Social sharing buttons in all their forms are one of the most common UI elements found on a web page.
And, it seems to me, poorly measured or studied.
Chris Coyier started a great discussion with this tweet:
I wonder how many people assume that an article is un-shareable if it doesn’t have sharing buttons. And if that’s any significant loss.
— Chris Coyier (@chriscoyier) March 12, 2014
You might be familiar with this conversation:
I’ve tried it all.
- AddThis and ShareThis (both offering analytics).
- Various plugins and lightweight implementations.
What Does the Research Show?
Placing sharing buttons on your site means you are agreeing to this exchange:
The extra clutter is a given, but are readers sharing more simply due to buttons on a page? Or do the people who share have their own methods, and don’t bother hunting for buttons?
It’s so difficult to measure because there is no objective measure of what makes content shareable.
Measuring shares against pageviews is not conclusive. Was it the buttons or was the content shareworthy?
- LukeW did some research and concluded that only 0.25% of pageviews lead to a share button click.
- This great post on UK govt sites is a good read (only 0.2% shares per pageview).
- Smashing Magazine found that removing Facebook Like buttons led to INCREASED FB shares.
- Chartbeat research showed that there was no correlation between sharing and people actually reading an article.
- GetResponse email marketing found that including share buttons on emails lead to a higher Click Thru Rate on the email. A closer look at their study shows it is quite ambiguous and the actual amount of shares is tiny (7 twitter shares per 10,000 emails).
- An AB Test on a Finnish shopping site showed that removing vendor share buttons led to a better CTR on their buy/shopping cart buttons.
Not much out there.
Test #1 – Share Counts: Buttons vs Overall
I looked at existing content pages, then tracked how often a Facebook share button was clicked. I then compared this to the actual share counts (querying Facebook directly).
The results surprised me.
|Page||Share Button Clicks||Actual FB Shares|
In this case I’ve only looked at shares (as opposed to likes or comments). Around half the shares were from mobile (this reflects the traffic patterns of the site).
As a percentage of pageviews share clicks were around 0.09-0.2% (consistent with LukeW’s research).
Before doing this test I was about to pull the share buttons, but I will leave them and do more testing.
Test #2 – Social Proof: Should You Display Counts?
I have been running an A/B test on a single page that is a good candidate for sharing. It has a generic share button and I measure the first click (i.e. The user’s intent to share).
There are 3 variations of the share button:
Two metrics stand out: time on page, and the number of clicks.
|Variations||Time on Page||Clicks|
|Share button + zero count (baseline)||5:09||40|
|Share button only||5:26||51|
|Share button + actual count||5:40||64|
- As a percentage of pageviews, actual shares initiated on a page are tiny.
- People do use share buttons and a reasonable proportion of overall shares may come directly from the source page.
- Consider removing Facebook Like buttons.
- Use lightweight implementations with minimal ‘noise’ and clutter. The extra page weight from heavier buttons is probably not worth the gains (also consider how Facebook use the data from share button implementations to record browsing history – ever wondered why Facebook ads can be so close to what you’ve visited that it’s creepy?)
- Be careful with displaying share counts. Maybe only do it on pages with reasonable numbers of shares. Social proof is real!
If a page is shared, it’s not due to share buttons, but because it is useful and interesting.
However, if well thought out and integrated with the site, social buttons can be beneficial in a small way.