free stuff and they’re often willing to do something for a chance to win. A pretty easy, if unimaginative and elementary way to get people to spread an idea or piece of content is to offer them a prize for doing so. This is perhaps the most surefire way to “go viral” and if you’re in a hurry to come up with a contagious idea, this is often your best bet. You’re essentially paying them to spread your content for you. The only concern with this tactic is that high-reach individuals will probably see right through it, and you’re unlikely to bribe someone with a size able audience this way.

When you directly ask people why they share things with their friends, the most common response is “relevance.” Things like “it seemed right up my friend’s alley” or “it made me think of so-and-so.” One fairly obvious CTA that exploits this motivation would be “send this link to your friends who’d be interested in it” or something to that effect. Its simple and obvious, but it might just work to a point.

My favorite, way to encourage this response, is the more subtle “combined relevance” technique that allows you to create content that seems personalized for a lot of people. Of course if you sent every one of your friends links to every piece of content that was relevant to them, that would be all you did. You’re going to need to hit on more powerful motivators than just relevance.

A 1983 article titled “On Viral Sentences and Self-Replicating Structures” the author, Douglas R. Hofstadter recounts a letter he was sent in response to a previous piece. This letter describes viral sentences beginning with the rudimentary: “It is your duty to convince others that this is true.”

The letter then explores a more subtle variation of that technique based on a simple structure: “The villain is wronging the victim.” If the listener believes this statement, and believes that the victim deserves to be saved and if the villain is bigger or more powerful than them they will realize that the only way to effectively challenge the villain is to recruit more people to help. Thus the motivation is powerful and the call-to-action itself is subtle.

For example let’s look at the title of one of my most popular blog posts ever:

Twitter plans to Mangle ReTweets.

The easiest way to make someone believe that he victim is worth saving is to make them identify with the victim, in this case anyone who ReTweets. ReTweeters were the target of this sentence, and given their contagious behavior, they make a wonderful audience. The villain of this sentence is, of course, Twitter. They may not be a huge company, but they’re larger and more powerful in this area than any individual user. The only way someone could hope to #saveretweets would be to recruit all of their followers in the fight.

I’ve written and talked about a bunch of different motivations and rather than detail each one of them here, I’ll list some of the best and link to more indepth studies of them.

Friction Reduction

As with any good call to action, you should have a very clear idea of the specific action you want your readers to take. Is it to share the blog post on Facebook? Is it to ReTweet you? Is it to email your YouTube video to all their friends? The most effective calls will likewise have a very specific aim.

Since you know precisely what you’re trying to persuade people to do, you should also be ale to understand how much time and effort it will take for them to do it. This time and effort is friction and the more friction presented to your potential spreaders, the less likely they’re going to be to share your content.

You can overcome a high friction action with a lot of motivation. If you want people to compose customized 500 word emails to 20 of their best friends with your link at the bottom, you’d better be offering them a very good reason for wanting to do so.

The other side of this problem is to reduce the amount of friction present. Asking people to click a single link to ReTweet or Digg something is asking very little of them. Creating a one-click sharing action is the holy grail of frictionless viral calls to action. But if you want them to do something that requires a bit more effort, like sending lots of emails, consider offering them some cut-and-paste text to use.

Timing & Placement

Consider for a moment, the tiresome marketing/dating analogy: if you go out on a first date, you wouldn’t propose marriage, so you shouldn’t ask someone to do something for you before you’ve provided them any value or built a relationship with them.

When we’re talking about viral calls to action there are two issues at play. One is that you want your calls above the fold in as prominent a location as possible, and the other is that you want to present your readers with the call to action exactly when they’re most likely to want to use it and share your content (which is typically after they’ve read it). The easiest solution is to put your buttons, links, whatever form your CTAs take in both places, that is at the beginning of your content and at the end.

In the case of a button, like the ReTweet or Digg buttons, the mere site of them becomes a form of social proof motivation. If a visitor sees that hundreds of people have already liked this content, they’re much more likely to percieve it as more valuable as well. With these types of calls to action you should place them where a visitor will see them upon first viewing your content.

There is also much to be said about the colors and look and feel of the CTA, especially if it is a button. I’ve written about that in more depth elsewhere, so I won’t go into it in this post.

Self-Replication

In many forms of social media, the element of a piece of content that is actually being shared by users is the title. On Twitter, ReTweets of an article typically contain its title, the same with social news sites, social networks and often email and IM. This means that when an individual spreads your content for you, they’re likely to be using the headline as if it were their own words, anything you say in the title borrows the authority of the sharer.

When we construct content we wish to spread, if we can bake in the viral call to action in such a way that it is contained in the title itself, we can make the act of sharing it self-replicating. A simple example is the blog post with the title “ReTweet this to win.” Again, while giveaways and less subtle calls-to-action like this can be effective there are drawbacks. A more sophisticated example would be the “the villian is wronging the victim” model from above.

Read More

Data Shows that Negative Remarks Lead to Fewer Followers

Continuing my series of TweetPsych based data points, this is based on analysis of over 100,000 accounts and looks at the “Negative Remarks” category. Negative remarks include things like sadness, aggression, negative emotions and feelings, and morbid comments.

As it turns out, nobody likes to follow a Debbie Downer accounts with lots of followers don’t tend to make many negative remarks. If you want more followers, cheer up!

Read More

Watch The Science of Social Media Marketing Webinar on YouTube

Update: Check out the newest version of the Science of Social Media.

A combination of much of the research I’ve been doing for the past 3 years, my “Science of Social Media Marketing” presentation is one of my favorite to give. It combines statistics, marketing, history, math, social psychology, memetics, epidemiology, steampunk, zombies and absinthe. I talk apply lessons from sources including urban legends, rumors, homeric poems and proverbs.

The first time I did it, was for a record-breaking 12,000 plus registrant HubSpot webinar back in December of ’09 and the most recent time was for O’Reilly Media. O’Reilly was able to get the whole, one-hour talk, (my slides and my narration) up on YouTube, so if you haven’t been able to attend a live version of it, you can now watch the recording.

Please do, and let me know what you think.

Read More

Data Shows That Self-Reference Does Not Get Followers

If you like this post, or any of my work, please, nominate me for a Shorty Award.

Following up on my last post using TweetPsych Data, I looked at a metric opposing social behavior: self-reference. This time the dataset is well over 60,000 Twitter accounts.

What I found here is pretty clear, accounts that have more followers do not tend to talk about themselves much. Want more followers? Stop talking about yourself.

If you like this post, or any of my work, please, nominate me for a Shorty Award.

Read More