| 11:20 pm on Dec 29, 2009 (gmt 0)|
I'm also seeing Dec 22 quite commonly. This is a new Google Labs feature - just introduced in Novemeber 2009 [webmasterworld.com], so I don't think we can say there's any pattern for updates that we can count on. In fact, that's true for all of Webmaster Tools, not just the site speed data.
| 12:35 am on Dec 30, 2009 (gmt 0)|
I was getting updates to my main site performance almost daily until December 22nd and then haven't seen an update since. What sucks about this is that I was in the middle of a major effort to tune and speed up my site and I made major changes on Dec. 22nd and I really want to see how much of an improvement they made.
I have noticed that one of my slower sites only gets updates on a very occasional basis so maybe Google only gives us an update when it has enough data to compile averages from. Christmas is a very slow period (at least for my site) and the speed tests depend upon the Google Toolbar so maybe Google just hasn't collected enough data for us over the past week or so.
| 4:32 am on Dec 30, 2009 (gmt 0)|
|I was in the middle of a major effort to tune and speed up my site |
Same thing here. I used Page Speed and was very curious to see the effect.
We'll see it - it's just that kiddie sensation that you want it now. :)
| 1:48 pm on Dec 30, 2009 (gmt 0)|
I don't think the average Page Speed shown for one of my sites is correct. It currently shows 2.3 seconds, when in fact the true load speed for the the most visited pages is less than 0.4 seconds even on a dial-up connection. These are static html pages with no scripts and no more than one small image of 15 kb or less. The total kb for all of them is less than 30 kb each. I've seen no problems with the server recently, and other information indicates that the pages consistently load in 0.4 seconds or less.
But Google's historical Page Speed graph shows load times ranging between about 2.0 and 3.0 seconds. This can't be right.
Also there is a statement that appears to indicate that any site not in the fastest 20% is classified as "slow". Very puzzling.
| 2:29 pm on Dec 30, 2009 (gmt 0)|
Maybe the page speed calculation was wrong and this is why they stopped showing us updates. Maybe they are working on fixing the calculations.
Even if the calculations were wrong, they were a relative bench mark. I could at least tell if changes I was making was having a positive impact on page loading times. Without the updates, I don't know how much impact the changes I made had on overall performance.
| 11:05 pm on Dec 30, 2009 (gmt 0)|
Today it says "updated on Dec 23, 2009"
| 2:10 am on Dec 31, 2009 (gmt 0)|
I just got an update for Dec 25th. Maybe they are catching up.
| 2:28 am on Dec 31, 2009 (gmt 0)|
Mine says Dec 26th with 2.2 seconds to load.
| 6:31 am on Dec 31, 2009 (gmt 0)|
Well, there you go.. Did all the recommendations by pagespeed I possibly could do and now the site has even slowed down(!) in site performance.
I guess thats reason it is still sitting in labs..
I think it might work as a long term tool, short term no. It fluctuates wildly and the stats lag. Not good for real time measuring and tuning.
| 6:39 am on Dec 31, 2009 (gmt 0)|
|Maybe they are catching up. |
Obviously. Same site now says Dec 28. No need to track dates anymore I guess.
What gets me is that the same site in a span from Aug through Dec (per Google's graph) is showing performance from being in the green area (fast one) to up to over 25 seconds.
Just two days ago, this particular site has been optimized by using suggestions from Page Speed and it is getting the overall green checkmark (nice job per that tool).
My thinking is that the next update should show the improvement.
Anyhow, going back to huge discrepancy - will people start blaming their hosts at some point?
I already have Google showing few seconds for all of my PHP redirects. At this moment I don't know if that would be PHP by its nature or slow server issue. Have to do more research.
What I know is that many people tend to blame something else (like hosting company), rather then their code.
| 7:04 am on Dec 31, 2009 (gmt 0)|
Since my earlier reply mine has changed. It's now dated Dec 28th with a load speed of 1.9 seconds.
| 8:15 am on Dec 31, 2009 (gmt 0)|
"What I know is that many people tend to blame something else (like hosting company), rather then their code."
Could be. I have a timer running with my php code and it shows consistently page build times in the milliseconds by the php engine of the webserver. So I know for sure that the page is fast enough but external factors are slowing it down.
| 3:01 pm on Dec 31, 2009 (gmt 0)|
To this time also affect connections to external servers like ads, external scripts, statistics systems...
I think it is wrong to mark all these sites as slow considering all these factors.
Thanks and have a very good new year.
| 4:03 pm on Dec 31, 2009 (gmt 0)|
Page Speed in WMT is based upon the load time of visitors to your site. So it takes in to consideration people with slow connections as well as very fast connections.
|Page load time is the total time from the moment the user clicks on a link to your page until the time the entire page is loaded and displayed in a browser. It is collected directly from users who have installed the Google Toolbar and have enabled the optional PageRank feature. |
| 1:52 pm on Jan 1, 2010 (gmt 0)|
Based on what i'm seeing there seem to be (or has been) a lag in the data that's reported by a little more than a week, as well as daily updates...
| 8:49 pm on Jan 1, 2010 (gmt 0)|
I suspect much of the lag I'm seeing from when I implement a design change until it is reflected in WMT is caused by various web caches. Not much can be done about this but wait it out. The figures could also be a rolling average that consists of several days worth of data. I say this because I parallized my images and css files to a new cookieless domain and now WMT is complaining about me serving those objects from two sources even though I know they are only being referenced to the new domain from all my code.