What’s really unique about Dan is that he makes science practical for the social media world. He takes a deep look at what’s going on with social media activity and he extracts knowledge or nuggets that are really amazing and very helpful for marketers.
Dan has contributed to major websites like Mashable, Copyblogger and ProBlogger. He has also developed a number of tools for Twitter, including TweetBacks, which is a great tool that actually helps bloggers display tweets related to a blog post on their website. He also works at HubSpot as an inbound marketing manager.
Mike: Dan, tell us a little bit about your background. How did you become so interested in things like Twitter?
Dan: By trade, I’m a programmer. My first real job was as a webmaster and I built a bunch of tools for the company I was working for at the time.
Then I moved into search engine optimization (SEO). My interest in social media in general started as an offshoot of SEO, generating content. Writing content with the keywords in it is pretty easy. Anybody can do that.
The hard part of SEO, of course, is getting links to your stuff. You can buy links, you can exchange links and you can do shady things like that, but eventually Google is going to catch you. It takes a lot of time and a lot of investment and it’s not really worth it anyway.
The only acceptable way to get links is by writing link-worthy content and using social media to promote it and getting other people to link to you. So I started playing with Digg and working on getting things on the front page of Digg because you can get a lot of links that way.
Then Twitter kind of blew up and I started getting into that initially, to be completely honest, just as a way to help my stories do better on Digg. But then once I got into it and realized the power behind it, especially with retweeting, which is one of my absolute favorite phenomena ever, I started focusing on Twitter.
What I think is so unique about Twitter, again, specifically retweets, is that people have been sharing content or sharing ideas for thousands of years. In fact, most of our culture is based on that concept. The chair you’re probably sitting in right now wouldn’t exist if somebody didn’t have the idea to build a chair and then tell somebody else about it, and then the idea caught on.
So retweeting specifically, I think, is a godsend if you’re interested in figuring out what types of content are contagious.
Mike: I want to get into retweets a lot, actually. But first of all, tell me what a social and viral marketing scientist is. What does that mean?
Dan: The best way for me to explain is if you think about the people who give a lot of social media advice or many of the gurus ‘in the space,’ a lot of them are doing what comedian Stephen Colbert refers to as “truthiness.” It’s advice that sounds good or sounds warming inside. I call it soft-focus sort of rainbows and unicorns stuff. It’s stuff that sounds right and it makes you feel good inside, like, “Be transparent and be honest and be nice and be helpful.” Those are good ideas. There’s nothing wrong with those.
But what I’m interested in, and being a programmer at heart, is learning more about what actually causes things to go viral or what actually causes people to take certain actions. Rather than just, “This sounds right,” I want to know what’s actually correct and what’s actually true.
Mike: You just want the data to prove that these things are accurate.
Dan: Yes, the hard data just to prove or disprove in some cases what actually happens and what’s actually good advice and what are really good best practices.
Mike: Where do you get your data as a scientist?
Dan: A lot of different places. The Twitter API is amazing. I have several systems built that capture every retweet that happens. Or I can write scripts very quickly that will draw whatever type of data I’m looking for from Twitter.
Then, of course, there are all the other APIs of the social web applications. Digg has an API that’s wonderful to use. Also, at HubSpot we have developed a number of grader tools. Twitter Grader is the one that’s probably the most well-known. Now that I work here, that allows me to dig into a database of over 5 million accounts and analyze all sorts of fun data there too.
Mike: For folks who don’t know what Twitter Grader is, it’s a great little tool. You can put in your Twitter ID and it will rank you up against other people in the Twitter universe by giving you a score from 0 to 100. I think it’s really very cool. It also helps set you apart by physical location.
How do people find Twitter Grader, Dan?
I actually didn’t build the tool. It was built before I got here.
I think one of the useful things about any kind of measuring or grading tool is that people love to compare themselves to other people. There’s a Twitter Grader badge and you can tweet about it. Of course, when you get a grade and it’s good or bad, you want to tell people about it. You want to say, “Look, this is what I am.” It definitely spreads word of mouth. There has been a bunch of blogger coverage [of Twitter Grader], and it definitely spread because people wanted other people to know how they stacked up.
