Does PageSpeed Actually Impact SEO? [New Experiment]

– Does site speed actuallyaffect your Google rankings? Well, I recently decided to put this ranking factor to the test, and the results might surprise you.

I'm Brian Dean, the founder of Backlinko, the place where marketers turn for higher rankings and more traffic.

And in this video, I'm goingto show you the results of my SEO experiment, keep watching.

You've probably heard that page speed is an important Google ranking factor.

There are a million blog posts out there that tell you how importantpage speed is for SEO.

Plus, a few years ago, Googlerolled out a page speed update that specifically down rank sites that loaded slowly on mobile devices.

I mean, even a large scale analysis of one million Googlesearch results that we did found a correlation between loading speed and first page rankings in Google.

But I haven't seen an experiment that isolated Page Speed asa potential ranking factor, which is why I decided to run one.

Here's what happened, a while back I noticed that a lot of our pages, eventhose that ranked number one for competitive keywords hadhorrible page speed scores.

For example, this page thatranks in the top three of Google for SEO checklist, scores an 18 out of 100 in Google PageSpeed Insights.

Now, I was pretty sure thatGoogle use loading speed as a ranking factor.

But seeing all those slow pages that were ranking for competitivekeywords made me wonder, is page speed one of those minor factors that don't really make a big difference? Or is it something that willmake or break your rankings? To find out, I wanted to improvethe speed of a single page and then measure how that page's rankings and organic traffic changed over time.

I ultimately decided to runthe experiment on this page, a list post of SEO tips.

Now this seemed likethe perfect page to test because it was already ranking at the bottom of the first page for my target keyword, SEO tips.

And if that page moved upto the top of the page, that will be a strong sign that whatever I did made a big difference.

So, before the experiment began, this page ranked numbereight for SEO tips, and was bringing in about 150organic visitors per week.

These numbers had been stableover the last few months.

And when I ran the page onGoogle PageSpeed Insights, it had a score of 28 ondesktop, and 13 on mobile, which are absolutelyhorrible scores all around.

That page's score onWebPageTest.

org wasn't any better.

According to that test, mypage took 40 seconds to load, 40 seconds.

So those were my benchmark numbers.

Next, it was time to actually improve the loading speed of that page.

So Backlinko CTO and I decided to dig in and figure out why thispage was loading so slowly.

On the surface, this pagereally shouldn't be slow in the first place.

We use an optimizedlightweight WordPress theme, compress all of ourimages and we use a CDN.

As it turned out this page loaded slowly, mostly because it had lotsand lots of high res images.

In fact, that page's totalfile size was 25 megabytes, that's bigger than thesizes of Contra III, Star Fox, Super Metroid, andDonkey Kong Country combined.

Now a study that we recently did that analyzed five million web pages found that large pages load alot slower than small pages.

No surprise there.

And no matter what else youdo to optimize it for speed, a big page is just goingto be slow, period.

So when it comes to page speed, there's no such thing as a free lunch.

So the first thing we did was strip out all of thosemassive images, that helped, but it didn't really makeenough of a difference.

For this experiment to holdwater we needed to go extreme.

In other words, to measure theeffect of page speed on SEO, we needed to take a pagethat was running super slowly and make it load lightning fast.

And to do that our CTO Lloydcreated a version of our page that was stripped down to the bare basics.

Specifically, here are all the changes that Lloyd made to thatpage to speed things up.

That's right, he even went as far as to remove the replyfeature from comments, because that slow thingsdown by a millisecond or two.

And here's how the new fasterversion of that page looked.

Now at this point, Ishould point something out, even after all those massivechanges to the site's code, the actual content was 100% the same.

The title tag, h2, keywords, and all that stuff was exactly the same as the old post.

The only real difference was that the page loaded much faster.

In fact, our GooglePageSpeed score went from 28 to 100 out of 100 ondesktop, a huge difference.

And according to WebPageTest, the page now loaded in 1.

68 seconds, a 40 times improvementcompared to the old post.

So once the new page was live, we used the Google Search Console to index the page and waited.

(clock ticking) I waited two weeks togive Google enough time to get used to the newer, faster version of the page.

And for them to measure anysort of user experience signals that might be different nowthat the page loads quickly.

After two weeks, the ranking for that page for my target keyword wasexactly the same, number eight.

Also organic traffic to thatpage was largely unchanged.

The slow page broughtin 305 organic visitors in the two week periodbefore the experiment, and the faster page broughtin 311 visitors from Google for the two weeks that it was live, essentially, the same amount of traffic.

What I found super interesting was that the bounce ratewas actually 48.

84% worse on the fast page.

This might be because usersget their answer quickly then bounce.

However, average sessionduration and page views were both higher on the faster page, which shows that, overall, people tended to interact with my site longer when they landed on afaster page, makes sense.

In fact, improving page loading speed made so little difference for SEO, that we decided to roll back to the old slow version of the page.

So what are the takeaway lessons here? Now it's hard to draw anysort of firm conclusions based on a singleexperiment on a single page, plus my target keyword, SEOtips, is pretty competitive.

So it could be that the SEO boost that we got from a faster page just wasn't enough toincrease the rankings for that competitive term.

There are lots of huge authority sites ranking above me for that keyword.

So optimizing for justa single ranking factor might not be enough to shootit to the top of the results.

So yeah, things might have been different if we ran this experimenton a page that was ranking for a low competition keyword.

That said, if page speed really was a super important ranking factor, I would have expectedto see some movement, especially considering the dramatic change in PageSpeed scores.

So yeah, this datasuggests that page speed might not be some sortof SEO magic bullet.

And at least according tothis one small experiment, improving a page's loading speed may not actually impactyour Google rankings, at least when it comesto competitive keywords.

So did you learn somethingnew from this video? Then make sure to subscribe tomy YouTube channel right now.

Also, if you want exclusive SEO techniques that I only share with subscribers, head over to Backlinko.

com and hop on the newsletter, it's free.

Now I want to turn it over to you.

Have you seen page speedimpact your Google rankings? If so, did it make a bigdifference or a little difference? Let me know by leaving acomment below right now.

This is a shot of me waiting, so we can just wait a second.

Okay, I think that's good.

Even all.

.

.

Okay, that's good for video number two.

.