Updated Google Mobile Speed Test - How You Rank Against Competitors

[Sally Stitts]

Have you tried the updated Google speed test for mobile?
There is a post from Barry at Search Engine land that provided the link -
[testmysite.thinkwithgoogle.com...]

I'm good.
TOP Sites = 4 seconds
My score = 2 seconds

Check yourself out.

[engine]

Yeah, thanks, that's a useful tool, although I always read it with a pinch of salt.

[Sally Stitts]

Google = 1 second
Bing = 4 seconds

HA!
.
I finally got my complete report in my email (took an hour) -

"See what went into your results
SPEED :
Excellent
Should fix
Consider fixing
Looks good
Reduce page weight

Compress images

Compress resources with GZIP

Minify resources
Reduce number of requests

Leverage browser caching
See how storing data in your local browsers can reduce network delays »

Eliminate render-blocking JavaScript and CSS in above-the-fold content

Avoid landing page redirects
Make a quick first impression

Load visible content before CSS and JS files

Reduce server response time"

ALL were green, except for browser caching. Too bad I have never been able to get it to work.
I find it EXTREMELY complex. Have you seen the maze graphic explaining how to do it? Mind-boggling.
Anyone know of a SIMPLE tutorial, explaining minimal, functional implementation?
I better go study it A LOT, and get it working.
.

[Wilburforce]

Anyone know of a SIMPLE tutorial

In .htaccess (edit as necessary):

<filesMatch ".(css|jpg|jpeg|png|gif|js|ico|ogg|mp3|wav)$">
Header set Cache-Control "max-age=2592000, public"
</filesMatch>

where the first line sets file types, and the number after max-age= is the time in seconds (so 2592000 is 30 days).

[ken_b]

2 seconds

But when it says the page beat the industry average there doesn't seem to be an apparent consideration of the toatal page size.

That has got be to a big factor in an accurate comparison.

My 200 kb page is faster than the other guys 200mb page ... how helpful is that to know?

(Or did I miss the size comparison somewhere?)

[keyplyr]

I loaded a completely blank page (no images, no scripts, no html) and the report said: Loading time: 9s poor.

I loaded a few different pages on the same server w/ rich content, several scripts, images and Adsense and the reports all said: 2 seconds or 3 seconds excellent

[iamlost]

I ran WebmasterWorld for the fun of it:

Testing on a standard connection (3G)...
70% of cellular network connections globally will occur at 3G or slower speeds through 2020.

...

webmasterworld.com
6s
Loading time: Fair
24% Est. Visitor loss
(Due to loading time)

It's interesting but the real test is not measuring your speed against competitors but against yourself. If you cut 3G load time from (using WebmasterWorld example) 6s to 5s to 4s to 3s how do your metrics (and especially your bottom line) improve each step of the way? At what point is the ROI telling you better to put your effort elsewhere aka where does it plateau...or does it?

[keyplyr]

70% of cellular network connections globally will occur at 3G or slower speeds through 2020

Well not where I live. 4g is standard and has been for years. In fact, the carriers here are laying the groundwork for 5g implementation. The modems are already being sold (I have one) the phones are ready for release and television commercials often mention 5g is coming.

However, WiFi is free all over the city and the only time I even use the mobile network is when I'm outside city limits.

[nomis5]

70% of cellular network connections globally will occur at 3G or slower speeds through 2020

And that's not true from a UK perspective either as far as I am aware.

The point is that Google is looking worldwide (if their statistic is true) and that has very little relevance to the situation in many developed countries. I'm certainly not developing my UK based websites on the assumption that 70% of my viewers will be on 3G connections.

What seems odd to me is that clearly G has the ability to determine how many cellular connections are 3G, 4G or even 5G so why isn't that information available in Analytics? They keep banging on about site speed, AMP etc.

The updated mobile speed test seems like some sort of desperate attempt to convince webmasters to speed up their websites. As someone in a previous post pointed out, comparison is meaningless unless the size of the page is taken into account. And not only the amount of text, the graphics can also have a huge impact on the page download speed.

Speed is good, I understand, but something about the way it's being pushed by G smacks of an additional ulterior motive.

