Reason for slow speed in 2018: Google Adsense & Analytics

the two biggest issues that are reducing page load speed in 2018 are Google Adsense Ads and Google Analytics code.

But what about bad site design?

What type of minimalistic designs are you running where Google Analytics code would impact anything load wise? It literally is millisecs to load.

I see your issue with Adsense, but you can also block things to help. More filters would make it better for us publishers.

It's well known that using Google Adsense Ads in combination with Google Analytics code is huge in slowing page load.

I removed Google Analytics code across all sites and saw significant reduction in load times.

There is really no reason to use Google Analytics. It's very poor in reporting accurate data and misses way too much.

There are so many more efficient ways to process server data.

All depends what you call "slow speed". Since Adsense ads are loading asynchronously, they are not slowing down your pages. Ads can appear late, but "your" content, is not affected, so your user shouldn't be impacted, and still see your page, and your content, at the speed your server is delivering it.

It's well known that using Google Adsense Ads in combination with Google Analytics code is huge in slowing page load.

Is it? I'm not aware of that. Did a quick search, found nothing. What would it be about the combination of the two that slows down a page?

I removed Google Analytics code across all sites and saw significant reduction in load times.

I'm curious, how did you measure that? In the browser? I don't see any difference between adding or removing Google Analytics, apart from the obvious tracking events, which fire asynchronously and only after DOMContentLoaded fires.

Some parts of the AdSense code do run at an earlier stage, and afterwards there's obviously a lot going on in terms of downloading, rendering and tracking that logically slows things down. Nobody likes ads, page speed being just one reason for that. It's a trade-off.

Very strange that Google is keen to add webpage load speed as an SEO factor, but the two biggest culprits are its own doing.

This sort of thinking always fazes me, because it's not strange at all. Search, AdSense and Analytics are separate products. I don't agree that Analytics slows you down, and of course any ad network is going to affect your page speed -- and there are limits to what the networks can do about that without significant revenue lost on all sides. Ultimately, page speed is your responsibility. Is AdSense too slow? You have plenty of options: disable display ads, block third-party networks, sell your own ads, or remove the ads altogether. Again, it's a trade-off.

Search, AdSense and Analytics are separate products

Indeed. It still surprises me when some are not succeeding to understand this. The example being about the debates about Adsense things being sometimes against Google Search rules.

Page load speed is not the major factor. Just make sure you are not getting server errors excessively. Site down will cause drop in ranking compared to slower loading page. If main site content loads in timely manner you are fine. People will stay.

I'm curious, how did you measure that? In the browser?

Google Pagespeed

Lots of sites measure speed including Google Pagespeed, GT Metrix, Pingdom, and Webpagetest.

For one of my Wordpress sites with 3 Google Ads, several of the above gave me about 15 JavaScript files associated with Adsesense and Analytics that need to be combined into "as few as possible". Mostly Adsense related files if I recall. I have used various plugins to combine JS files, remove render blocking etc... but nothing works for the Adsense .js files.

Also, many other suggestions for improvement from the website speed measuring sites I listed above show issues with Adsense and Analytics.

My guess is 90 percent of the problem is with Adsense (including JavaScript files and unoptimized large images), and 10 percent of the problem is Analytics.

I host another site on the same host with the same theme and plugins, but no Adsense or Analytics, and the speed for that site is consistently in the 96-99 percentile of all websites range.

My guess is 90 percent of the problem is with Adsense

Adsense is definitely the main offender. That's the price we pay for being in the ad revenue game.

My point above with Google Analytics is there's really no need for it to be on our pages. It is not very accurate. It mistakenly identifies many bots as human and is generally inaccurate.

I've had this discussion a dozen times with web developers. They all agree that Google Analytics is a drag on performance. Of course they put it on client's sites, but wouldn't use it on their own.

Using it on a dedi or VPS with HTTP/2 may not be much of a performance hit, but on some shared hosting plans, it really counts.

There are many other more efficient ways to process server activity. I do all this off-site. Nothing to slow page loads. If you are moderately tech savvy and have access to your raw server logs, that is the best way to get actual data without being "interpreted" by stats (analytics) software.

Download your log a couple times per day and manually examine it with a text editor. In time you learn what to look for. Investigate who and where your visitors come from, and what they do on your website.

If you find looking at a raw log overwhelming, once it is on your local machine you can process it with Analog [mirror.reverse.net] which is free, accurate & highly customizable web log analysis software.

Also, many other suggestions for improvement from the website speed measuring sites I listed above show issues with Adsense and Analytics.

Suggestions given by page speed tools are based on fairly black-and-white performance guidelines, and there are other things to consider when you run an ad network that serves billions of requests per day. For example, while I'm sure that in an ideal situation you would combine those 15 javascript files into one, in this case those scripts probably stem from different parts of the ad request chain (one of many different possible chains), so they cannot be combined even if you wanted to. This will still trigger a simple performance alert, but there's nothing you can do about it. Google has some of the most knowledgeable people in the industry, my point being: if AdSense could be faster, it would be faster.

Where are you conducting these speed tests? Your machine? Another machine? Have you had someone in a different locale run the test and report? Is it consistent?

I have used a few different personal and client machines as well as my cell phone and tablet.

Most of the speed measuring sites that I listed earlier allow you to check speed from different locations around the world. They are all extremely well known and reputable sites.

Can confirm.

GTMetrix (Dallas) speed test: 1.32s without ads, 7s with ads.

Analytics page speed reporting is affected by AdSense as some AdSense ads keep on loading "connecting, waiting for...".

