Site Performance After Hosting Upgrade

The last week of July was a very painful week for this website. Our previous hosting company was having a world of difficulty through an ongoing Denial of Service attack and, combined with a particularly alarming security breach, it was time to leave. Despite lengthy periods of outage and a slow ftp connection, I managed to migrate the site to a new host and get the DNS sorted by the 30th July 2010.

I thought I’d share some interesting data which should encourage us all to rethink our hosting from time to time.

Average time taken to load SEOgadget.co.uk in seconds (monthly)

average monthly performance of SEOgadget page load

Time spent downloading a page – note the outage

time spent downloading a page - Google

Webmaster tools performance overview

Webmaster tools Performance

Search engine visibility – Advanced Web Ranking

Search engine visibility - Google.co.uk

We monitor SEOgadget’s top 200 industry and traffic driving search terms on a daily basis. This chart shows the search engine rankings expressed as a percentage, which works a little like this (assuming one keyword):

Position 1 = 30/30 (100% Visibility)
Position 2 = 29/30 (96.6% Visibility)

Dedicated IP

Believe me when I say our old IP was hosting some bad websites (not the hosting company, I might add – they were just reselling). A bad neighbourhood would be putting it mildly. At the same time that server performance has increased, we’ve also moved to a dedicated IP address. That’s an important point to remember.

Does a significant improvement in page load in some way explain the improvement in search engine visibility?

Assuming all other factors remained the same (no significant link acquisition, no major algorithmic factors, etc), you could say that the improvement in site performance and / or IP change correlates to an improvement in the rankings. Was it the IP address, the performance, or both? Correlation is not causation, and I fully understand that. It’s just these charts show super compelling evidence to justify a proper study, in my opinion.

I can’t even begin to imagine the amount of data and analysis you’d need to prove / disprove a causative relationship between page load time and rankings. I wish someboady with a lot of time and interest the very best of luck.

Site Performance After Hosting Upgrade, 5.0 out of 5 based on 1 rating

Comments

  1. Nichola Stott

    Hi Rich – really interesting stuff and big thanks for sharing. We have recently seen a similar correlative relationship between site speed and aggregate rank, on a client site, though in this particular case other factors have been at play and it’s still early days. I have a question for you – your search engine visibility ranking graph above; which engines and markets inform the data? Is that just Google UK, or is there and US data in there. Would you be willing to break out and share if so?

    About a month ago I had a contact ask Maile Ohye about site speed related issues at SES Toronto (I couldn’t attend) and I’m sure at that time she said speed wasn’t a factor anywhere outside of the US. My memory might be failing me however.

  2. richardbaxterseo Post author

    Hi Nichola, nice to hear from you!

    The rankings are UK based, Google.co.uk only. Maile said the same thing at SMX Advanced Seattle in June. Not sure where that leaves us in August :-)

  3. Nichola Stott

    Thanks, that’s really, really interesting. In that case we have seen at least one correlative relationship between site speed and rank improvements in Google UK, during the month of August; following a fix that might have been compounding the site speed issue. In my case I am more inclined to think the fix is more directly relevant to the rank increases; however it is interesting that the two cases show really similar patterns, across a similar number of terms, at the same time! ;-)

    Oh.. and good news on getting the hosting issues resolved.

  4. Duncan

    Great post thanks. Sometimes proving with empirical data is not necessary – our brains are the most powerful analysers we use and this just seems to make sense. Why would Google not give more love to sites that performs well and has their hosting in order? I’m moving to a dedi server pretty soon and expect to see some nice boosts.

  5. Peter Bowyer

    Hi Richard,

    Very interesting to see – I have the first three graphs for my sites, but what do you use to monitor your “Search engine visibility – Advanced Web Ranking”? That’s data I’ve yet to find a way to reliably gather. As I’ve done a lot of optimisation for site speed and server performance, I’d be curious to see if it did the same for me :)

  6. Branko

    Nice. We have significantly improved the page load time on one of our sites recently, but have not seen any difference in traffic or rankings. We did see a significant difference in the number of daily spidered pages but nothing more than that.

  7. Allan Stewart

    Hi Richard,

    Glad to see the server move went well. I’ve always beleived that hosting is one of the most important factors in running a good website. As you know RB, there are many nice monitoring tools on the market one I like particularly is Keynote. Anyway, more interestingly, what is that monitoring tool that you use for Keyphrases? I very much would like to know?

    Kind Regards,

    Allan

  8. Tattoo

    Very interesting,though I’m not sure the move is worth it for the majority of websites. The ranking, traffic, and most of all roi increase is often not worth an impacting change

  9. Roger S

    We changed our server to use gzip – didn’t see any change in rankings and webmaster tools *still* tells us to speed the site up by gzipping the files!

  10. Matthew Brookes

    Hi Richard,

    was not going to post but thought i would let you know your site seems supper fast now!

    I think that if site speed is having an impact on user experience as well as what webmaster tools etc are tell you then a move is justified, even if you don’t get a significant boost in rankings / traffic you don’t want to annoy your current user base too much. Most moves can be managed effectively to minimise any negative impact as well.

    I have also seen some positive signs from Google when improvements to either the site code or hosting architecture have been made.

  11. richardbaxterseo Post author

    Hi Matthew, nice to hear from you – yes, the site is just slightly over 1 second on most pages. There’s a few tweaks on the back end and I’ve removed teh Quantcast tracking to try to speed things up. My target is sub-1 second to load.

    We’re working towards making the Kessel Run in less than twelve parsecs….

  12. Neal Dougan

    Very interesting Richard! You always come out with useful info like this. Will be looking into this myself as I am with Dreamhost. I pay about $15 a month and I host about 6 or 7 of my site on there. I haven’t really noticed slow download speeds or any dip in my rankings because of it. But it’s very interesting what you have found with your site and will be taking an interest in my websites performance from now on.

    Thanks

    Neal

  13. Friendly

    I completely agree with this post. I’ve recently upgraded my hosting and I instantly saw faster loading times. Can’t wait to see how it affects my rankings.

  14. WhiteSites

    If you website is running asp.net and was running on a Server 2003 box, then this doesn’t surprize me. The biggest problem with most hosting companies is they use a shared app pool, and then all the sites get to fight over the memory. I highly recommend you try to get on a server 2008 box where each site gets its own app pool by default. I now run my own server ( colocated at Internap in Houston ), I am still amazed by the number of shared hosts that still struggle to pull a 10 second page load time. Their system admins should hang their heads in shame.

  15. Seni Bana Yazmislar

    Very interesting Richard! You always come out with useful info like this. Will be looking into this myself as I am with Dreamhost. I pay about $15 a month and I host about 6 or 7 of my site on there. I haven’t really noticed slow download speeds or any dip in my rankings because of it. But it’s very interesting what you have found with your site and will be taking an interest in my websites performance from now on.
    Thanks I completely agree with this post. I’ve recently upgraded my hosting and I instantly saw faster loading times. Can’t wait to see how it affects my rankings……..