[keyplyr]

One thing I noticed with this tool that is also true for the older PageSpeed tool is that caching is recognized if done at the page level, but not recognized when done at the server.

[seobox]

After the test, I think I need to change another hosting provider. :(
14s
Loading time: poor

[lucy24]

70% of cellular network connections globally will occur at 3G or slower speeds through 2020

“Globally” does not mean “in the UK”. “Globally” does not mean “in a major American city”. “Globally” means globally. That includes the vast proportion of the Indian population whose only Internet connection is via an elderly Android. (Dunno about China, since I block them, but I can't imagine it's all that different.)

You can see slow connections in logs. They're the ones who load up a page, and some of the supporting files--and then close the browser window before the rest has even been requested. (Most of the time admittedly because they've had time to realize that you're not the site they're looking for, so the sooner they clear out, the better.)

[keyplyr]

...because they've had time to realize that you're not the site they're looking for...

Really? That's never happened to me. Maybe you could put up some cute pics of kittens playing with yarn?


(joking of course)

[lucy24]

Maybe you could put up some cute pics of kittens playing with yarn?

Hey, you've seen my site! Joke.

In the case of my Indian visitors, what they appear to be looking for is hardcore pornography involving {weird combination of search terms which all happen to occur in the course of a 250-page novel} or {bizarre misspelling that happens to correspond to actual word in Old English}. I single out India because it's one of the few areas where Google search terms are still visible (in the referer) more often than not -- and their connections are often so slow that they really do have time to close the window before everything has been requested.

The site linked in this thread says I took 2 seconds--admittedly for a page that only has about six pictures--and the chance of people packing up and leaving is “Low”. Now, if I were designing this Results screen, I would not have the word “Low” in huge print indicating a positive result, but that’s me. So I got lucky and tried it at a time when my server was on an upward hiccup.

[RedBar]

That includes the vast proportion of the Indian population whose only Internet connection is via an elderly Android.

I don't know who your Indian visitors are however the vast majority of mine are using up-to-date smartphones with very fast connections even out in the countryside. I communicate immediately 24/7/365 with many factories and very little downtime ... read monsoon periods.

Insofar as China is concerned, the largest smartphone market, I also have very few issues.

Over a third of the world’s population is forecast to own a smartphone by 2017, with China predicted to hold over 30 percent of this market. Both India and China have rapidly growing economies and they are expected to hold almost 50 percent combined of the global smartphone market share in 2017, in comparison to the estimated 20 percent they held in 2011.

Check that, China and India 50% of the global market:

[statista.com...]

[EditorialGuy]

Our site is a lot faster than the top performers in its industry, according to the test. (CloudFlare probably deserves a lot of the credit for that.) I'll look forward to seeing my free report.

I do have one question: Isn't this really a test of a page, rather than of a site?

[keyplyr]

Isn't this really a test of a page, rather than of a site?

Yes, you're testing the web page, not the entire site. Most of the speed tests tools test individual pages.

[JS_Harris]

After you get a good baseline don't forget to turn off 3rd party stuff(ads, tracking etc) and run it again to see how much latency they add to your page(s).

[MayankParmar]

Tested using new Google's Page Speed. My site speed is 7s, I will be abandoned lol.

[edited by: not2easy at 4:41 pm (utc) on Aug 3, 2017]
[edit reason] merge cleanup [/edit]

[MayankParmar]

"According to the company, about 40% of website visitors leave a site if it takes more than three seconds to load."

I really don't think 40% of the visitors will leave the site. There are only a few sites which load under 3 seconds.

[romerome]

This is strange. So if I test a page on pingdom I get 1/2 of a second (san jose as destination). Even if I test in Australia I get 1.73 seconds.

For for google I get 5 seconds for the same page.

For me the page loads more or less instantaneously.

Is G really testing the page speed or analyzing the page and assuming it takes this long to load.

If not why would pingdom and G have such different numbers

[Peter_S]

If you are running adsense or other ad network, can you tell if the ads are loaded when testing with Pingdom? It's possible that ad networks are blocking Pingdom's IP.

Also, which "page speed" are you using at Google? Because Google might shows the speed on mobile device using 3G.