Rand Fishkin of Moz wrote a fascinating Whiteboard Friday article recently, making the claim that except for the top 10% of content producers, social media sharing appears to have no impact on SEO. Go read the article first.
This struck me as a very broad claim. I wanted to see how I might find that out for my site. As with so many things in digital marketing, your mileage may vary, and I felt instinctively certain that my experience is different from the aggregate.
Using one of the many SEO tools at my disposal thanks to SHIFT Communications’ data-driven marketing technology toolkit, I took a look at my website’s inbound links vs. social shares. I did a Spearman regression with Rand’s hypothesis that there should be absolutely no correlation between social shares and inbound links.
What did I find?
Above, we see a modest positive correlation between social shares and referring domains (which are the domains that contain inbound links to my site). While it’s not amazing, it’s also not zero.
I even took a look at Rand’s own site, Moz.com (and Rand, that link with equity is on the house):
Above, an R value of 0.445 with a p value less than 0.001 is far from no correlation. In fact, it’s a moderately strong correlation.
What we’ve proven is that the aggregate statement “social has no bearing on inbound links” is an extremely broad statement. As with anything in metrics, analytics, and statistics, there is no substitute for doing your own work, with your own data, and analyzing for yourself. You can start with a hypothesis derived from a broad, aggregate statement, but do not make the fatal mistake of assuming the aggregate whole also represents your business. You’ll drive your business into the ground.
You might also enjoy:
- You Ask, I Answer: The ROI of Data Quality?
- Cómo decide Google Analytics el seguimiento de atribuciones - Christopher S. Penn - Orador principal de ciencia de datos de marketing
- How to Think About High Bounce Rates in Google Analytics
- You Ask, I Answer: Best Language for Marketing Data Science, R or Python?
- Advanced Content Marketing Metrics: Reading Time, Part 1
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers