Automated web performance optimization supports providers of e-commerce in finding the right balance between size and format of the graphical material used and the performance of your website.
When the Frankfurter Allgemeine Zeitung (F.A.Z.) printed colour photos on the title page for the first time in its history in 2007, it was a revolution for the well-known newspaper and its readership. It was the year the first iPhone was introduced to the market, which redefined digital photography. That was the beginning of smartphones suitable for the masses with really good digital cameras; today, as many as 2.5 billion people use such “telephones”. And they take many photos with them, very, very many: 1.2 billion photos were taken in 2017 alone.
This huge quantity of photos is increasing constantly; social networks above all play a key role in this. Not only can photos be simply shared via Snapchat, WhatsApp, Instagram and other networks, at the same time they are – besides videos – the most important means of expression for a great many users.
A picture forms the first impression
This does not only apply to the personal filter bubbles of social networks: Images, whether photos or graphics, are a key element for most websites. E-commerce providers especially rely on photos as the most universal of all “languages”. If users were not shown the product in question in detail, then only a few of them could be turned into paying customers.
Photos thus have to be placed on the Web, everywhere. Whilst an Instagram user can simply get started because the app takes care of most of the work in processing and uploading, users that manage an online-shop via a CMS are confronted by (at least) four problems when working with photos:
1. File size: Modern cameras take increasingly better digital photos, but at the cost of increasing file sizes. Embedding a 10-megabyte photo on a website is usually not reasonable; the file has to be made smaller so that it can be displayed quickly to users with a slow Internet connection.
2. Image format: To make sure a photo is not distorted, the width-to-height ratio in relation to the presentation on the website must be taken into consideration and thus frequently needs to be altered as screen sizes and resolution evolve fast
3. Missing caching information: If there is a lack of image caching on complex websites, or none at all, the loading time increases. That would be easy to rectify but is frequently ignored.
4. File formats: There is an unmanageable number of RAW formats where photos are stored as raw data of the respective camera. And it goes without saying that with JPG, PNG, GIF and WebP there is more than just one standard format for online usage. And each has its own good reason for application. What should be used when, where and how? It just depends…
(Yes, besides these there are other problems. We are focusing on these four here.)
Problems with photos can have a negative impact on sales in e-commerce
Getting these four problems under control simultaneously and sustainably is no trivial matter. Many cases can indeed be simply automated, such as the conversion into a certain file format or the reduction of the Dots-Per-Inch (DPI) number, but in most cases, such tasks and others are taken care of by humans. Such manual workflows between photography, graphics and marketing cost working hours, i.e. money and are one of two reasons why the problem with photos is also a financial one.
The other reason, especially for e-commerce providers, is the more important one: Problems with image files frequently lead to longer loading times of a website. This has a negative impact on sales – and, in fact, a direct one, not one that doesn’t appear until the next quarter or year.
“In retail, we see that for every one second delay in page load time, conversions can fall by up to 20%”, is an insight by Google. And because the captain of all search engines especially wants to recommend websites to its users that offer a good user experience, slow websites are displayed further down the list of results. The visibility goes down and that can have a huge impact on sales.
Even Amazon is doing its own research into the subject of “loading times” and arrives at the result: “Amazon calculated that a page load slowdown of just one second could cost it $1.6 billion in sales each year.” It’s obvious that, besides optimising the check-out process, Amazon is mainly investing in as fast a website as possible.
Providers should thus find the balance between the scope and format of the graphical material used and the performance of their website. Too few photos or such of bad resolution are just as bad for sales as long loading times.
Two different methods of resolution
But there is hope after all! Artificial intelligence in neuronal networks will solve the problem once and for all!
Nope. (There may well be first approaches, but the journey there is still a long one.)
Until we have got that far, it will remain a great challenge to keep photo-laden websites up-to-date, optimise them for conversion and deliver them as fast as possible – all at the same time. Instead of focusing on solving the photo problem, it stands to reason to think bigger and to optimise the speed of the website straight away, irrespective of how many and which photos in which formats are placed on it. “Web performance optimization” is what this is called, and it not only addresses graphical material but scores of other factors that slow down websites.
Providers that find speed AND photos sales-relevant and who wish to confront the problem have two possibilities:
1. Manual optimisation: There are many developers that are able to optimise the code of a website in such a way that the performance is increased sustainably. However, a provider should get them to commit themselves as far as possible as it is a long-term and recurring task. What a “fast website” means in practice is determined by customer requirements and technology – and both are developing at a tearing pace. One needs to be able to afford permanent manual optimisation.
2. Automatic optimisation: For this reason, automated web performance optimisation is frequently the better alternative here, for, instead of constantly updating the code of the website, tools such as “wao.io” can conduct adaptations relevant for performance even “on-the-fly” during transmission to the users. This is a completely new approach and in most cases, more efficient than manual optimisation where new efforts are needed each time for development, testing and roll-out.
Which solution is the better one depends on the initial situation. Amazon, for example, develops new technologies particularly for this purpose and relies on a mixture of internal manual and internally developed automated optimisation. Amazon, however, plays in its own league; this is out of the question for 90% of all e-commerce providers.
This is where automated optimisation, which is not yet widespread, comes in. It does, however, have the potential to bring good performance (and security!) to the standard for every online shop. Sevenval relies here on “wao.io”, for the activation of which a provider only has to adapt the Domain Name Server (DNS) entry for their website. An algorithm developed specifically for wao.io significantly reduces files sizes of JPG files without these suffering the loss of quality. “Many good photos” and “short loading times” do not rule each other out.
This article is based on the presentation “Photography curation in a digital age” by Jonothan Birkett at “CGNwebperf #22” in March 2018. Jonothan is part of the Sevenval team and co-responsible for wao.io, the new on-the-fly service for web performance (& security) optimisation for websites.