Does Site Speed Affect Rankings?

The short answer is yes, but let’s have a look at the reasons why

Back in 2010, Google’s Matt Cutts announced that the search giant had altered its algorithms, in order to account for the loading times of websites. At the time, he said that, as just a tiny part of Google’s 200+ elements to the algorithms, speed didn’t account for very much, but does this still apply?

It does, although I’ll go into how much and why a little later on. At the moment, the popularity of responsive design sites are adding to the issue of slow-to-load sites, as many people, including designers, don’t realise that they tend to load a full desktop site, whatever the target device.

This is something that a competent developer can easily overcome though and there are a variety of techniques that can be employed to speed things up. These include conditional loading, using media queries and reducing image size and scripts, to name a few.

The best user experience

The reason that site speed is an issue at all is that 40% of users will abandon a site that takes more than 3 seconds to load. This means that not is the site then considered not as useful to visitors as it could be, but potential customers are also leaving and finding the competition.

Due to the time in between Google’s announcement, all of the changes to the search engine’s metrics that we’ve seen over the past couple of years, and the rise of responsive design sites, many web professionals are asking just how much relevance does speed have in 2013.

In an attempt to answer this question, Matt Peters, a Moz data scientist, recently teamed up with Zoompf.

The study

In order to begin researching the issue, Matt and Zoompf came up with a list of 200 search queries from the 2013 Ranking Factors Study. They then selected a “representative sample of queries with between 1 and 5 search terms and extracted the top 50 result URLs for each query.” This resulted in a list of 100,000 pages in total which could then be used for evaluation purposes.

Using Google Chrome, with an empty cache, and using the WebPageTest open source tool they “we launched 30 Amazon "small" EC2 instances running in the Northern Virginia cloud”.

The results

For each URL, the study collected more than 40 different page metrics, most of which appeared to have no significant impact on ranking. Looking at both document load and rendering times, as Google isn’t clear on what page load time actually means, the data showed:

“[N]o clear correlation between document complete or fully rendered times with search engine rank”.

“We would expect this graph to have a clear "up and to the right" trend, as highly ranked pages should have a lower document complete or fully rendered time. Indeed, page rendering has a proven link to user satisfaction and sales conversions […] but surprisingly we could not find a clear correlation to ranking in this case.”

However, the researchers didn’t stop there and the results, although surprising, are most likely to be connected with how Google’s metrics work and on what aspects, rather than that page speed doesn’t impact rankings.

Time to first byte

The next task that the researchers carried out was to look at ‘Time to First Byte’ (TTFB), which looks at how long it takes a browser to receive the first byte of data. This means that it looks at:

• Network latency with regard to sending query to specific URL
• Amount of time server spends processing the query
• Amount of time it takes to send first byte from server to browser

This time “a clear correlation was identified between decreasing search rank and increasing time to first byte”.

In order words, sites with a lower TTFB respond faster and as such, return higher ranking in search results. This, they found, was the “strongest correlation effect” which means that it’s likely to have an effect on search ranking.

Page size and images

Whilst there was a correlation between page size and ranking, the researchers pointed out that they can’t prove their theory that businesses with low page sizes didn’t necessarily rank well compared with those that do. This is because, they think, that larger brand sites have the resources and funds to fully optimize a site, whilst SMEs don’t.

However, with regard to images, there were no surprises, although the results did suggest that these are not overly important to ranking in the great scheme of things. Overall, it was found that page load time is not currently affecting search ranking.

Does this mean I shouldn’t worry about load times?

Not really, in terms of good practice and business sense, it’s wise to have your site as speedy as possible as, regardless of how it affects rankings; conversions depend quite heavily on how quickly pages load.

Additionally, although the study showed no direct link with page sizes and load times, the correlation between higher ranking and TTFB shows that there is still a link. Larger sites for organisations were more likely to be ranked higher using this metric, as they tend to have dedicated servers and a back-end infrastructure.

This means that, contrary to popular belief, the back-end infrastructure of a site is potentially more important than the front-end when it comes to performance.

Many small businesses rely on shared web hosting for their sites, so perhaps spending a little more on a virtual dedicated server (VDS) would help this issue to be addressed for the SME.

Despite this, the research team said that they “suspect over time, though, that page rendering time will also factor into rankings due to the high indication of the importance of user experience.”

Improving performance

For the SME, this is something that needs to be addressed with their webmaster and hosting company. It’s important to ensure that not only is the back-end speedy due to the use of dedicated servers, but network connections and database servers are also taken into account.

This means:

• Optmizing application code
• Optimizing database queries
• Ensuring web servers are fast and responsive

To look at how a site tests for TTFB when looking at your own and competitor sites, a tool such as the aforementioned WebPageTest is necessary.

The results are very interesting as many people believe that it’s load times that affect the rankings with regard to un-optimized images and documents. However, the speed of the site is measured before we even get to that stage, so it’s important for businesses to take steps to look at TTFB times.

As mentioned, there are things that can be done about it and the basis of this will be in using cloud servers. This levels the playing field a little, as cloud tends to, in terms of accessing high quality IT services, without costing the earth.

Cloud application services are becoming increasing important for this reason. The small business no longer has a reason to be ranked lower for not having a sophisticated infrastructure, as this can be addressed easily with a VDS and some further optimization.