It has long been accepted that page speed and SEO go hand in hand, but I ran an experiment last month that shows strong evidence that page size also directly affects search rankings.
It started a couple months ago, just by adding a simple chat widget to our home page. It’s purpose was to allow you to get any questions answered in real time. After I added it I was immediately alerted (by MachMetrics’ alerting features) to the increase in doc load time (of course MachMetrics monitors its own site). It didn’t alarm me too much since the visual complete didn’t change, it was only adding some post-interactive load time, which I was willing to accept.
It also added about 200 KB to our page size. That’s normally not that much, but when your base site is about that size it’s almost doubling the page size.
However, after a few weeks of it being on I noticed our page ranking dropping quite a bit on a few of our top search terms. Take a look at these two graphs, which have about the same time scale:
You’ll notice a bit of a lag in the effect. Perhaps, if this implication is true, Google gives you some slack if you have a few temporary large page loads, but then penalizes you after it’s consistent.
The bump up after we removed the widget was also delayed. Perhaps time time it takes to spider your site also weighs in here.
Could it just be a coincidence?
Absolutely. Perhaps there were some blog posts that gained traction around this time. Perhaps other high ranking sites made changes that affected their SEO. Or perhaps it’s not specifically to page size, but page speed – look at how the page size directly influenced how long it took to load the page:
But here’s what makes me think it’s not a coincidence – this didn’t just happen on 1 keyword. It happened across many. Check out some others:
Perhaps not as striking, but there’s no denying a period of high ranking following by low ranking – which then bumps up higher again.
Regardless, it’s important to keep your page size lightweight
I know there’s going to be people pointing out the million other factors that could be at stake here, and they’re not going to be wrong.
But just keep in mind that everyone benefits from having an efficient site that loads with the least size and amount of resources. Ignoring the SEO relation, it usually results in a faster site, less strain on your servers, and happy visitors/customers.
Page size monitoring is included as part of the page speed monitoring service of MachMetrics. I’d love to hear if other have the same experience – let us know in the comments below.
How are you monitoring the SEO rankings?
Good question – we’re using the search ranking widget within Cyfe.
I’m not sure this is very scientific! There’s not enough data here to draw any kind of conclusions. Would love to see and in depth post about it though
You’re correct – it’s still up for discussion, but the data definitely trends towards the correlation here.
It would be nice if you had a reference in the y-axis scaling – its unclear whether baseline is 0.
Good point. Y-axis on page ranking graphs is actually reversed – from position 100 (bad) at bottom to position 1 (best) at top.