But all of this is not slowing down "your" content. Here, you are all considering that what matters is the time for the page + images + ads to be fulled loaded. But in real life it doesn't matter. What matters is that your content is presented to the user quickly. It doesn't matter if the content bellow the fold is taking some more times to load for example.

If you're bothered by the time that it takes for AdSense ads to display, having fewer ad units ads might help. Sometimes less is more.

Also, the AdSense Ad Balancer lets you reduce the overall number of ads served with minimal impact on revenue. (See Webmaster World's Google AdSense Forum for discussions of the Ad Balancer.)

I have occasionally seen AdSense ads fire a request for the Google Analytics script, even though I'm not entirely sure what the point of that would be. Anyway, if you see something like that, don't let that fool you into thinking Analytics is slow.

What matters is that your content is presented to the user quickly. It doesn't matter if the content bellow the fold is taking some more times to load for example

The time it takes for the entire page, and all of its components to load, is especially important now we're going into the mobile-first index where page speed is an even more significant ranking factor than before.

Bottom line... if the page doesn't rank well, fewer users will see it.

It's one ranking factor among hundreds. It can make a difference in extremely competitive situations, where lot of pages are providing the same relevance for a given search, but in everyday situation there are much more important things coming first.

Google never mentioned which kind of page speed factor(s) they are taking in consideration. TTFB, First Paint, Time for the content to start showing to the user is more important than the time a whole page takes to load with all elements and ads.

Each ads are not loading at the same speed either, a third party ad will load slower than an ad directly served by Adsense. A text ad will load faster than a static display ad image, which itself will load faster than an animation. So will Google load the page several times, to find the average speed load? or is the speed load achieved at the level of the whole site? What if Adsense is under DDos, and ads being served slowly than usual, will all pages serving ads be downgraded bellow other pages which contents are less relevant?

I am sure that if you check the speed of pages which are ranking above you, they are not loading faster than yours. And it won't change, even with the full mobile index thing.

Google never mentioned which kind of page speed factor(s) they are taking in consideration...

All their tools rate page speed on the time it takes the page to complete loading. I don't see any room for interpretation.

If there are no official statement that Google uses the value returned by its page speed tool, then it's an interpretation/guess.

It's not because Google's page speed presents a simple and single information that it means that Google search engine is not taking in consideration more detailed information like the resource loading waterfall for example. Chrome's developer tool provides uses it, I see no reason Google Search wouldn't.

I would consider Google's Pagespeed results to be more definitive of how Google judges page speed over anyone's opinions... but believe what you like.

I have to agree with TravisDGarrett; the full load time as a metric is not as strongly related to the user experience as other metrics like DOMContentLoaded. PageSpeed Insights does not measure speed, and a low-scoring page can be faster than one with a high score. It's just a tool that gives you insight into possible optimizations. With the addition of data from the Chrome User Experience Report, however, it now also provides a distribution of real-world load times, but only for First Contentful Paint and DOMContentLoaded, and I think that's highly indicative.

The worst offender is usually the time to first byte (TTFB; also included in the DOMContentLoaded metric, of course), and we know from the Search Console's "Time spent downloading a page" graph that Googlebot keeps track of (roughly) that. The data from Chrome will still be the most detailed and reliable, though, and I'd be surprised if they weren't using it. It's all about snappiness; get your content on the user's screen ASAP.

I am speaking specifically to Google's ranking factor regarding the speed of a page loading.

I give more merit to what Google reports [developers.google.com] than someone's interpretaion of how a page loads.

On this page :

[webmasters.googleblog.com...]

Google says :

Today we’re announcing that starting in July 2018, page speed will be a ranking factor for mobile searches.

The “Speed Update,” as we’re calling it, will only affect pages that deliver the slowest experience to users and will only affect a small percentage of queries.

Here they talk about "user experience" and that only sites which are really-really slow would be impacted.

a slow page may still rank highly if it has great, relevant content.

And that relevancy of the content will always outperform page speed by far.

Although there is no tool that directly indicates whether a page is affected by this new ranking factor,

So this is the proof that Google is not using PageSpeed Insights, since they mention there are no tool for this.

Then they list 3 tools to help appreciate the page's performance. So PageSpeed Inights is not the only tool, and it comes only in third position.

How do you define a really really slow website?

@TravisDGarrett - I only said I trust Google's page speed results over the assumptions & opinions of what people think. I never said Pagespeed Insights was the only tool used.

Again... believe whatever you like. This is only becoming a pointless argument and not contributing to the OP.

Pagespeed insights + ThinkWithGoogle Speed Test is probably the best combination. GTMetrix/WebPageTool is great for monitoring speed from Geo location.

According to pagespeed, I should Leverage browser caching on AdSense (https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js) :D

And optimize CSS delivery of Google Fonts. If we optimize these two, the score would be 100 :D

Slow response time plus slow loading AdSense makes the site really really slow. While I can ignore AdSense, but the server response time is never stable, there are abnormal spikes at certain hours on GA and Pingdom.

So this is the proof that Google is not using PageSpeed Insights, since they mention there are no tool for this.

Well, not really. They're only saying there's no tool that will indicate if you're affected by the ranking change. But the Chrome User Experience Report will give you a pretty good idea of how you're doing.

I give more merit to what Google reports [developers.google.com] than someone's interpretaion of how a page loads.

That's fine, but can you be more specific? The link just goes to developers.google.com.

the server response time is never stable

Are you on shared hosting? Oversold servers or noisy neighbors can make your site unstable.