BLOG ARTICLE
September 28, 2018

How slow is your website? (part 2 – eCommerce Conversion)

This is part 2 of 3 part series of articles about how site speed can affect conversion rates, search traffic, and mobile experiences for eCommerce sites, and what you and your developer can do about it.

If you missed part 1 – it’s here: Why eCommerce owners should care about site speed

So is your site slow?

Actually this depends on a number of different factors including:

  • How competitive is the specific market for your online store?
  • What are your eCommerce traffic sources? For example site speed will be more critical for your end-of-funnel traffic who are looking to transact.
  • What are the pages that you are testing (e.g. product pages / shopping cart pages)?

A good starting point is to to benchmark pages-types against your competitors

Why? Because different markets have different tolerances to site speed. Even in one market, there will be big differences between segments. In fashion retail, visitors to a unique boutique may be OK waiting for high quality images of beautiful people and products to load. But what about visitors to a fashion retailer selling multiple brands? How long would you wait on ASOS for a page to load to buy Nike trainers, when you know you can get them at The Iconic or Nike Store, in 2 clicks?

How to benchmark your site

I’m assuming you have developer/maintenance person you are going to talk to about speed issues. So the point of benchmarking is to empower you to have sensible discussions with them about how the site is performing. Different tools test different aspects of the your online store page load (server, front-end etc). You don’t need to understand these too much – your developer will be able to dig in further. These are the leading tools:

  1. Google PageSpeed Insights – used by Google as one indicator in its ranking algorithm, so a good starting point. This tool doesn’t show site speed in seconds, but indicates whether it considers speed optimisations have been done. https://developers.google.com/speed/pagespeed/insights/
  2. Pingdom – gives a reasonable estimate of real world page experience times. You can set your tests to run from different locations including Melbourne (unlike most free tools which run from US/Europe). https://tools.pingdom.com/
  3. GTMetrix – combines Google PageSpeed and another speed tool called YSlow. The tools also gives an indicative time in seconds for a fully loaded page. Fully loaded means the page has fully loaded including all content below the fold. So this data is useful for your developer but does NOT represent real world experience, which is when a users can see and use content above the fold. (It also tests speed from Canada only on free version) – https://gtmetrix.com/
  4. WebPageSpeedTest – Detailed data your developer/maintenance person can review. You can also run from various location, although we experienced problems on some pages which block the tool (403 error) http://www.webpagetest.org/

For testing real experiences that may impact the return on investment on your Paid Search, I suggest you first do a Google search for an exact product your online shop sells, so you can test the relative experience of a visitor with intent to purchase. For example we searched for AirMax shoes (see above) and then picked two eCommerce product pages to compare.

I suggest you create a table like this and share it with your developer:

SitePage Speed (Desktop)PageSpeed (Mobile)GTMetrixWebPage Test.orgPingdom

ASOS.com.au

http://www.asos.com/au/nike/nike-air-max-90-essential-sneakers-in-black/prd/7704704

64/10052/100

Fully loaded 17.1 sec (Can)

 

Page size 1.83mb

 

236 requests

12.45 sec (doc loaded)

3.07 sec (Melb)

Page size

1.7mb

 

238 requests

Nike Official http://store.nike.com/au/en_gb/pd/air-max-90-essential-shoe/pid-10064377/pgid-1110094971/10064/100

Fully loaded 14.2 sec (Can)

 

Page size

1.51Mb

 

149 requests

Blocked

(doc loaded)

5.5 sec (Melb)

Page size

1.4mb

 

154 requests

Things to think about when benchmarking

  1. One of the things you will notice is the impact of location where tests are performed from. So while NIKE looks better on GTMetrix, tests run on default server from Canada, the Pingdom result from Melbourne gives a 2 second faster load time to ASOS.
  2. These tests are done against a random landing page only. To understand your customer’s journey better you also need to consider the site speed for the next steps, especially the Shopping Cart and Checkout. You can get more data about your client’s experience  from your Google Analytics.
  3. Remember the purpose of this is not to fully diagnose the problems – most of these will be fairly technical, but to give yourself an idea of how you stand in the marketplace, and no single tool gives you a clear picture.

Testing for real experience (real user monitoring)

Once you embark on improving your eCommerce Store performance, you can use more accurate tools to test your improvements, but also to assist with analysis of where problems might lie. Real User Monitoring is an approach to web monitoring that analyses all (or a sample) of visits to your website, and requires installing some code on your site.

One of the key tools you probably already have at your disposal is Google Analytics. I’ll talk more about how you can use Google Analytics to help diagnose speed problems that may be impacting your sales, in the 3rd article in this series.

Next

In the next article I’ll provide we’ll look at specific action you can take to improve your website performance to get higher conversions.

envelope-o