While I have been working in FiloBlu, I was the person in charge to develop CRO tactics and analyze business data for delivering actions to improve the e-commerce performance sales.
Helping all departments in the company, I documented how to uplift the Conversion Rate and deliver better results for several national and international fashion brands (i.e. marketing budget optimization, tactics, merchandise machine learning based, mail marketing automations, etc.).
Lotto Sport Italia had some important improvement and it was a challenging project too!
Understand through deep dive analysis into Google Analytics and Customer Base for solving business gap and uplift sales.
Solve funnel issues in less time as possibile (aka, yesterday!). Such a constraint!
I started with understand what it was done, until that day. I was looking for drop-off issues evidence, investigating inside analytics (quantitative data). After a matching with UX heuristic, there were some UX improvement to deploy.
Because all starts from data evidence, I created personalized reports to make them obvious, avoiding bias and personal perceptions/opinions of creatives.
The best way to proceed is create a plan experiments for a continuous e-commerce improvement.
Looking for specific indicators of customer behaviours ( through quantitative and qualitative analysis), I created new experiments for testing hypothesis, and coming out with solutions to deploy.
A small part of visits that drive an high conversion rate came from search navigation. I proposed to:
- Create a bigger search text input for improving better conversions;
- Improve search results in order to deliver a better experience;
- Up and Cross selling, suggested by recommendation algorithm improving AOV (Average Order Value) and reduce abandonment rates in some funnels.
I rolled out by Google Optimize all tests, splitting the traffic in a small part to preserve the business, in case it goes wrong.
The search bar was enlighted much more by a strong outline along text input, and a visible placeholder. It was pretty much more evident and fits really nice in e-commerce design.
Furthermore I propose to stretch a little bit, leading to redesign menu architecture: usually when you edit the menu architecture, it's not a easy task! Fortunately I had plenty of data to suggest a better structure.
For a better performance then an autocomplete search text (usually, in Magento, it's implemented Solr o some similar text based technology) I implemented an external product which it delivered recommendations to search results, based on customers purchasing and navigations insight.
Search bar with autocomplete intelligence engine
As you can see, in the following picture, I wrote "scarpe" (shoes) it's a general term but the search engine show me the most probably interesting results analyzing in real time my cookie.
This live search growed up the conversion rate more than 30%, only for sales coming from this starting point to checkout process.
The machine learning algorithm make us to create a more engaging user navigation, lowering the bounce rate (product, checkout, etc ..) and it reduced the abandonment rates, replacing some e-commerce section page.
Other important improvements needs more time to be delivered, such as implementing a complete visual personalized merchandising in order to create a real personal shopping experience, creating an hyper-personalization which deliver show a product based on customer behaviour, integrating other data sources too.
More engaged customers through machine learning recommendation engine.
I notice that it works really well when a user starts his purchase process in the search bar; in the product page, the cross-selling section works well too uplifting AOV (average order value) since in this phase because of the machine learning engine recommend good alternatives to current the product, fitting their needs and driving customers to checkout page.
+200% uplift conversion rate
For all customer sessions which starting the purchase only from the search bar
During the test the conversion rate increased by as much as 12.5%. On mobile devices, the performance was even better. The tested changes improved the mobile conversion rate by as much as 29%!
While all the updates had an impact, the new and improved shopping cart proved to be the biggest contributor to the uptick in conversions.
To translate these percentages into monetary terms, improving the conversion rate by 12.5% means that if Lotto had monthly sales of 100,000 €, the increased conversion rate meant that they brought in an additional 12,500 € each month.
Not bad considering it only took six weeks for getting these results.