A new e-commerce chapter is coming to our blog, inspired by the launch of our newest analysis module entirely dedicated to tracking e-commerce performance. Today and over the next few weeks, we’ll be sharing our vision of what efficient, data-driven e-commerce looks like. We’ll examine the challenges and opportunities available to e-retailers who use data as a catalyst for e-commerce action.  We’ll give tips for analysing the user experience, optimising conversions, and building customer loyalty  thanks to automation.


To kick things off in this first article, we’ll look at why an analytics strategy is a natural and necessary element for any online merchant looking to sell efficiently.


Measure your site’s performance (before investing more)

Often, we equate an e-commerce site’s health with how efficiently it can acquire traffic. We think of the following equation: more traffic means more turnover. We all know the main acquisition channels: SEO, display, affiliation, social networks, marketplaces… but while acquisition spending is skyrocketing, ROI isn’t. More than 9 out of 10 visits to e-commerce sites create acquisition costs without generating any ROI. The “leaky bucket” metaphor nicely sums up the ineffectiveness of this simplistic approach: Why continue pouring water into a bucket full of holes? (And why not focus on patching up those holes first?)


Of course, traffic acquisition is important (if you have no visitors, you’ll have a hard time building loyalty with anyone), but before focusing on any acquisition goals, first evaluate your website’s performance. In addition to your main performance indicators like turnover and number of visits, be sure to check conversion rates, repurchase rates, product return rates, etc. There are many optimisation levers when it comes to strengthening a website’s ability to sell: the product catalogue, personalising the purchase journey, the UX at critical phases of the funnel (product pages, cart, payment, …), growing customer loyalty, etc.


This optimisation, based on results of your data analysis, is absolutely key to making your online store stand out from the crowd, to improving customer engagement, and to creating greater efficiency.


Optimise your merchandising 

Even if your merchandising’s impact seems to go unnoticed (especially compared to the impact of a media buying campaign, for example), optimised merchandising should not be forgotten in an e-commerce strategy. Examples from the physical retail world are inspiring for their ingenuity and efficiency: signage, arrow-guided pathways, strategic in-store product positioning, olfactory triggers, and special additional services…

In the digital world, there are 2 major challenges:

  • Present a catalogue of products or services that’s consistent and relevant
  • Incite the visitor to fill his/her cart from the product page

To conquer these challenges, you have several optimisation levers available: internal search tool, product categorisation, chatbot-type tools, product lists, user experience, product pages, customer reviews, cross-sell and upsell opportunities. Behavioural analysis of visitors, made possible with analytics data (visits, searches, product views and purchases), is therefore crucial to be able to make an offer that’s appropriate.

Let’s stop for a second to talk about your site’s internal search tool… and in particular, searches made on your site that don’t lead to anything. Your analytics tracking enables you to recover all keywords entered into the search bar, so you can ensure that a typo isn’t causing an absence of results. And especially considering that visitors who use a site’s internal search tool tend to convert 3 times better than other visitors, you’ll definitely want to ensure your site search is returning results.


Strengthen your test & learn strategy

You can’t think about merchandising without thinking about testing, too. A/B testing is an efficient optimisation technique which many e-retailers use. Test variations are practically infinite for an online store: you can test visuals (add-to-cart button colours, size or shapes), different navigational steps, highlighted prices or stock, shipping fees, placement of reassuring elements, information included in a form, payment process, and much more…

But trying to manage a test & learn approach while flying blind can penalise you. Whether it’s loss of time, useless optimisations, or incorrect data interpretation, there are many risks.

Linking your testing (or CRO) tool to an analytics solution is therefore highly recommended. Upstream of your waves of tests, you can easily identify your site’s weak points (where you’re losing traffic, for example) and better target your tests and corrective actions. With data, you can also detect which pages have strong potential for tests focused on promotions or similar products, for example. Your analytics solution also enables you to verify and qualify (with great precision) the results of your A/B tests. Should you decide to target a specific market in particular, simply isolate data from a single country, thanks to geolocation analyses. Find more information in this article on combining A/B testing and digital analytics.


Turbocharge conversion rate

Nearly 30% of visitors abandon during the checkout phase because the payment process is too long or too complex. Your purchase funnel is essential to conversion. A customer who engages with your funnel is an immediate sales opportunity. Your priority is to limit cart abandons as much as possible by making the purchase process extremely simple. And you can do this by having a clear understanding of buyer behaviour at each step of the order process, which will enable you to then tweak your site to offer a streamlined, fast and reassuring experience. This customer knowledge is available right here in front of your eyes in your analytics data. Thanks to this data, you can break down each step in the customer journey. For example, you can:

  • Measure contribution to turnover from related products recommended on the cart page
  • Examine the impact of delivery options on non-converted visitors
  • Define a fee tolerance threshold for shipping
  • Compare the effectiveness of different payment options depending on number of orders placed
  • Detect technical errors with different payment options (slowness, breakdowns in secured systems, etc.)

This little list is far from exhaustive. Many insights will allow you to refine your offerings at different levels of the funnel. The result? Your conversion rate will grow noticeably.

And let’s talk about this famous conversion rate for a minute. Generally speaking, conversion rate is calculated based on visits: it’s the ratio between visits that converted and the total number of visits. But you can also focus on conversions per visitor. A “visitor-centric” analysis takes into account the (temporal) granularity of the purchase cycle. For example, a drive-through-grocery-type website might expect to see weekly orders, whereas a site selling jewellery might have a much lower frequency of conversion. When you take a “visitor-centric” angle, your analysis will be more accurate, close to reality, and easier to act on when making decisions.


Create greater customer loyalty

Here’s one number to keep in mind: 27%. That’s the probability that a customer will come back after making a first purchase. In e-commerce, the secret to creating loyalty with your customers is customer knowledge (again!). And that knowledge can very well come from automatic segmentation technology. You’ll save precious time and avoid manual errors that can be potentially damaging. So how does it work? Machine learning algorithms are integrated into your digital analytics solution, and they can automatically detect different buyer populations who are then categorised into different groups (recent buyers, loyal buyers, sceptics, dormant customers, thrifty spenders, etc.). The system is based on activity detected on your site and different thresholds of purchase recency, frequency and amounts. But do take note, even if automation can save you time, your good analytical sense will always still be necessary to:

  • ensure you’re not blindly inundating your customers, and to take their purchase history into account when retargeting them (we’ve all received an email or been shown an ad for a product we’ve just bought… #fail)
  • be attentive to all legal aspects relating to retargeting actions (the GDPR says so!)
  • take a product’s purchase cycle length into account before reactivating a campaign


If you only remember one thing, think “site optimisation” before investing massively in all your traffic acquisition channels. Bet on your data to improve your sales efficiency. But remain alert to the quality of your analytics data. In e-commerce, reliable data isn’t just a luxury – it’s an absolute necessity. Sampling, for example, is not acceptable when working with payment and turnover data. Loss of traffic due to adblockers and to the technical limitations of certain analytics solutions (such as losing the browser on the order confirmation page) are major roadblocks to tracking your performance. Be sure to examine all these issues carefully before embarking on your data-driven e-commerce adventure.

Looking for guidance, best practices, tips and expert recommendations? Check out this guide:


Editorial Manager. Bernard is responsible for the content strategy of the AT Internet brand. He has almost 10 years’ experience in technical and marketing writing in the Software industry. A word specialist, Bernard’s job sees him working on many different mediums including blogs, white papers, interviews, business cases, press, infographics, videos etc. His specialist fields? Marketing and digital analytics content of course!

Comments are closed.