The Connection Between CDNs and Conversions (Does It Exist?)
August 28, 2015 | Robert Gibb
Attracting users to your website is one thing, but turning them into customers is a science in and of itself.
Converting users into revenue depends on a variety of factors from your marketing strategy to your website’s layout. And while each of these factors plays a critical role in generating conversions, two of the most overlooked contributors to conversions is page load time (PLT) and the CDN that helps improve PLT.
In this post we’ll provide evidence that load time matters, citing real world examples and research by respected companies like Intuit. We’ll then challenge you to implement MaxCDN – free of charge – and see if an improvement in PLT caused by MaxCDN leads to an increase in conversions.
How Page Load Time Impacts Conversions
Generally speaking, an improvement in PLT results in an increase in conversions (subscriptions, purchases, downloads, etc.). This trend isn’t always linear though.
Case Study: Intuit
Intuit discovered a stair-step pattern where the rate of increase in per-second conversion rates fell into three distinct ranges. This stair-step pattern is based on research that Intuit gathered while taking on the bold mission of decreasing PLT from 15 to 2 seconds across its portfolio of websites (TurboTax, Quicken and Mint, to name a few).
Here are the findings Intuit presented at the Velocity 2013 conference:
Information source: Moz
- +3% conversions for every second reduced from 15 to 7s
- +2% conversions for every second reduced from 7 to 5s
- +1% conversions for every second reduced from 4 to 2s
While this graph relates specifically to Intuit, below you’ll see that PLT affects conversions similarly for other business types. Whether people are loading a web page, downloading an app, buffering a movie, or waiting for a pre-roll video ad, they want the content now – not a few seconds from now. This fact is universal across all business verticals.
The benefits of improved PLT aren’t limited to sites owned by Intuit that market and sell software. They extend across many different business verticals. Here are a few more examples:
- Mozilla increased its volume of Firefox downloads by 60 million per year simply by getting their pages to load 2 seconds faster.
- The Obama campaign pushed a complete redesign of its platform for online donations, leading to a 60% increase in speed and a cool $34 million in additional donations. By optimizing their frontend code and moving their static content to a CDN, the platform was able to handle donation surges as high as $3 million per hour.
- Walmart discovered that for every 100ms increase in site speed, their revenue increased by 1%. With every second of reduced load time, their conversion rate increased by 2%. This goes to show that you don’t need a huge increase in performance to see a noticeable boost in sales.
- Almost half of users expect websites to load in two seconds or less (Akamai)
- More than half of users abandoned websites that took more than 3 seconds to load (Akamai)
- Converted shoppers receive pages that loaded 2x faster than non-converted shoppers (Web Performance Today)
- A one second delay in response can result in a 7% decrease in conversions (KISSmetrics)
- Conversions increase 74% when PLT decreases from 8 to 2 seconds (ConversionXL)
- 51% of online shoppers in the US will not complete a purchase if the site is too slow (Radware)
How CDNs Fit In
Everyone knows that CDNs improve PLT by caching content on servers closer to end users, but there’s hardly any documented proof of this. There’s also no documentation of the direct impact CDNs have on conversions.
In a previous post we showed you how MaxCDN alone improved page load times by 15% and assisted other web optimizations to cut PLT overall from 4.5 to 2 seconds. But we failed to measure its impact on conversions. Also, much has changed as far as web optimization best practices are concerned since that post was published.
For instance, parallelizing downloads across hostnames (or domain sharding) is no longer a best practice with HTTP/2. Additionally, this test focuses on a Magento ecommerce store (one that no longer exists!) rather than popular web services offered by advertisers, publishers, software distributors, and game developers – our largest clients.
This is why we’re calling on you.
Take the CDN-Conversion Challenge
While the numbers related to conversions and page load time (PLT) mentioned above are true, they are also business-specific. The only way to know how a CDN impacts your page load time and conversion rate is to implement a CDN and measure the results.
You can do this for free with MaxCDN.
Here’s how we recommend testing a CDN’s impact on PLT and conversions:
- Measure the PLT of a popular page on your website (one with static assets like images) using Pingdom.
- Measure the conversion rate of the same page for a given week (the definition of a conversion will vary from business to business).
- Sign up for a free test account with MaxCDN.
- Schedule an integration call with MaxCDN.
- Measure the page’s new PLT after MaxCDN has been delivering content for a few days.
- Measure the page’s new conversion rate for that same period of time.
- Submit results to me at email@example.com and I’ll include them in the chart below. (You can also email me with any questions.)
|Date||Web Service||Web Service Type||PLT (without a CDN)||PLT (with MaxCDN)||Conversion Type||Conversion Increase|
|9/23/15||RebatesZone||Coupons||3.23 seconds||1.12 seconds||Coupon Clicks||433%|