That didn't take long, and it's a good idea.
Adwords blog post on load time [adwords.blogspot.com]
To clarify this from the page on how load time is evaluated [adwords.google.com], though:
Is it now just the HTML content that's being evaluated, with the other components mentioned not being factored in yet?
Except in extreme circumstances is this really warranted? Really?
It's 2008, large segments of the population have broadband and those that don't are familiar with the load time of an average page.
If it were 2001 this would make perfect sense...but at this stage of the game does it really make sense for webmasters to worry about the load time of their pages (unless there is a drastic amount of load time that is out of the ordinary).
The last time I paid conscious attention to the load time of a page was probably 4-5 years ago...it feels like a step backwards to worry about getting all those gif images to be 2kb less just to please the adbot.
I feel this a change that no one really asked for...advertisers with obscene load times will be pushed out anyway as customers bounce instead of waiting for the page to load.
I suspect this has nothing to do with broadband speed, but rather the design of the target-page, in most cases a shop system, I assume.
In general, webmasterworld-readers have always paid considerable attention to this very, very important issue loading-speed. But this is definitely not the case for many online-shop-owners, who probably paid a hell of a lot of money for a dubious system with loading times of 10 seconds and more on SERVER side. Often an important reason why they had to use adwords at all;)
I'm relatively new to adwords: Does this mean, that my fast-loading page will be placed higher than a lame competitor's, who bids an equivalent sum? Does quality outweigh money with respect to placement?
Landing-page loading time should be kept short to avoid annoying users, regardless of whether they happen to be coming via AdWords.
It's thoughtless to so load up a page with so much crud, often in the form of a large number of graphics, scripts and CSS files to fetch on that first hit, that it takes an age to display on broadband, dial-up or mobile where the user may be paying by the byte and where the CPU is less powerful.
I don't always succeed but I try to keep my page loading times under 100ms if possible and certainly under a second or two except in extreme cases.
So I think this is a GoodThing(TM) for user experience.
[edited by: DamonHD at 11:18 am (utc) on May 9, 2008]
BTW, is there some way of ensuring that this information is readily available for "Conversion Optimiser" campaigns, which most of mine are, since there is no magnifying-glass link to the keyword analysis stats and the Quality Score column text does not mention load time...?
also, as with the other diagnostics, it would be usefull to be able to pull a report on this information, rather that having to hover over every keyword
So will QS actually say "Poor landing page response" or will it be another "BS" reason to jack up may rates under the guise of "slow response" or whatever email hack of the day google wants to use as a response?
Until QS spells out what quality is, QS = BS.
How can they know where the bottleneck lies? ... what if the load time is being affected by isp or other hops along the way?
I think this is a good move, and not just for those people who still have slow connections. It might encourage webmasters to think harder before crowding pages with all sorts of moving images, sounds, or other clutter.
There are some industry news web sites I read regularly that are going to have to get busy. Thank you, Google.
>>>Except in extreme circumstances is this really warranted? Really?
Agreed. Seems like grasping at straws. While I can see that long load times might effect the overall user experience for some small segment of the population, it certainly has little to do with the quality of the content on the page. And I'd expect that the primary focus should be on quality.
This is likely such a small issue that it seems to have been a make work project/got a hammer looking for something to hit.
> This is likely such a small issue..
wheel, I'm sure you are doing a really good job in a highly competitive industry. And most of your competitors as well. You're doing so fine, that many of you probably lost contact as to how much crap is out there calling itself a "website."
No, this is not a small issue.
I recall another posting a while ago, I think it was from googleguy. He was complaining about parsing - problems and similar issues. More than 90% of the html out there is completely rotten, and even google's own pages do not pass the w3c-validator, though this doesn't really mean anything.
Loading speed and parseability are two really important issues, not only for the searchers. And I'd bet there are strong statistical correlations between the quality of these two on the one hand, and "quality of the content on the page" on the other.
Well I think it's mostly a good idea; sometimes it's hard to accurately judge the loading times of some of my sites and my client sites, since they're sitting on my own network. Most of the slowness seems to come from Google Analytics and/or AdSense; hope they're taking that into account. And also, tying it in with the keyword is a little clumsy, specially if you have thousands of products and thousands of keywords for them. I'd sooner see it attached to the ad.
I am curious how to Google would determine the page load if the landing page doesn't have any callbacks to Google (adsense, analytics, conversion tracking, etc..)? And if this is the only way to determine it, what is to prevent webmasters from putting all the calls to Google at the top of the code so that they are loaded first?
Also, there are a couple ways to measure load times (which can measure different metrics). The two most common are: first byte received and last byte received.