Mike: Let’s talk a little bit about retweeting because I know this is something you’re very passionate about.
I know you’ve done a lot of research. First of all, talk to us about the findings. What makes folks retweet? What are the main characteristics of messages that tend to go viral?
Dan: One of the first things is that a lot of tweets—I think it’s something like 20%; I have the exact numbers on my site—have links in them. That’s just random all tweets, not just retweets. Retweets have more of a tendency to have links in them. So, clearly, if a tweet has a link in it, it’s much more likely to get retweeted.
Another one that is kind of a no-brainer, but is really, really powerful is if you ask people to retweet a tweet for you. If you say, “Yada, yada, yada,” with a link and then you say, “Please retweet,” you’ll get a lot more tweets than if you didn’t.
Mike: What’s the psychology behind that?
Dan: I think there are two things. There’s a call to action there. With any sort of marketing, you always want to tell people exactly what to do and what action you want them to take. I think that’s probably the thing that’s most powerful.
The other thing is that people who are going to see something that says, “Please retweet,” are following you, which means they like you for whatever reason. So if you’re asking for a favor, it’s not really a big deal. Most Twitter clients have a one-click kind of retweet thing and so they’ll do that for you.
Mike: So far, we’ve talked about how retweets often have links and often have a request to retweet it. What are some of the other characteristics?
Dan: More recently, I’ve done research into the linguistics of retweets. Actually, some of my findings here were sort of surprising.
I assumed that with most kinds of viral content outside of Twitter, the message had to be simple and people had to get it in order to retweet it. So I ran the Flesch-Kincaid, which is a standard test of readability, against random tweets and retweets, and I found that retweets, surprisingly, have a slightly higher grade level required to read them, which means they’re more intelligent or they’re more complex.
Mike: What does that mean for a marketer, just out of curiosity? How does that translate?
Dan: It means you shouldn’t be afraid to not dumb down your content.
Another thing I found, and I think this might have to do with that readability, is I looked at word uniqueness. I took every word in my database of millions of tweets and millions of retweets and I looked at how many other times those words occurred in that data set.
It turns out that retweets, the words used in them, tend to be a lot more unique and novel.
Mike: Do you have any examples off the top of your head?
Dan: Novelty is huge in all type of viral marketing. That’s one thing that’s pretty important.
Another thing I looked at is the actual breakdown of parts of speech in tweets versus retweets. Retweets tend to be very noun-heavy and speak in the third person.
What I think that has to do with is one of the most retweetable phrases is actually “new blog post.” What that comes from is in WordPress and a bunch of other publishing platforms, there are features where you can automatically publish or automatically tweet the newest content. So if you’re doing a WordPress blog and you write the blog about whatever and then the system automatically tweets it for you, typically it will start the tweet with the prefix “new blog post.” Then that gets retweeted a ton.
What that means is that a lot of retweets are actually headlines. Headlines, if you think of some newspaper headlines, tend to be very noun-heavy.
Mike: So “new” is obviously one of those keywords there. “Blog post” implies some valuable content that they’re going to be clicking on. And the link, of course, is going to be present because that’s going to be pointed to the blog post and then maybe asking for the retweet.
Do you also think some of the tools that are now on blogs like TweetMeme are going to help increase the retweets?
Dan: Absolutely. To compare this to some other data that I’ve looked at, when a tweet gets tweeted, the more it’s retweeted, the more likely it is to be retweeted, which means that even if you’re not asking for a retweet by saying, “Please retweet,” merely having “RT” or “retweet” in something is sort of an implicit call to action. Just bringing up the idea of retweets and reminding people that that’s something they can do will lead to more retweets.
TweetMeme has a little button where you can see the number of times it has been tweeted. Then it says ‘Retweet.’ So just reminding people, even if it’s subtly or implicitly, that retweeting is something they can do is a great way to get them to do that.
