We recently did some work on a large eCommerce website and within a week we were able to see an improvement of 50% in their overall conversion rate. No, this is not another story of an accounting error – this is a direct result of a solid ecommerce website with a ton of potential getting a proper tune up. We will share the details of what we did and what we plan to do next.

Before we start, let’s review: what is ecommerce conversion rate? We recently covered the 3 most important ecommerce metrics, one of which is conversion rate. A brief description is the number of transactions over the number of visitors over time. This number indicates the success rate of converting users into customers.

Ok, but really? 50% improvement in conversion rate with a single tune up? Yes! Really! The proof:

2.4% to 3.6% is a 50% jump!

ecommerce conversion rate increase

Incredible right? We had to cool our jets too when we saw that. Ok, let’s get to the meet and potatoes here: how did we help achieve this?

So before we cover what we actually did, it is important to realize that this particular customer has tremendous value to offer. They got down most of their eCommerce business: prices, product pages, product images, product search, online marketing, customer support, and order fulfillment. Even their shipping is great: low rates and fast shipping is the norm.

It is simply impossible to scale a business that does not have a great foundation. Ok, now that we got that out of the way, let’s take a look at how we increased their conversion rate by 50%:

1. Improved server response time. This is all about server side performance and in order to achieve this we have put some effort in improving cache and reconfiguring some aspects of the code to run smoother and faster. Page load time is super important for SEO these days. We were able to bring down the average server response time a lot but not optimal, still at around 500-700ms on average. Regardless we decided to continue for now as we found lower hanging fruits all over the place.

2. We added lazy load images to about 30% of images. While reviewing their site we found that there were lots of images that were not setup with lazy loading, in particular on mobile. This is a huge performance bottleneck. Since they already had the lazy load mechanism it was relatively easy to implement.

3. Move unnecessary javascript to load after the page is fully loaded. These days, every page needs a facebook pixel, adwords remarketing, listrak tracking (email marketing), snapengage, bounce exchange, and a ton of other services to help you market better to your users and customers. We moved most of these external libraries to load after the page has fully loaded. This helps render the pages super fast because otherwise these 3rd party libraries are blocking that process. No data is lost when done properly so the site is fast and functional if implemented right.

4. Fixed a bunch of javascript errors. So, over the years sites have plugins, extensions, and libraries added and removed to them. If you do not have a dedicated team of developers, you are likely introducing some minor JS errors. In most cases errors would slow you down or even block later JS from running properly. Cleaning up any JS errors helps streamline the rendering of web pages in modern and old browsers.

5. Removed old javascript that was updating CSS with workarounds. Ok, what was this? unbelievable! we found JS code that updated css styles for no other reason as perhaps the original developer could not find were the CSS rules were defined. This was causing the page to render twice and JS had to load into the page for this. We just hit the delete button on all of that.

6. Optimized images in bulk. We found so many images that were uploaded by non technical personnel that were way oversized and not compressed. Instead of optimizing images one by one, we used the awesome jpegoptim tool. There was no way to directly install it on the server so we used jpegoptim on docker and it worked great. We had tens of thausands of images optimized in less than an hour! Oh, and for the non technical staffers – the customer is now enforcing an online manual tool for image optimization that seems to work really well.

7. Improved UI/UX on mobile. We found several areas that were super clear to be clickable but in fact nothing happened when you clicked. Now this can be due to changes over time that were simply overlooked or poor UI/UX design decisions. Regardless, once you identify those they are really easy to fix. Needless to say, every eCommerce site these days must pay attention to their mobile sites.

8. Removed a block of hidden images which were loaded on every category page load on mobile. Again, what is this? where did this come from? ok, at this point we stopped asking questions. Old block removed. No design change. No functionality change. Saved in number of requests and bandwidth. Improved mobile performance.

Boom! My team made it a point to work in an agile way. So we finished and pushed live as each one of the above fixes were done. We sent everything to QA overnight and it passed. Done.

Ok, that was fun! can we do this again?

So what next? Now that we have shown our customer that a bunch of small & targeted fixes can bring huge benefits we have scheduled one day a week to work on performance in the next couple of months. We’ll keep you posted on whatever we find.

What increased ecommerce conversion rates for your sites?

Ron Peled

About Ron Peled

Builder of things. Builder of teams. Passion: eCommerce & Marketplaces. Magento expert. CTO Mentor.

Leave a Reply

%d bloggers like this: