Trak.in is a popular Indian Business, Technology, Mobile & Startup blog featuring trending News, views and analytical take on Technology, Business, Finance, Telecom, Mobile, startups & Social Media Space

Speeding up the web-2: site speed and why hosting in India becomes a no-no henceforth

8

Google is apparently quite keen on making the web faster. First there was the announcement of SPDY protocol and now there is some talk about how the underlying algorithm of Google has been changed,under the hood,  in caffeine, to bring you faster and better search results.

Google’s Matt Cutts recently gave an interview to Mike of WebProNews, and in that, he discussed, amongst other things,  an internal debate going on in Google about whether site speed should be factored in into the search rankings that Google uses to display search results.

Apart form the famous page ranking algorithm, Google currently uses about 200 factors to determine the order in which the results are displayed for a query term. The new idea is to use the time it takes to load a site as also a factor in the rankings, such that faster loading pages are pushed higher up the rankings while slower loading pages penalised.

The intentions are noble- to make the web faster and to force webmasters to take a hard and fast look at the design and structure of their websites. But I foresee practical problems.

First there should be some benchmark of how to measure site speed independent of the measuring location and the hosting location. Else if the loading speed of a site hosted in India is compared to loading time of a site hosted in US, where the test runs in US, then perhaps the site form US would have an unfair advantage.

I am concerned, because for ex, the site Cutts mentions WebPagetest, gives only US,UK or NZ  locations as test locations; this essentially means that sites hosted in India are at a disadvantage.

Even if the testing is done locally in India, India’s poor Internet infrastructure may hinder the page loadings experienced by users and thus skew the results.

Does it mean Indian websites have no choice but to host their sites in the US if they want to enjoy the new SEO dynamics of increased site speed based page ranking?  If Google implemented this, would you host your site in US to optimize site speed based SEO, even if your target audience is Indians, or would you still host it in India? Do let us know what you feel about this issue via comments!

  1. arun dudee says

    One is response time and one is load time and response time will be ALMOST same independent of location you doing testing and i think Google is quite smart , they will rely on response time and not load time

    1. Sandy says

      Arun,
      Would you elaborate how you are going to calculate the response time independent of the location? how do u distinguish latency from load time ? If you juts foucs on no. of bytes transferred and no. of requests across which this is spread, you are being too simplistic. Perhaps then , it wont really matter if I host with a good dedicated host with good connectivity to the internet or a shared host with poor connectivity, because my latency is irrelevant! or perhaps there would be no advantage someone would get from having association with CDNs and being able to serve content locally. Knowing the response time of a server independent of client and server locations is very tricky and I wish you/Google best of luck if you plan to go that way! In any case it wont reflect the real page loading time or user experience, something Google purportedly wants to improve! I would still maintain that it is nothing but a ploy to get webmasters to optimize their sites for chrome.

  2. Venky says

    Relevance is more important. What if some “not so relevant” site comes up in the search results, just because it is able to respond faster? And the sites which we are looking for end up in page 2 or 3 of the results? In most cases, we look for results in the first page. If we don’t find any, we will rather change the search term than go to the second page of results.

    1. Sandy says

      Venky,
      The fact that people only click through the first search page is a very important fact you have brought up; if pages that are relevant to the query term loose their ranking due to site speed, then it would definitely be a step backwards to the user experience , rather than a step forward.

      As they say, sometimes it is not important that you go fast,its equally important what direction you are going in. If efficiency (speed- how you are searching the web) wins over rationale ( relevance- why you are searching the web), it would be akin to difference between ‘doing the things in a right manner’ and ‘doing the right things in the first place’. Efficiency(managerial goal of doing things the right way) should never in my opinion win over the leadership/visionary goal of doing the right things in the first place. If one makes the web faster , but it is a web that is no more relevant and searchable as per relevance, perhaps one is missing the point. Time for Google to take heed.

  3. Madhav Shivpuri says

    Loading speed is an interesting parameter to consider for a website’s pagerank. Would you then think which browser is used? For example, my Mozilla Firefox loads faster than IE; and Chrome loads faster than Firefox. How would you pagerank it?

    Would you consider the number of RSS subscribers, email subscribers, retweets, number of comments on a blog post etc., from a site to rank it? In fact these parameters might mitigate the server location bias. What say?

    1. Sandy says

      The choice of browser for speed test is an important point you bring up. Assuming that Google performs site loading test on Chrome and uses that as a result, that would mean that all webmasters would have to optimize their sites for loading in chrome and not in IE/Firefox. That adds another dimension to the browser wars. Perhaps with sites optimized for chrome, people will start using chrome more and Google wins on another angle.

      I’me sure that the simplistic page rank algorithm needs to be supplanted with the other metrics like RSS feed subscribers etc that you mention; I am afraid some of these metrics may already be used in ranking results…but all these factors would not mitigate the effect that location based site speed ranking would have as that is an independent dimension and poor infrastructure countries will continue suffering a disadvantage of site speed is introduced as a ranking factor.

  4. Rakesh Waghela says

    I am little disturbed about the view point about the Faster Loading Web Page.

    The benchmark may not be the BANDWIDTH as you are describing it !

    Perhaps the better way to Estimate the site loading time is to analyse the amount of objects a page contains ! I.E, Number of images , flash objects ,applets !

    Summing Up Your Voice :-
    Page loading time is always RELATIVE and UNFAIR aspect to look as far as Search RANKING is concern !

    Alternative : Let Google Keep the SEO Ranking Intact with their Results unchanged, but every result could be indicated with the HEAVINESS of the page.
    May be with a graphical icon or some tag next to result snippet !

    1. Sandy says

      Rakesh,
      That is a very nice metric you have brought up- but even if that metric of how many objects are present in the page is used, I would still like Google to go by your suggestion of keeping the search ranking intact, and maybe displaying a logo or graphic alongside the search result to indicate the site load speed. That way leave the decision to the user as to whether he wants to visit a slow loading website or not, but don’t penalize the site for the site load speed by lowering its ranking. Displaying the site speed while not factoring it into search ranking is the middle path that Google should walk. Lets hope they read your suggestion and pay heed to it!

Leave A Reply

Your email address will not be published.

who's online