In this case study, we will show you how (an appliance eCommerce store) was able to boost their conversion rate by 25% in just 6 months through our A/B Testing tool and interactions in 2018.

The eCommerce Appliance Industry – A First Look

The large appliance industry has a particular behavior when it comes to online purchases. Based on consumer insights from Google Analytics, 66% of the of large appliance purchases (such as refrigerators, stoves, and washers) are made to replace broken appliances (39%) or to upgrade old ones (27%).

In the case of broken appliances, timely delivery is a very important criterion when deciding what retailer or brand to choose[1].

01 1

Another important aspect is the influence online stores have on offline buying. People are researching online stores before buying appliances from brick and mortar stores.

The tendency to make this type of purchases in physical stores is backed up by data: according to a study conducted by PwC in 29 countries[2]. 56% of the household appliance purchases and 51% of the consumer electronics and computer purchases were made offline, in brick and mortar stores. *

Key Takeaways from the Appliance eCommerce Industry:

  • Almost 50% of the online shoppers are researching and purchasing in less than 2 weeks;
  • 51% of potential clients are undecided on what brand/manufacturer to choose from;
  • 86% are mostly searching for products by category name;
  • Pre-purchase research is made 56% of the time online, on the store’s website or app;

02 1

About the Customer, a Romanian eCommerce store, offers a large variety of top-brand household appliances from washers and fridges to ovens and microwaves. It is also a customer-centric brand that comes with full-stack customer service.

In Romania, the main criteria that shoppers consider when buying products are price, brand trust (based on shopping experience, feedback), and product range.

Around 25% of potential consumers first “showroom” products of interest to receive more information about the related products.

03 1

The Challenge’s competitive advantage lies in strong customer support and low prices for large home appliances.

The real struggle was to compete against other marketplaces or brands that offer a larger range of product pages and have a higher market share.

Other challenges they faced with their website included:

  • Creating a suitable user experience for their target audience;
  • Great usability experience;
  • A good showcase of the shipping cost;
  • Having good information architecture (on ToFu, MoFu, Bofu)

04 1

The Testing Methodology

In 6 ½ months, we followed a structured methodology starting with qualitative & quantitative analysis and UX research. We put them all together in an optimization strategy.

The first part of the optimization process consisted of two important phases that generate long-term results and have the role to create the basis of the upcoming plan: setting up the technical aspect of your website and correctly running your research.

05 1

In general, our advice is to not jump into testing in the first months.  Check the tracking of your website and invest in discovering information about visitors and customers.

It is crucial to set your Analytics correctly so that you properly track the activity of your visitors, understand the customer loyalty and how website visitors interact at each touch point.

This way, you will build an optimization strategy based on accurate and reliable data.

From our qualitative research (surveys) we found out the following: what people like & dislike about the brand, what are the obstacles they met on the path to making a purchase, what is their consumer behavior and what are the demographics that defined them.

06 1

The survey conducted for visitors who purchased something on the website uncovered the reason behind buying more appliances in a certain period of the year. The high interest was due to the need to furnish their new home. Therefore, based on this insight, the team created a specific Facebook campaign that was one of the top best performing campaigns of the year. From quantitative data (Google Analytics, Mouse tracking & Surveys), we found out how visitors interact with the website and how each part of the website performs.

We then started to optimize the homepage, testing iteration after iteration, which has eventually led to a final design (currently running in their website)

#1 Optimization for the Desktop Version

For the desktop version of the website, we looked at the homepage and category pages and analyzed what could be improved. The changes ensued as follows:

  • Homepage on a desktop before testing:

Screenshot at Apr 27 15 48 19 1

  • Homepage on desktop after testing:

Ideall homepage after

  • Category page before testing:

screencapture web archive org web 20160910223053 https www ideall ro magazin aparate frigorifice 2018 04 27 16 11 13 1

mock up variatii ideal pag categorie var 04 1 1

The types of A/B tests conducted on the desktop version of the category page were::

  • Filters on the left side
  • Buttons from the product image
  • Product alignment
  • Product information
  • Comparing functionality
  • Discounts

#2 Optimization for the Mobile Version

In terms of mobile devices, we tested almost the same hypothesis to keep the content consistent on desktop & mobile ecommerce platforms, but we took another approach in terms of adjusting the elements.

The category page was transformed to include a lighter presentation of the products recommendation, introducing some elements that help visitors decide faster and narrow down their decisions:

Before vs After

0eff9fa832556da38990020e37ef0dee 1Screenshot 5 1
On the homepage, the main focus fell on the slider (initially not user-friendly), and then, the transformation continued with the menu and the sub-categories.

Before vs After

d1d418109455f4d8d910895d8213bf17 1Screenshot 6 1

Types of A/B Tests Conducted on Mobile Devices:

  • A/B testing the filter display (very important feature for the website)
  • A/B testing product “SEE details” button
  • A/B testing the price
  • A/B testing the “Fast delivery” icon
  • A/B testing the discount label
Desktop13 A/B tests6 Interactions
Mobile5 A/B testsNo interactions

After each test adding was validated, the layout was implemented by the client in order to have a new layout to build on the next hypothesis. We conducted 24 different tests (A/B tests & interactions) split between desktop and mobile devices.


