In one of our recent blogs, we discussed the implications of UX / UI design or backend engineering and how this can improve your e-commerce conversions.

A small recap:

As business owners, we measure certain key e-commerce metrics to evaluate the overall performance of the store. For example Bounce rate, channel conversions, cart abandonment rates among others. However, these metrics give you a holistic picture of store performance. This can be a combination of multiple parameters such as your site design, backend engine, pricing, catalogue, marketing communication among several others.

Therefore, it is challenging to identify accurately where the problem lies. Should you improve your offer strategy, marketing, pricing, catalogue, store design or app engineering? E-commerce business owners usually rely on intuition or internal testing/website checks to solve this. However, this might not be enough, since your users are using the site in very different contexts.

Where lies the problem?

While there are enough well documented resources to measure and improve the impact of pricing, marketing communications and other business / marketing aspects on E-commerce ROIs; there is limited documentation to single out and understand e-commerce UX and  e-commerce backend application performance and its impact on your e-commerce ROIs.

My previous article focused on how to evaluate and identify problems in the UX / UI design of your store. Today, we are evaluating your e-commerce store’s backend engineering.

Measuring and benchmarking your site speed

For E-commerce business owners page loading time is the acid test of application engineering. A good application ensures your pages load quickly at various level of traffic and brings you the right data quickly at the right time and place.

There are online tools such as yslow (powered by Yahoo) and pingdom that you can use to understand if your site is running slow, compare it to global benchmarks and a few action points to improve the site speed performance.



While Yslow leverages a set of coding best practices and evaluates your site on those parameters, pingdom gives your insights on your site speed when compared to global benchmarks and what elements on the site are taking the most time to load.

Following best practices

If the above tools show that your site speed performance is not up to the mark, again there might be multiple reasons why that might be happening. Although a thorough root cause analysis will involve expert’s review of your code (Btw, we do this at Kuliza and you can contact us here); I have listed a few common reasons below that e-commerce business owners should know:

  • Leveraging CDNs: CDNs are a collection of servers that deliver content based on the geographical proximity to the users. Content Delivery Networks (CDNs) improve your page loading speed by ensuring user’s proximity to the server delivering the files. CDNs are especially useful when you have a large number of images or rich content files. If your e-commerce site page speed analytics show high loading time for images, CDNs are a must have.
  • Leveraging smart caching: Smart caching can really improve your e-commerce site speed. The important thing here is to implement caching at every level – presentation layer (frontend), business logic (backend) and database. You can also use combinations of caching systems such as Memcached, Redis and Varnish to achieve this caching.
  • Allow progressive page rendering: To gauge your users’ attention, it is imperative that the page is loaded progressively, and the browser displays whatever content it has as soon as possible. This means that your style sheets does not hinder the progressive loading of the page. This also means loading the most important components of the page initially or pre-loading components while the browser is idle.
  • Optimizing Images: This optimization can happen at multiple levels, starting from converting gifs to PNGs, since they are more compressible. With the modern browsers, you need not worry about PNG support. Optimization can also happen at the server level, depending on the resolution of the device requesting the image, the delivered image can be of a higher or lower resolution
  • Leveraging flat databases: Flat databases (also known as no SQL or unstructured databases) have often been in the news because of their use in Big Data exercises. However, what is relatively lesser know is that flat databases can be used to improve your e-commerce site speed. We recommend replacing EAV (Entity Attribute Value) model, that is often used in E-commerce databases and are not most efficient, with Flat databases to simplify queries and fetching of data.
  • Separate the business logic from presentation layer: We preach and practice at Kuliza for the entire business logic (backend of application) to be separated from the presentation layer (front end) and deployed on separate environments and integrated through webservices. This ensures that both the presentation layer and business logic can have their own growth trajectories and load on one layer does not impact / hang the other layer.

If you’d like to know how you can identify when the frontend is the culprit of your slow conversions, you can read more about it here.