The stampeding sound you may have recently heard was not caused by a bank panic or rumors of an impending astroid strike.
It was just the new Like button from Facebook.
The promise of this new offering is that visitors will click the buttons on your site, pump up your Facebook mojo with their friends and return fresh new traffic your way, all for free. What could be better than that?
Free link-love from the world’s largest social network was too good an offer for most to pass up, and no sooner had Facebook’s big presentation ended than site owners worldwide scrambled onto the bandwagon. The herd-on-the-move sound was their collective response to Facebook’s call, and small, blue Like buttons are now multiplying across the web faster than you can say “pandemic.”
But are they really appropriate for all sites? I decided to find out.
I began by testing my own integration as well as several different WordPress Like button plugins on my personal blog. When I had become confident that the buttons wouldn’t cause any problems, I felt it was safe enough for a limited rollout to a few of the blogs I manage.
I tested Like Button performance on two sites for a one-week period. Site #1 is a stock forecasting blog that gets about 20k-30k visitors each week. Site #2 is blog for fans of a popular singer and receives about 2,300 visitors a week or one-tenth the traffic of the other blog. Both are updated with fresh content daily.
So, two sites, with very different traffic levels and audiences, both running the new Like buttons. Which do you think saw the greatest number of button clicks and return traffic?
If you’re thinking as I did – site #1, by a landslide – you’re wrong. And it wasn’t even close. By week’s end, site #2 – the fan blog with light traffic – won the contest handily. This certainly raised my eyebrows. The number of Like votes it received easily exceeded the number received by the far busier site, which received none – none.
Piling humiliation atop insult, The Little Fan Blog That Could also received a fair amount of new visits back from Facebook. This was great news, because adding a Like button did, in this case, boost traffic measurably, tending to confirm that the Like buttons can be beneficial to more than just Facebook.
But why did the blog with the higher traffic mysteriously receive absolutely no love from Facebook users?
I believe there are two reasons: first, that visitor demographics play a strong role in how things will be shared, that some social bookmarking tools are more effective than others in certain settings; second, that the type of content dictates whether it will be shared at all.
To see if these explanations make sense, I needed a much larger set of data than was available the sites I tested. Since the important data is available to anyone who cares to look, we’ll examine stats for articles on a few heavily trafficked sites.
First, let’s focus on the visitor demographics question and whether it affects social media sharing preferences. I’ll use Facebook Like and Share button counts versus the number of retweets each page received to see if a trend emerges.
The graphic below shows the preference of visitors to a Techcrunch post on May 1st. As Techcrunch appeals to a tech-savvy audience, Twitter is favored by a wide margin. (Techcrunch posts are also widely auto-tweeted by those trying to establish a Twitter presence and build a follower list.)
Now look at site with a very different audience. This is a photo of a skateboarding baby chick from icanhascheezburger.com, where the results clearly favor sharing on Facebook by an almost 10-to-1 margin.
Here’s a second site with wide audience appeal: a break.com video of a tornado forming in front of a couple watching from a parked car. The total number of Likes and old-style Share clicks dwarfs the number of tweets it received.
Finally, we have these eye-popping results for Oprah’s No Texting Campaign, where a predominantly female audience seems to overwhelmingly prefer Facebook. (Ironically, the Twitter users might be the ones doing the texting in the first place, perhaps contributing to the disparity).
The limited data suggests to me that audience demographic strongly affects whether something is Liked (or Shared) on Facebook instead of Tweeted.
Of course, one example from each site is hardly enough data from which to draw a conclusion; so I examined a random sampling of different content on each of the same four sites to see if the preference ratios were similar. Sure enough, on each site the general preference was indeed the same. From this, it seems clear that different audiences prefer Facebook over Twitter and vice-versa.
Now, what about the content itself? Certainly, it isn’t front page news to state that some content is more shareable than others. But how much effect does it really have? To find out, I chose a selection of articles from The Washington Post, all published May 1st. By focusing one site I can be assured of a relatively consistent audience. At the same time, I can be assured of a diverse range of content since it is a general news source.