Are we to assume it is last byte received so that it will take page size and bandwidth (more so) into consideration?
Hmm, well, a quick review shows that all my sites say they're faster than average for the geographical area (!) and one just says the landing page speed is great (nothing about about the geo area) So I guess there are levels, but so far I've only come across two.
Pretty good. I had always worked on making the websites faster, good that it will be counted :). Yahoo's page by Steve Souders is good, and now Steve is with Google.
The question is do I think they’ll implement these load time penalties against those monster sites like Amazon, Shopzilla, or a thousand other comparsion sites that Google receives a hefty cash flow from. Well actually no more than they saddle those sites with bogus penalties in the regular search as compared to everybody else.
The devil is in the details and I’ve heard a thousand things that sound good. In the end it just sounded good nothing else.
My opinion is Google will set an artificial low load time that probably can be implemented on the fly even quicker than increasing rates on keywords. In fact they know in many cases implementing cost increases on many accounts can take up to a month where this can be done with pinging every 5 minutes. Plus the system is less exploitable as far as revenue is concerned when Google assigns costs to one variable. The problem for the average webmaster is this load time figure is highly volatile based upon a hundred variables.
Also the only reason I can see for implementation of this system by Google is the load times may effect Google being credited for a click with slower BB and DU that “breakaway” during the load.
Again I don’t see Google implementing this policy against its buddies. These companies will quickly let Google employees know at their daily luncheons together that they design for broadband and its unfair for Google to implement this policy against them in particular. Bottom line behind the scenes the general rank and file, who have no voice, will be saddled with the costs.
I’d like to see an opinion from the FTC on this for a change. Other than that it sounds “hunky dory”, “jim dandy”, wonderful.
Think about it. Your page loading time affects quite a few down the pipeline. It also directly affects Google's resources as it is waiting to serve ads and your pages are taking forever to load. That's just one in the pipeline that is affected.
Its really cool to see this feature. I'm hoping it will wake a few up and let them know that those 500k+ pages may present some issues in the near future. ;)
You ever hit those sites from the SERPs? You'll sit there and wait 8, 10, 15 seconds and nothing. C'ya!
Buried away in the help files is this:
|The threshold for a 'slow-loading' landing page is the regional average plus three seconds. |
In case anyone was wondering...
|Until QS spells out what quality is, QS = BS. |
By far the pages with the longest load times are those dynamically-generated shopper sites that pull 20 products with affiliate links to make a "page".
Incidentally these are also the pages that right now ride high on Adwords and seem immune to any sort of quality score.
I doubt much will change.
But if you're not them- rates are going up again.
Google's been peeing on my leg for a year now and telling me it's raining.
Some of us don't mind waiting an extra second to see quality graphics.
Ebay loads fast but is has flea market quality.
|By far the pages with the longest load times are those dynamically-generated shopper sites that pull 20 products with affiliate links to make a "page" |
They should implement better systems in place to make it faster. Fast is better.
Landing page load time (according to Google)? Well mine may be different than what Google Displays.
AdWordsAdvisor I have an advice for Google. Just like the Keyword Suggestion Tool why not have a “Google Load Time Speed Tool”. This should clear all the confusion.
Landing page load time as part of quality score? I might have to rearrange many pages. Some of my pages contain 10 or more pictures.
> Some of us don't mind waiting an extra second to see quality graphics.
@54 Mb/sec I don't notice graphics loading delays any more.
For anyone who's curious, Google has already been tracking "Connection Speeds" in Google Analytics for years. Theoretically the only QS penalty should be against sites visited by the Dial-Up crowd whose pages load slowly (read: noticeably slowly).
> Some of my pages contain 10 or more pictures.
BrandNewDay, that won't matter. Each image produces a separate http-request on browser-side. The landing page itself will be loaded much quicker. Unless you have a very questionable page structure, the html-code will be loaded before your browser even sends those images-requests.
> The threshold for a 'slow-loading' landing page is the regional average plus three seconds.
OK. That would allow for more than meg of raw code, because I doubt google's hardware works with modem speed;) So the key issue is database-request-design, I suppose. And this can quickly become critical on shared hosting.
Let me put it this way: As soon as you're beginning to make money with your shop, you'll need a dedicated server. Any objections?
|defined as the amount of time it takes for a user to see the landing page after clicking an ad |
Is sounds as if they are focusing the measure on when a user can see the entire page, not a technical measure of when the html is loaded. I dont think Google will be fooled and will know when all the images have loaded for us to see.
Shouldn't there be a discount for fast load times?
| This 36 message thread spans 2 pages: 36 (  2 ) > > |