Mike: Talk to me a little bit about the usage patterns of people on Twitter. I know you’ve done a little work on this. From a marketer’s perspective, what time of the day should we be tweeting our best content? When are people most likely to be viewing tweets? What other kind of time or usage patterns can you share with us?
Dan: I’ve graphed both the volume of tweeting and the velocity of tweets overall and then the volume of retweets. They both fall into very similar patterns.
Twitter is a worldwide tool, but it’s very, very prevalent in English-speaking countries, so clearly the U.S. leads in Twitter usage right now. Specifically, the coasts lead it.
East Coast time starting from say 11 am until 5 or 6 pm are the peak hours on the East Coast for tweeting and retweeting. It varies up and down in between that. At 11:00 am, I think people on the West Coast are starting to get into work or waking up.
My best tweets are the things that I want to get retweeted. I’ll do, later on in the day, maybe one or two because the West Coast is up at the same time and the East Coast hasn’t really gotten into being ready to go home or anything.
Remember that people use Twitter differently. I’m on Twitter 24 hours a day, but I’m a huge geek. Other people use Twitter when they get home or before they go to bed, or they’ll use it early in the morning or they’ll use it at work all day. If you have a bunch of followers, they may be using Twitter at completely different times. There’s the morning crew or the afternoon crew or the night crew.
So if I have some piece of content that I think is really, really good and I really want everybody to see it, I’ll retweet it a couple of times at different times because people’s tweet streams move so fast that it’s easy to miss things.
If you do it a few times, don’t go crazy with that because you will annoy people if you’re tweeting the same thing over and over again. But if you throw a few different things in between, I think that’s perfectly okay.
Mike: Talk to me about days of the week. Do you have any thoughts on that?
Dan: Yes, I think, like everything else on the web as far as business kinds of things go, Monday through Thursday are probably the best, with Tuesday and Wednesday being THE best.
Especially the stuff I’m tweeting about. I’m a marketer, but I also publish marketing content, so I’m marketing to marketers. Clearly, they’re all during business hours during the week.
It’s the same thing with email open rates and things. Monday, people have a lot of stuff because they’re catching up from the weekend. They’re at work, but that might not be the best time to catch them. And then on Friday, they’re thinking about the weekend and they’re just ready to go home. A lot of people even leave early on Fridays. So Tuesday and Wednesday and sometimes early in the day on Thursday are the best times.
Mike: What are some of the new projects that you’re working on that you can tell us about?
Dan: The book is the one that’s holding my life right now. I’m writing a very introductory sort of PowerPoint as a book for O’Reilly publishers called The Social Media Marketing Book.
The idea with that is if you’re new to the web but you’re an old-school marketer or if you’re new to marketing but you’ve been on the web for a little while, like on Facebook or whatever, this book will introduce you to the concepts of social media marketing.
It’s not really meant for the super-advanced users. It’s meant for people who want to know, “What can I do on Facebook for marketing?” or “What can I do on Twitter or on forums?” and that kind of stuff.
That should be out at the end of the year, hopefully before Thanksgiving. I think it will be really good.
Mike: Where can folks go if they want to learn some more about you, Dan?
The thing that I probably update most often is my Twitter account, so @DanZarrella is where people can get a hold of me and talk with me.
Mike: Excellent. Do you have any closing words of wisdom or any last thoughts on social media marketing or anything in general that you would want to share with marketers?
Dan: Yes, I think that with the web in general and with social media marketing in particular, we don’t have to just rely on what sounds right anymore.
If you are making a decision and it’s important to you, you should actually be looking at data. If somebody is giving you advice and it sounds right but you’re thinking, “I don’t know if I believe that,” ask them where the data is coming from. If they don’t have an answer, try to figure it out and get your own answers through the data. It’s easy to do now and everybody should be doing it.
Mike: Excellent advice. Dan, thank you very much for taking some time out of your day to talk to us.
Dan: Absolutely. Thank you very much.
Listen to the rest of this interview (below) and learn how to set up Twitter account descriptions, whether your personal name or business name is better, how to create automated direct messages, how to use social media analytics and more.[audio:DanZarrella-interview3.mp3]