At the end of the day, it is always about how much product a site has sold and how profitable its business venture has been. The game always comes down on the conversion rate: the number of visitors versus the number of customers that purchases product.
And all this breaks down on how well the checkout page is perceived by customer. Being the last step towards selling product, there is quite some pressure on customizing it to suit customer demands. Choosing between quick and multistep checkouts is a critical question as well. Let us look at a case study to learn from practical example.
Case Study: How Insound boosted conversions to 54%
Insound managed to boost its conversion rates by 54 percent by simple changing a single call to action button on their checkout page. That’s right, a single button. Let us look just how by altering such a minor aspect of their checkout they managed such positive and drastic results.
Insound was among the pioneers of online music stores. Its products range from silk band posters to band shirts and vinyl records. After launching a new, revised checkout boasting bold field validations on form, they actually experienced a decrease in conversion rates instead of rise.
The shortfall in conversions posed a serious risk to their business as they already operated on slim margins which were highly reactive to risks among other probabilities. After a thorough analysis it was predicted that their conversions would get back to normal but it was critical to stabilize them for the short run as well.
Since their Holiday Promotions were nearing and Insound was not looking to spend money on failed marketing, it was critical to solve this decisive issue.
What was the issue?
Testing softwares pointed towards a likely culprit: the continue button being used redundantly. This necessitated modification to the buttons. Deemed a rather minute aspect of the overall checkout page, it was highly unlikely that such a small element would be causing such immense effects.
The checkout funnel, starting from the landing page to the finalization of sales at checkout page were using just continue buttons. This posed a serious turn off for customers who were starting to get the wrong ideas about whether they were progressing through or not.
Moreover, the new checkout was user tested and had great feedback on tablets and mobiles. It looked very dashing and captivating on mobiles and tablets. Even after the test had proven it to be successful for launch, the page was underperforming compared to its predecessor.
A/B Testing – Implementation and Results
To choose the best model for assessment, A/B Testing was chosen. Various changes were implemented and practically brought to field. During live testing over the course of two weeks, the various formats were brought on page and their progress was closely monitored.
Statistics of the models were precisely noted to evaluate their impact on page so that the best model could be chosen. Their analysis was taken which is given below:
- Review Order
Purpose: To show order is ready but has not been confirmed.
Conversion Rate: 54 %
Purpose: A straightforward explanation regarding the purpose of button.
Conversion Rate: 30.5%
Purpose: Gives the customer understanding that information is about to be submitted.
Conversion Rate: 29.7%
- Almost Done
Purpose: Boost customer confidence about order completion.
Conversion Rate: 27.5%
Results derived from A/B Testing
Although there are many models to predict customer behavior, at the end of the day they are still unable to compute the exact nature of customers since there are millions of visitors on a page every day. Predicting the behavior of such an enormous crowd is an extremely hard task with varying accuracies.
A/B testing yields the most accurate results since they provide statistics derived from on field, live testing of models on real customers, which is as accurate as it can get. If a model proves successful, it can immediately be shifted to serve as a field replacement until a better, improved version can be thought of since it has already proven its success.
The Four lessons of boosting Conversions
Following are 4 main lessons derived from this case study.
1. Choose A/B testing for accurate results
Testing is a very complicated process. Models might be able to predict performance but there is nothing like field testing with live audience. Choose A/B Testing for best results because it places prototype checkouts in front of real audience whose response is the perfect indicator of its success ratio.
A/B Testing allows data to be read in real time. This is extremely beneficial for calculating customer trends and actually predicting the future. The checkout model can be revised or even scrapped for other options once its health has been evaluated.
With a model working on field, A/B Testing boasts the benefit of permitting one model to continue functioning while a better, more successful and optimized checkout model can be built on the current feedback and statistics.
2. Build according to customer platforms
Platform to platform (desktop, tablet, smartphone), even the same customer can have different reactions. This means that during checkout development, care should be taken to keep them optimized as per device. For example, smartphones should have a minimum number of fields since they are small in size and customers are looking to purchase product fast on them.
Even though the new checkout designed by Insound was deemed successful by user testing, it failed to deliver results. This was due to unoptimized development which did not cater to the majority of shoppers who use desktops and laptops.
3. Keep the customer informed about progress
During the entire checkout funnel (landing page to confirmed sales), the user should be informed about what is their progress and where is their current position. This can be concluded from the redundant use of “Continue” button during the checkout funnel. It greatly confused customer and just by changing it, the conversions were at 54 percent.
Modules such as FMM Presta One Page Fast Checkout are well suited for such actions, coming loaded with a function at back end from which customers can be allowed to review their cart just one step before they finalize purchase.
4. Testing isn’t as hard as it seems
Always test your ideas in front of live customers for best results. Live field testing is the best option since it provides the most accurate feedback about models.
A/B Testing is done mostly on large scale models due to misconception that they require lots of time to yield results. This is totally untrue as this case study was conducted in 2 and half weeks. The A/B Testing accounted for 2 weeks only.
This proves A/B Testing is a highly versatile option with the capability to scale as per user requirements and deadlines.