Pilulka: Average Cart Value Increased by 4.21%

Steady place on several markets

Pilulka is a technology company operating since 2013. Initially, it was an online pharmacy and a network of brick-and-mortar stores in the Czech Republic. Over the years, however, they expanded their business to Slovakia, Hungary, and Austria markets. Their web domains offer thousands of products across pharmaceuticals, nutritional supplements, sports, and children’s products.

From the beginning, the founders of Pilulka aimed to commit to meeting customer needs. Therefore, they’ve constructed a comprehensive platform that seamlessly links manufacturers, distributors, service providers, logistics, and all relevant information within this segment.

One of the great advantages of an online pharmacy Pilulka is the same-day delivery via its own courier service, Pilulka Express or the Pilulka Boxes and Pilulka Points.

Roots of cooperation

Pilulka started using Luigi’s Box back in 2017, when we reached out to them with the offer of conversion rate optimization for their e-shop. At first, Pilulka tried out Luigi’s Box Analytics.

Later, in 2019, the first conversations about implementing Luigi’s Box Search began. But first, they tried it out in an A/B test on the CZ domain.

First A/B test on Pilulka.cz

On the Pilulka side, they gradually realized that their in-house search was not enough to achieve the desired results and provide a great customer experience. Between July 27 and August 25, 2021, the first A/B test of Luigi’s Box Search and Autocomplete took place on Pilulka.cz.

The A/B test yielded great results in search usage:

The result was the deployment of Luigi’s Box Search and Autocomplete on other Pilulka domains as well – Czechia, Slovakia, Austria, and Hungary.

Margin A/B test

At the end of 2023, we started another A/B test on the Pilulka.cz website, this time focused not on search usage and conversion rates but on margin.

The main goal of the Margin A/B test was to measure the impact of using margin signals in search of user shopping behavior on Pilulka.cz. So we took the original search and the modified search, where one was with margin signals and one without it, and compared them.

Primary goals

• uplift in Average Cart Value
• uplift in Average Cart Margin

Secondary goals

• change in Average Clicked Position
• change in Converted Position (within Autocomplete)

We evaluated the impact of margin on several metrics based on data collected between December 7, 2023, and January 7, 2024. Analytics included only non-bounced sessions (this involves visiting multiple pages, spending more time on the site, or completing desired actions) with direct conversions from search results. The traffic between the search with margin signals and the one without was split with a 50:50 ratio.

Results

Cart value and Margin value both play an important role in e-commerce businesses. Cart value, or Average order value (AOV), provides insights into customers’ purchasing behavior. Margin value, or profit margin, indicates how effectively businesses manage costs and generate profits.

Both metrics are essential for understanding and optimizing the financial performance of e-commerce businesses.

The results of the Margin A/B test on Pilulka.cz website showed an uplift in Cart and Margin Value by several percent with the search version with active margin signals.

Cart Value

Average and median cart value is calculated as the average/median cart value per session (total sum of prices of items added to the cart within the session).

Margin Value

Average and median margin value is calculated as the average/median margin value per session (total sum of absolute margins of items added to the cart within the session).

This internal A/B test, where we took into account margin, showed results with a considerable uplift in cart value and cart margin value using margin signals.

A/B tests and experiments in the app

One of the new features that Luigi’s Box will bring in the second quarter of 2024 will be the possibility of A/B tests directly in the application.

The procedure itself will be very straightforward. Just a few clicks and you will be able to define, run, and evaluate the A/B test yourself. And all this without programmers or any programming skills.

Since users behave slightly differently in each segment, we are considering unique profiles for different segments (e.g., fashion).

Further, there will be a possibility to manually adjust the importance of individual signals (e.g., popularity, freshness, personalization).

Finally, as for the evaluation, you will be able to see the A/B test’s impact on key business metrics, such as sessions, conversions, AOV, CVR, etc., over time.

Summary