Here are the Like button results from five articles I picked, all of which were published the same day:
- An article on the gulf oil spill received 34 Likes at the time of my sample.
- A story on shifting immigration views drew 12 Like votes.
- This opinion piece on Obama and immigration reform received 10 votes.
- Another piece on the US economic recovery had 21 Likes.
- Finally, an article discussing a humorous video made by US soldiers in Afghanistan got a comparatively whopping 55 votes, the most of any item in this limited sampling. Additionally, the linked page that contained video itself had 68 Likes.
(The numbers for the examples I’ve cited are dynamic and will have changed by the time you read this.)
The top vote getters were the articles about positive or uncontroversial subjects. The two that discussed immigration reform (currently a hot-button issue in the US) were the least recommended. From this I conclude that the nature of the content plays a major role in whether it will be shared at all; that, particularly in the case of Facebook, people won’t click ‘Like’ on something they don’t want their friends to see, perhaps due to concerns about privacy or because the nature of the content may inspire unwanted debate or attention.
The fact is that there are things that humans simply don’t care to publicly share with others, things they may like but prefer to keep private. As an extreme example, do you really want your Like preference for this photo (warning: NSFW) showing up on the Facebook pages of friends, co-workers or even your mother? Contrived? Yes, but dumber things have happened in the age of Facebook. It also shows that not every site or page is a good candidate for Like buttons.
Some sites, in fact, may find the buttons do harm instead of good. Personally, I find it a little creepy when I land on a page and discover that someone I know also happens to like the same thing. And while, yes, the similar practice of ad targeting goes on every day, it happens out of sight and doesn’t affect user perceptions. But the in-your-face, we-know-who-you-are aspect of Facebook’s buttons are a different matter, one that may well work against some sites that use them.
So, for example, if you’re selling “male enhancement pills”, cures for balding, creams that purport to clear up mysterious rashes in unmentionable places or books ‘For Dummies,’ chances are the Like button will be ineffective at least and damaging at worst. People either won’t click on them or their mere presence may inhibit visitors who would much rather blissfully maintain the illusion of anonymity.
If, however, you specialize in cute animal photos, silly videos, top-10 lists or something else innocuous, you’re probably going to do well – very well.
Now, to return to why I think blog #1 did so poorly in the Facebook popularity contest. That site specializes in neither the cute nor the overtly offensive, and it has a loyal, growing and diverse following. The writer who runs the site, however, does have very definite political opinions; and there, I think, is the problem.
Writing that touches on religious or, in this case, political issues is bound to offend someone. It’s just a fact of life. Ultimately, I have concluded that visitors to his blog are choosing not to take a public stand by clicking Like, which would tend to expose their political sympathies. After all, what is a Like button if not an expression of agreement or approval?
So, this seems to explain the horrid Like button results on a site I thought would surely benefit from it. It’s also consistent with my limited sampling of other sites, where the humorous, feel-good or uncontroversial items received the most votes.
For me, this has been an illuminating exercise. Though I could be mistaken on the precise cause of why things turned out as they did, it’s clear to me that Facebook Like Button functionality is inappropriate in some circumstances and won’t perform the way we all hope it will.
When deciding whether to add Like button functionality to your site, consider that:
- Neutral content for general audiences seems likely to deliver excellent results due to its wide appeal. Humorous content will do even better.
- Pages covering controversial subjects will probably fare very poorly.
- Items that interest a narrow demographic won’t be shared widely on Facebook. Tech items, for example, will probably do much better on Twitter.
- Your user demographics will determine whether Facebook, Twitter or some other form of social bookmarking suits their needs.
- If you run a site that sells products that are shipped in plain, unmarked boxes, your visitors are expecting a level of discretion from your site that the Like button may perceptibly diminish.
Have you implemented the Like button on your site? What have your experiences been?
Read more from Michael Johnston on The Well Run Site or connect with him on Twitter.