15 of the total 24 tests won and were implemented on the website. From the inconclusive tests, we learned a lot of things that you could not normally figure out without ecommerce ab testing.

07 1

Results from A/B Testing with Omniconvert

Preliminary results after 6 and a half months (including almost 2 months of research) proved an increase in the conversion rate of no less than 22.26% (including just winning A/B tests and interactions) and an increase in revenue/visitor data of 14.23% with over 95% statistical relevance.

08 1

Metrics performance
+22.26% in CR+14.23% in Revenue/ visitor+ 9.90% in Average Order Value

Conversion rate with Omniconvert  vs. without Omniconvert tests

pasted image 0 34

In the case of all the experiments, including inconclusive experiments:

  • the conversion rate has increased by 9%
  • the revenue/ visitor has increased by 8.32%.
  • the average order value has increased by 1.39%.

09 1

The old version of the website’s interface was replaced with the winning tests just by using Omniconvert scripts, HTML, and CSS.

(A Side Note about) Incremental Uplift

When you decide to conduct optimization tests on your website and evaluate the impact on the business of the testing process, there are some basic calculations you need to make. These calculations will determine if your testing idea paid off in terms of costs and work involved.

The incremental uplift describes the total uplift optimization experiments can bring to a business in 1 year in terms of revenue (considering the cost of testing). The revenue is calculated based on the projected revenue for 1 year brought by the most efficient experiment(s). First, we have to divide revenue to the number of days to find out the revenue/day.

In our case, we used the minimum impact, meaning the best performing experiment from desktop devices and the best performing experiment from mobile devices. Cost of testing includes the cost of the testing platforms, the cost of human resources and the cost of researching tools (surveys, user experience, user testing, user recordings tools) .

After you calculate the revenue and the cost, the return on investment (ROI) can be deduced.

A raw calculation of the ROI uplift would be something like this:  (Minimum impact in revenue – Omniconvert cost)/ Omniconvert cost


Revenue Impact from best 2 performing experiments for 12 months103949 euros
Cost of testing for 12 months4500 euros
ROI in 12 months2209%

What was the incremental uplift for testing with Omniconvert?

The incremental uplift = [ (total revenue brought by the best performing test on desktop +  total revenue brought by the best performing test on mobile) X 12 months – (total costs x 12 months)] / total costs x 12 months

Minimum impact: Value brought by the most successful experiment on desktop + the most successful experiment on mobile multiplied by 12 months (1 year).

Omniconvert cost: the cost of services x 12 months

The raw uplift on ROI for this particular client was 7561%. This means that in a normal environment, this could be the uplift of the revenue in the next year, considering that all the changes have been implemented.


[clickToTweet tweet=”When #A/B testing the layout of your website you have to progressively test changes, but they have to be visible and #data-driven to prove their importance. ” quote=”When A/B testing the layout of your website you have to progressively test changes, but they have to be visible and data-driven to prove their importance. “]

For example, if you found out that the comparison function brings a higher conversion rate, then emphasize it and make it easy to use to increase the navigability.

Testing and conducting qualitative conversion research brings a lot of benefits – such as gaining and encouraging you to make changes to your website (precisely because they are validated by data).

Benefits of Auditing Your Website:

  • Finding the undiscovered behaviors and patterns  of your audience
  • Finding the specific and uncovered needs of yourwebsite visitors
  • Finding why visitors are choosing other competitors over you
  • Finding usability problems on the website

Benefits of Testing Different Layouts on Your Website:

  • Validate your ideas
  • Discover pieces of information that were missing from your website
  • Discover which is your best-converting segment of visitors on the website
  • Address some information just to a specific kind of visitors or clients
  • Discover which is the best spot to place information
  • Find out what is the minimum discount that works on the website
  • Find out where the visitors are losing interest in forms

Key Takeaways:

  • Discover the ups and downs of your website through qualitative and quantitative analysis
  • Do research about your competitors
  • Don’t hurry to draw conclusions on your tests if you do not have statistical power
  • Await your returning visitors to get accustomed to the changes
  • Measure, measure, measure everything
  • Implement the tested changes that won and build the next hypothesis

What’s Next?

If this scenario appeals to you, then make the first step towards optimizing your ecommerce business. Omniconvert offers you the tools that will give your business the magic (data-driven!) power to help you grow.

You can opt-in for a FREE 50K tested views account or for a full managed services plan that can handle everything you need for your optimization strategy, design, implementation, and so on.

Get your Free Omniconvert account today and boost YOUR conversion rate too!


Get in touch with our team and learn more about our managed services pack!

[1] The Role of Digital In the Large Appliance Shopper Path to Purchase, Google/Compete U.S., May, 2012