New!Schedule Threads!Threads scheduling has arrived!Threads scheduling has arrived on Buffer! Find your community now.Threads scheduling has arrived on Buffer! Find your community and keep the conversation going.Learn more

7 Powerful Social Media Experiments That Grew Our Traffic by 241% in 8 Months

Mar 16, 2016 8 min readContent Marketing

If you’ve asked this before…

“How can we get more visitors to our website?”

… You’re certainly not alone, as increasing traffic is often the number one problem faced by marketers today.

The bad news? Saying “get more traffic” is easier said than done. You could write guest posts (Leo wrote 150 articles in 9 months when Buffer first launched), optimize for SEO traffic, or drive visitors through social media. The options are endless. This article focuses on the latter, though.

In this post, I’ll share the seven most powerful lessons we learned at Hubspot from running social media experiments to increase our social media referral traffic by 241%.

Ready to dive in?

pablo (35)

On the Sidekick blog (which now redirects to the HubSpot Sales Blog), we ran a series of social media experiments to drive more traffic. The result? A 241% increase in monthly blog traffic over eight months:


During this eight month period, we ran hundreds of social media experiments (shoutout to Brian Balfour on creating an experiment-driven culture) to grow our users and increase traffic. And throughout our journey of building a growth machine, we discovered a few learnings related to social media sharing along the way, and I’d love to share these with you below.

Note: HubSpot is a publically-traded company, so we’re not allowed to share actual data. As a result, all of the numbers here are fake, but I promise the learnings and experiments are all very real.  ?

1. Giveaways 2x shares per blog post

When we A/B tested a blog post on body language, we split it between two variations:

  • Version A had a giveaway at the end
  • Version B did not

The result? Version A doubled the amount of shares. In other words, instead of 1,100 shares, we got 2,200. Here is the simple CTA we used at the end for version A:


We kept running these tests and seeing the same results — insert a giveaway, 2x the shares on the article. The compounding effect on traffic for this experiment was monumental for us.

2. Inserting Click-to-Tweet throughout articles would boost shares

In the below spreadsheet, we analyzed articles that used Click to Tweet (CTT) links versus articles that did not:


Notice a pattern? In column E, you’ll see articles that included Click-to-Tweet links were amongst our most-shared articles on Twitter.

Seems obvious, but it was refreshing to see data that backed up our hypothesis.

At a glance, you might be thinking, “Awesome! I’m going to add CTT links to every article now!” … but I’d recommend testing everything. This might work with our audience but could have a profoundly different impact on your audience.

Our philosophy for CTT was if it doesn’t feel natural, don’t force it. The most common CTT links we used were quotes or interesting statistics.

3. Twitter was our “most valuable” share source

Just because you’re getting more shares, doesn’t mean you’re getting more traffic.

For example, after someone (let’s call him John) Tweeted our article, the goal was for someone else to click the link that John Tweeted. If someone new did NOT click that link, someone new did NOT visit our blog. If that happens, John’s Tweet isn’t delivering any return, because we’re getting zero new visitors to our content.

Thus, the more people that click a link after it’s shared, more “valuable” the share becomes. Make sense?

By using a tool called Filament, we could automatically calculate our most “valuable” social media source through this simple equation:

Total Pageviews Per Network
Total # of Shares Per Network

During a 30 day period, here were our results:

In other words, that means:

  • For every share we get on Twitter, 2 people clicked that link they saw in their Twitter newsfeed
  • For every share we get on Facebook, 1.5 people clicked that link they saw in their Facebook newsfeed
  • For every share we get on LinkedIn, 0.75 people clicked that link they saw in their LinkedIn newsfeed

For example, if we got 1,000 shares on Twitter and 1,000 shares on LinkedIn for the exact same article … we could expect 2,000 visits from Twitter (1,000 shares * 2 views per share), but only 750 visits from LinkedIn (1,000 shares * 0.75 new views per share). Despite them, both have 1,000 shares.

That means getting shares on Twitter is more valuable than getting shares on LinkedIn, even though LinkedIn has the highest number of shares. Interesting, right?

This opened our eyes to not only track the total number of shares but also keeping in mind our most valuable share source.

4. Visualizing our “share retention” over time

In this experiment, we sought to understand the rate that shares decrease over time per article. For our blog, the data looked like this:


The Weekly Cohorts (on the left column) are the average number of shares for all articles published that week (we published three articles per week on average). For example, in Weekly Cohort A, the average number of shares for three articles after one week was 203. After two weeks, it was 264. Etc.

The average shares for ALL weekly cohorts (highlighted in yellow) is the most important part of that data. The average shares an article would get one week later was 309. Two weeks later was 414. Three weeks later was 479. And so on.

This graph visualizes how our shares decreased over time:


5. Influencer sharing articles results in significant shares spike

In March 2015, we published an article about Benjamin Franklin’s daily routine. On July 20, 2015, that article had a total of 181 shares. One day later that article received an additional 244 shares … in one day:


Huh?! What happened?!

HubSpot’s CEO, Brian Halligan, retweeted the article. As a result, the total shares on an article published three months ago nearly TRIPLED. Next thing we know we’re getting thousands of new people reading that blog post, discovering Sidekick.

Yes, seems obvious. Get an influencer to share your article and you’ll get more shares. Duh. But since this happened, we started reverse-engineering our writing, constantly asking ourselves, “What influencer would share this article?” then working backwards to make sure it appealed to them.

For details on content promotion strategies, check out this incredibly helpful article.

6. “The # Habits of ____ People” article

I was initially hesitant of giving away this secret …. but here goes nothing.

There is a magical headline and blog post framework that you might not be aware of yet. It follows this structure:

The # Habits of ___ People

Can you guess what our two most shared articles of all time were at Sidekick? Yup, they were:

  1. The 13 Habits of Hyper-Productive People (2,500 shares)
  2. The 9 Habits of Insanely Likable and Charismatic People (2,000 shares)

I’m not alone here. What some of’s most shared articles?

What about
Or how about this one book you might have heard of before?

Why does this headline and framework consistently get shared?

People want to emulate the “habits” of those who have characteristics they desire. We aspire to be more productive, more liked, more mentally tough, more successful … so when we’re given the habits of other people who have those qualities, we pay attention.

But since we aspire to have these qualities, we share the advice with others, since that’s how we want to be perceived. The New York Times ran a fascinating study on the “Psychology of Sharing” which dives deeper into WHY we share content online (such as “The # Habits of __ People”). I’d highly suggest giving it a read.

David Khim wrote a similar article following this framework on the HubSpot Marketing Blog: The 15 Habits of World-Class Content Marketers. The result? One of the most shared articles of the year.

Try this article framework on your blog and see how your readers react. It works.

7. More shares doesn’t necessarily mean more product registrations

Time for a curveball.

Sidekick is a free email productivity extension for Google Chrome. So our content strategy was centered around one goal: Get more people to sign up to Sidekick via our content marketing.

A long-standing question had been for us whether increasing social shares is a valuable goal for product registrations. In our case, the answer was clear …. social shares does NOT contribute to more Sidekick registrations:



Data uncovered that more social shares isn’t necessarily contributing to new users. As a result, we decided that focusing on social shares isn’t the most effective use of our time for getting new Sidekick user registrations. Instead, we pivoted our content strategy to focus on other tactics.

These are other tactics are covered in my coworker’s, Anum Hussain, recap of building a B2B2C content strategy. Examples include building a “Hub and Spoke” strategy, doubling-down on content more related to our product, and creating co-marketing campaigns with other companies.

The irony of optimizing for social shares, then realizing they don’t have a direct correlation on product signups might make you think, “Well what is the point of getting social shares?”, which is a fair question. This all ties back to the core disclaimer of the entire article:

Different businesses have different customers. What works for our company might not work for yours.

Perhaps social shares might correlate to product signups for you. But unless you run the experiments and analysis yourself, you’ll never know.

Let data guide your decisions, listen to your audience, and you’ll be surprised how fast you’ll grow.These results were powerful for our audience, but YOUR audience will be different. Make sure you treat this as inspiration to run experiments, not prescription. A/B test these ideas on your audience and don’t simply borrow these lessons without running experiments yourself. Different customers react differently.

Over to you

These results were powerful for our audience, but your audience will be different. Make sure you treat this as inspiration to run experiments, not prescription. It’s always important to remember that different customers react differently and what works for one company may not work for another.

Thanks for reading! And I’d love to hear your thoughts on these experiments or any of your own in the comments below.

Brought to you by

Try Buffer for free

140,000+ small businesses like yours use Buffer to build their brand on social media every month

Get started now

Related Articles

Content MarketingJun 20, 2024
Reddit is My Surprising Secret to Marketing Success: Here’s How I’ve Used it to Grow My Business

Learn exactly how this founder used Reddit to build his profile as a thought leader and gain new customers.

Content MarketingJun 13, 2024
Content Distribution: Your Everything Guide, From Psychology to Strategy

Without a solid distribution strategy, even the best creative can get lost in this vast universe of content. Here's everything you need to know about building yours.

copywriting formula
Content MarketingMay 23, 2024
10 Copywriting Formulas to Convert Skimmers to Buyers

In this article, you’ll learn about ten copywriting formulas that can help you package your social media content in a more engaging way.

140,000+ people like you use Buffer to build their brand on social media every month