Influence of Page Speed on Google Rankings

[NYCTech]

---

System: The following message was cut out of thread at: https://www.webmasterworld.com/google/4851947.htm [webmasterworld.com] by goodroi - 2:38 pm on Jun 1, 2017 (utc -5)

---

@goodroi, you're suggesting page speed seems to be an issue. Is that page speed as shown in webmaster tools (ie time spent downloading a page), page speed as in Google Analytics (page speed timings), or both? If a site was slow at one point, but then improves, how long does it take to reflect that? We were targeted by numerous negative SEO attacks that slowed down both these metrics at various points, but have implemented a new firewall that seems to be working now. I'm curious when this is likely to yield improvements.

[RedBar]

In my industry in the USA, the site seen mostly at #1 and ranking in the top USA 6,000, has a site speed slower than 93% of other sites. Yep, it is seriously slow however since they are a near door neighbour of Google I assume they get a special closeness bonus alowance!

[goodroi]

I would say technical SEO issues have a direct impact on rankings but I haven't done enough research that isolated just page speed. When I have been working on improving page speed it usually impacts other aspects that can also be influencing rankings.

Page speed can definitely impacts many other SEO aspects and cause some ripple effects. Here is one possible nightmare scenario. Imagine someone launches an attack on your site which then makes your your site inaccessible. While your site is inaccessible, a website that is linking to you runs a broken link report and finds the link to your site isn't working so they just remove it. Once your site returns to normal it now has lost backlinks. I don't think this is common. It think it is more common that slower loading pages will lose users and the users that wait for the page to load won't be impressed enough to link back or refer other people to your site.

I don't think page speed is a magic cure for ranking. There are many other factors involved in Google rankings. You need to make sure you also have good content, links, etc.

Even if page speed has no influence on rankings, you still want really fast pages to make your users happy and more likely to convert.

[Peter_S]

The thing is, there are several kind of speeds. The TTFB, the time to download the content, and the time to render the content (loading images, and external elements).

[robzilla]

It's likely a small ranking factor, if it does anything at all except in the more egregious cases where a page is unreasonably slow and, as such, makes for a particularly poor user experience right out the Google gate. How they measure this is unknown. I tend to focus mostly on the Document Content Loaded measurements in Google Analytics. Big improvements in performance are likely to improve your traffic figures one way or another, simply because it's a much better user experience, and all of that feeds back into your search rankings.

[keyplyr]

It certainly has an affect on mobile. If your site is slower than Google wants it to be, Google will suggest a faster version without images or ads for users on slower connections.

[lucy24]

If your site is slower than Google wants it to be, Google will suggest a faster version without images or ads for users on slower connections.

.... which, for some webmasters, is worse than not listing the page at all, right?

[keyplyr]

IMO some sites look really bad with weblight & the other utility that strips the ads & images. It's almost like Google knows this and is using it to get webmasters to build faster, leaner responsive sites.

As far as a factor in SEO & ranking, most definitely. If your page is slow & the user is on a less-than-fast connection, there is a label next to your listing telling the user your site is slow and offering them the choice to use weblight.

Also, if your site is not responsive, there is a message offering the mobile user a tool to "make this is mobile friendly" that also strips most images & all the ads.

Solution - make optimized, fast loading mobile responsive web pages and test with Pagespeed.

[glakes]

In an effort to deal with Google's zombies, my company addressed site speed. We moved to a faster server with more resources and effectively reduced load time by 75%. Months later I can look back with complete confidence and say it made no difference in either the quantity or quality of the traffic Google is sending. In more competitive industries site speed may have more influence, but definitely not in my industry. Regardless, a faster loading site is good for users and that is what matters most.

[linkbuildr]

Top ranking sites in my vertical are way slower but have the links. Links >

[mack]

There is also an indirect effect of slow loading pages. If a page is taking longer than a user is used to, they may well click back. Bounce rate may influence ranking.

Mack.

[Peter_S]

A page which loads fast, have nice layout, no excessive ads, in other word, is pleasant and offers a "good user experience" (it's the new keyword that we see all over), is also likely to be linked , shared and liked on social networds, which can help too.

By the way, about "page speed", do you think that Google considers the time its crawler takes to download the page? Or does Google Chrome send statistics back to Google, or anything like that ?

Another aspect to take in consideration, is the location of your host/server compared to the location of your visitors.

[keyplyr]

do you think that Google considers the time its crawler takes to download the page? Or does Google Chrome send statistics back to Google

There are 2 UAs calculating the page load time. The PageSpeed utility when you manually run a test, and Googlebot itself. A third if you test for mobile.

