Navigation
Page speed problem in digital advertising

Page speed problem in digital advertising


Wojtek Andrzejczak
Wojtek Andrzejczak
Page speed problem in digital advertising

The fact that Page speed is critical it is a well-known fact. But what impact does it have for our digital advertising? How much revenue do we lose, and how much we could gain?

Page speed vs. Advertising

The website page size matters. But if we would need to have 100% optimize the website, we’d have to remove all images, JavaScript, and leave an only black text.

But we want to do business and advertise it with an in-depth analysis of user behavior and campaign traffic. So we need to put some JavaScipt, images, animations, and tracking tags/pixels. And this makes our website slower.

Question

If our business relies on digital advertising, how much money we could lose if our website is slow? And how much we can increase our revenue if we improve our website speed?

What the statistics say

How crucial page speed is, shows companies reports regarding this topic: 

  • Amazon found that each 100ms delay costs them 1% fewer sales.
  • Walmart: for every 1s quicker page load time, we experience 2% more conversions.
  • Google research: Page load time goes from one second to 10 seconds; the probability of a mobile site visitor bouncing increases by 123%.
  • Akamai research: every 100-millisecond delay in website load time can hurt conversion rates by 7%.
  • Pinterest increased search engine traffic and sign-ups by 15% when they reduced perceived wait times by 40%.
  • COOK increased conversions by 7%, decreased bounce rates by 7%, and increased pages per session by 10% when they reduced average page load time by 850 milliseconds.
  • The BBC found they lost an additional 10% of users for every additional second their site took to load.
  • DoubleClick by Google found that 53% of mobile site visits are abandoned if a page took longer than 3 seconds to load.

If the time to the interactive page takes more then 6-10s, you have a problem. When considering that the majority of our traffic comes from mobile devices, then the problem gets even more significant.

Drop-off rate problem

Landing page drop-off rate tells us how many users click on our ads, but did not land on our website.

Users can abandon if they accidentally click on the ad, or when page loading time takes to long.

Page speed vs. digital campaigns

I’ve seen campaign landing pages where accidentally someone has uploaded to big PNG files and whole website weight 12-18MB. Our campaign performance KPI (CTL – Click To Landingpage) has dropped to around 15%. After fixing problematic images, page size dropped to 2.3MB. Our CTL raised to 45-50% just after one day.

Page speed vs. Google Ads

Google Ads uses something called the quality score for our ads. The higher the score we have, then less we pay for the click, even up to 50%!

The quality score includes multiple factors, which one of them is how fast the website is loading. If too many users abandon a website before it loads, it might reduce our Quality Score.

Lower Quality Score means not only higher the cost per click, but also the fact that our ad displays more often on the bottom positions.

Page Speed / Google Ads / Quality Score
Page Speed / Google Ads / Quality Score

Website optimization problem

Each website is different, so the list of the required performance changes can be different very.

However, besides the website itself, we can make some adjustments within the website server to reduce loading times.

Is my website slow?

You can check your website speed by using multiple websites. I’d recommend checking the first Google PageSpeed Insights, which shows excellent website performance. You can also check what percentage of your users is affected by the slow loading times.

What numbers you should look at?

  • Desktop: above 80 points
  • Mobile: above 60 points
Google Page Speed
Google Page Speed

How much money do we lose?

Google has built a website calculator to let you estimate how much revenue you are loosing based on your page speed. Or how much you can gain if your website would load faster.

Page Speed / Google Estimate how much do you lose
Page Speed / Google Estimate how much do you lose

Server-side performance tweaks

Performance tweaks described bellow currently recommended and most commonly used. Implementing them to your server immediately reduce website loading times by a very significant factor.

HyperText Transfer Protocol – HTTP/2

HTTP is the protocol that allows us to transfer files between the server and browser. Currently, the most common version of the HTTP protocol is 1.1. But when you need to load multiple files, it is very, very slow. For every file browser request, it has to establish a connection. This process requires many communication cycles with the server. And it takes plenty of time.

The second version of the HTTP can send multiple files at the same time to the browser. It needs to establish only once the connection with the server. It saves plenty of time.

Switching to HTTP/2 is not complicated. The server administrator needs to install the additional module and enable it support. Your website does not require absolutely any changes.

