When you launch a website, there are many factors that must be optimised, like checkout, but speed is the one you should really focus on. At the end of the day, if your website is optimised on all key conversion pages but your website is slow then it is wasted effort, believe me. Can the speed of your website really have that much of an effect on your conversion rate? Even if your site isn’t loading too slowly, can it still be improved?

According to surveys done by Akamai and Gomez.com, nearly half of web users expect a site to load in 2 seconds or less, and they tend to abandon a site that isn’t loaded within 3 seconds. 79% of web shoppers who have trouble with web site performance say they won’t return to the site to buy again and around 44% of them would tell a friend if they had a poor experience shopping online.

This means you’re not just losing conversions from visitors currently on your site, but that loss is magnified to their friends and colleagues as well. The end result –lots of potential sales down the drain because of a few seconds difference.

How fast should be my website?

This is a question I get asked a lot. I have naively always thrived for the magical number of 2s but let me put it out there: ‘Not possible’! I know the stats and the graphs but for a retailer this number is fantasy. So what should it be?

Well in my opinion, you should do two things. Firstly, make sure that you are faster than your previous year. if you build a new website, it is not to be slower. There is nothing wrong to compare YoY. However, remember that it might not be LFL because you have changed the amount of content on a specific web page (i.e. larger images on your product detail page). Secondly, you should benchmark your key competitors. At the end of the day if you are faster than them then you can gain a competitive advantage.

Speed and google

Let me clear this one out. A slow website doesn’t currently harm your organic ranking. Will it change? The speed of your mobile pages currently doesn’t impact your mobile rankings, but soon it may, says Gary Illyes of Google. Read this article.

What tools to use?

There are plenty out there. For me, the best free ones are webtestpage and GA. I use them on a weekly basis and find them incredibly rich in insight. I know a lot of web developers use Lighthouse but for me it is rubbish. You can run the same tests 10 times in a row and you get huge variations.

Also, if you can afford it, find an excellent monitoring partner. I use NCC and they are fantastic. This is their area of expertise and they have some excellent tools. One that I am currently testing is their RUM tool that can predict how much cash you are loosing or potentially could gain if you were going to increase speed.

How to measure speed?

Speed Index. Yes you need to monitor your speed index on a monthly basis if not weekly if you have just replatformed. Read more.

There are all the KPIs you must look at:

• First BYTE – The First Byte time is the time from when the user started navigating to the page until the first bit of the server response arrived.  The bulk of this time is usually referred to the “back-end time” and is the amount of time the server spent building the page for the user.
• Render start – The Start Render time is the first point in time that something was displayed to the screen.  Before this point in time the user was staring at a blank page.  This does not necessarily mean the user saw the page content, it could just be something as simple as a background color but it is the first indication of something happening for the user.
• DOM load – DOM ready means that all the HTML has been received and parsed by the browser into the DOM tree which can now be manipulated. It occurs before the page has been fully rendered (as external resources may have not yet fully downloaded – including images, CSS, JavaScript and any other linked resources)
• Visually complete – Visually Complete measures how long it takes to display the content visible in the user’s browser: content “below the fold” and non-visual content (like third-party tracking beacons) is excluded.
• Fully loaded – The metrics grouped together under the Fully Loaded heading are the metrics collected up until there was 2 seconds of no network activity after Document Complete.  This will usually include any activity that is triggered by javascript after the main page loads
• Page weight – this is how heavy your page is with all the content

and then leave all other metrics to your web developers. Optimise all of the KPIs above and your CR will start to improve. How much will CR improve depends on how bad your website is.

Also remember to tests in different environments. For example, how fast is your website in a slow 3G environment vs. a 5mbps WIFI? Also adjust latency.

Conclusion

Ignore website speed at your own peril as the CR gains are potentially huge. You haven’t got time then appoint a partner. If you have invested in a Ferrari but it is as fast as a Skoda then you are burning cash away! Monitoring takes no more than an hour when you select key pages (i.e. homepage, category landing page, product detail page) and will give you a great set of focus with your web developers. Good luck.

Benoit Mercier