While Chrome does gather some info, there's no evidence it is used to gather page load time data.

More importantly, for the last year or so, the load time for each page is included in the indexing data so when a user on a slow connection accesses the link in the SERP, they are offered a Google Weblight version of the page if that page is slow loading.

Pagespeed could be a huge factor if your market is rural or from 3rd world countries with slower connections.

[mosxu]

@NYCTech

I wonder how is recovery going and still have Ips slow loading your website to mislead Google?

Did you make any progress?

[keyplyr]

I wonder how is recovery going and still have Ips slow loading your website to mislead Google?

@mosxu - Did you intend to say ISPs (Internet Service Providers)?

ISPs intentionally "misleading Google" is a silly idea and makes no sense.

There is the possibility however, that since Net Neutrality has been removed in the USA, the networks will develop data speed tiers that charge customers different subscription pricing for faster speeds. This has already begun on mobile networks.

As far as ranking, everything will change soon with the adaption of the Mobile-First Index [webmasterworld.com]

- - -

[SEOLeeds]

My website scores over 80% of the google speed test but is poor on sites like gtmetrix (69%).

Should i try to fix it or is the fact that it scores well on the google page speed test enough?

thanks in advance

[mosxu]

@Key.plyer

I was referring to an IP in server logs. The IP scores from 10-60 seconds by simply asking slow the elements on of the site. Asks for html first, after 20 seconds asks for a picture than after 10 seconds one more picture and the loading time is increasing

The IP has good bandwidth so some sort of script is doing it in Chrome and Google takes it for granted as a result the site may drop temporarily !

Just asking what the sever logs are saying for NYCTech

[keyplyr]

@mosxu - So you're referring to an IP Address (Internet Protocol Address.) Which is basically saying the user's computer.

Some connections are much slower than others and that will be reflected in your server logs.

However, what you are describing is likely a bot. Many scraper bots will pace their requests in an effort not to be detected. Good work on catching it.

[Peter_S]

My website scores over 80% of the google speed test but is poor on sites like gtmetrix (69%).

Should i try to fix it or is the fact that it scores well on the google page speed test enough?

All efforts you'll achieve to make your site faster will profit you and your users.

Now I still wonder how Google estimates the speed of a site, for its ranking. Because, if Googlebot comes exclusively from the USA, sites hosted in North America are favored compared to site hosted in Europe or Asia. Or may be Googlebot is running / sampling from different locations to better estimate how fast a site is in average.

[martinibuster]

I believe there's an influence but that there are many more important factors that will outweigh page speed.

Ultimately the page that will rank at the top is the page that Google's algorithm determines solves the user's problem, (for most cases) regardless of page speed.

[Awarn]

I believe it has little influence. My results were similar to Glakes. I had pages scoring 97%. Never saw any effect. Don't worry about speed unless you know you are slow.

[mosxu]

Running a pagespeed tool on a site is a joke

To be in the mobile index any page of the site has to load under 1 sec and do not put this down to bandwidth.

How many IPs are loading pages under 1 sec looking in server logs?

[keyplyr]

Running a pagespeed tool on a site is a joke

Maybe to you... but to Google, these are the qualities it considers, or will consider, in the index*.

Don't forget, Google has a documented history of badgering site owners long prior to bringing down the hammer. Just look at HTTPS Warnings [webmasterworld.com].

In a month (October 2017) Google will be showing all visitors blatant warnings if your pages aren't secure, starting with the LogIn & Payment pages, but moving to all pages soon after. They've been strongly pressuring this for well over a year and now it's here.

PageSpeed [developers.google.com] has been around for a couple years, but has been updated several times lately to adhere to Google's new guidelines for Mobile... this in advance of the upcoming *Mobile-First Index [webmasterworld.com] which will rank both desktop & mobile according to how your mobile page performs.

[mosxu]

@key.plyr

Are you sampling all your pages for speed in analytics? If so how many pages load under 1 sec?

[keyplyr]

Are you sampling all your pages for speed in analytics?

If we're talking about my *personal* website, then yes I test all of the pages using PageSpeed (I do not use Google Analytics) since all (about 300) are indexed and draw direct traffic. While I don't get perfect scores, my pages do very well and I'm satisfied.

...how many pages load under 1 sec?

None of my pages load under 1 second. I don't have that type of website with those types of pages. My pages are content rich articles with charts & graphs usually accompanied by photos and images.

If you have the type of pages that do not require mixed media and heavy content, I recommend considering AMP [ampproject.org]