Page Load Latency VS Google Rank
After spending many hours looking at the crawl data from Google’s new Crawl Rate feature within the Webmaster Tools I believe that I have built a case to support my theory that latency of the crawl has a direct correlation upon rankings. I have recently made a few large changes on a few of my sites. 2 of my sites in particular have very similar content and are organized slightly differently. One specifically has just had a full makeover to it, actually making the pages lighter and more appealing to the users. But for some reason this new code takes a little longer to load increasing latency from an average of 300 ms to over 500 ms. My other site had an average latency in the high 300’s.
I have seen a cannibalization effect from one site to the other in a large scale motion. I could be crazy but I think this has an effect! I have another site that had dramatic ranking changes about a year ago; this site has maintained an average latency of 1000 ms or worse. Has anyone else seen this?
I have seen a cannibalization effect from one site to the other in a large scale motion. I could be crazy but I think this has an effect! I have another site that had dramatic ranking changes about a year ago; this site has maintained an average latency of 1000 ms or worse. Has anyone else seen this?
Labels: seo
by Aaron Shear
5 Comments:
Aaron, is the concept of page latency related to what Brett Tabke claimed many years ago about building light pages (under 10k)? Page latency actually seems more on target as you have described lighter pages taking longer to load... At the time, I think he was referring to server response and it was discussed that G could favor this due to savings on bandwidth, measured in nanoseconds...
What would you say is the reason for Google apparent liking of fast crawlable pages? Like in any murder, without a motivation we have no case...
Hey Bob,
I have been thinking about this for years, it makes sense to me to rank sites based upon their ability to respond. The main reason alone being accessibility, if you come across information that you are looking for in a search and it takes along time to appear you may decide to hit back and find something else. I think the same thing works for large sites that are ranking for highly searched terms.
For example if you have a keyword, lets say mortgages and there are hypothetically 100k searches a day for this term. Let’s assume a 40 percent click through rate, that’s a lot of load to put on a crappy web server. If Google see’s a high latency when it’s crawling it can assume that the user experience is not nearly as good as another site with a much lower rate. This is all purely hypothetical, the data I am looking at spans over 12 sites. Many of them have tough problems with cannibalizing each other. So when one site does better than the other I am always looking at the reasons why.
Just recently I had some extremely coincidental data to certain changes made to the site that affected latency. But of course I may be totally out of my mind. ;-)
Hi Aaron,
One of our analysts did a correlation piece for a few of our sites on a bunch of obvious and not-so-obvious ranking factors and we found that page load time had a strong positive correlation with ranking.
As you mentioned, a high page load time usually does indicate a poorer user experience, which would not be difficult to factor into G's secret sauce...
- shor
It's good to hear confirmation on these issues.
hello,
i have one question for all
my website on no.3 in google.co.uk but i had to input some images on homepage (around 110 kb). after input a image my web page size will be 150kb ..so my question is that
my webpage size will be more than 100k then its effect on my current ranking in google?
Post a Comment
<< Home