We should get an improvement of at least 20% in the loading times. But more files your website requires from the server, the more significant difference in loading times. Sometimes even up to 40-60%.

HTTP/3 QUICK protocol

The next version of the HTTP protocol is slowly rolling out. The main performance upgrade is reducing required ping-pong requests between the browser and the server.

Google is already using it for all its services. They noticed an improvement of 8-13% over HTTP/2.

Page Speed / HTTP/3 QUICK protocol
Page Speed / HTTP/3 QUICK protocol

Brotli compression

Brotli is a compression algorithm used to compress text files from the server. The server is compressing files like HTML documents, CSS, and JavaScript before sending them to the browser. It allows reducing the amount of data to transfer, so the page is loading faster.

The server administrator needs to install the Brotli package and enable it. When done, your website starts serving compressed files without and additional change.

Page Speed / Brotli HTTP header
Page Speed / Brotli HTTP header

Brotli vs. GZIP

The most common server compression is GZIP. However popular these days is Brotli compression. Which compresses more efficient, and it is quicker.

What compression benefits you should expect?

  • 14% smaller than gzip for JavaScript
  • 21% smaller than gzip for HTML
  • 17% smaller than gzip for CSS
Page Speed / Brotli vs GZIP compression
Page Speed / Brotli vs GZIP compression

It is worth to mention that Google is also using Brotli compression for all of its services.

Time To First Byte problem

Or also common known as TTFB. Represents time between requesting a resource from the server and time when the browser receives it.

For example, when you enter a website, the browser requests the HTML document page from the server. If you wait for 1.0s for the server to send back the page, it is precisely what the TTFB describes.

You wait for 1s just to let the browser receive the first resource. CSS, JavaScript, and image files are not yet loaded. So for 1s, nothing happened, it is not good.

Depends on what language your website built on, you can find multiple cache mechanisms to improve loading performance. 

You can reduce the time server need to send back your page from 1s to 20-100ms, which is a very significant improvement.

Page Speed / Time To The First Byte time
Page Speed / Time To The First Byte time

Consider the use of the WebP format

WebP is a new web optimize format for images. It is a replacement for the JPG format. The main difference is that it allows compressing files with similar compression rates but around 30% lower file size. And that is quite a lot, to be honest.

Unfortunately, WebP is still not supported by Safari, and it is hard to say when Apple adds support.

The downside is that you would need to generate the second file besides the original PNG/JPG files. And some server configuration would also be needed.

Let’s quickly compare how WebP has an effect on my website. I’ve compressed JPG files with 82% compression.

Page Speed / Images without the WebP
Page Speed / Images without the WebP

And then converted them to WebP with the same level of compression, so images look identical. Difference? 412KB. And yes, you could try to compress JPG’s more, but still, you will see a significant difference in the file sizes.

Page Speed / Images with the WebP format
Page Speed / Images with the WebP format

The answer

You should put your page performance in the first place. With a few tweaks described above, you can improve your website loading times by at least 40% in the worst-case scenario. And this is before you even touch your website code, which could reduce loading time even more.

The revenue generated by the improved website covers the invested time and money very quickly.

Links


  • Contact Me
    Contact me if you need advice or if you need help. Would you please choose the most suitable contact channel for you?
  • My Services
    If you need a technical specialist to technical support to fix your problems, you are in the right place.
  • Who I am
    Hello, my name is Wojtek, I’m an experienced Software Engineer, with many many years of experience in the advertising industry.

Subscribe to receive updates about new articles.

[newsletter_form button_color=”#E74C3C”]
Show Comments (1)

Comments

    Related Articles

    5 Reasons Why Programmatic Digital Out-Of-Home is the Future of Advertising
    Performance

    5 Reasons Why Programmatic Digital Out-Of-Home is the Future of Advertising

    Welcome to the future of advertising! Programmatic Digital Out-of-Home (DOOH) advertising rapidly transforms how brands and marketers engage with their audiences. This exciting...

    Posted on by Wojtek Andrzejczak
    8 facts why IP location targeting is not accurate
    Ad operations

    8 facts why IP location targeting is not accurate

    IP location in advertising is nothing new, but if and how accurate it is? How do we know where this user using an IP is? What IP exactly is? IP stands for Internet Protocol...

    Posted on by Wojtek Andrzejczak