You may think life is %90 routine,but in fact we don’t realize how frequently we are subjected to a “test” or we act based on a result from a “test”
No I’m not referring to your overdue essays, or driving licence exam.
Okay okay, Let me visualize it further.
Imagine you woke up on a regular working day with your alarm set for 6:30 am
-You woke up with your alarm because you tested If you don’t set your alarm you may oversleep. (A conducted test result)
-You woke up at 6:30 am because. It is the best time for you to both get ready in comfy setup and not to be late for work. (A conducted test result)
This list can be longed to a full day of your “routines” but the bottomline is simple.
“A meaningful action must rely on a proven result.”
So as CRO experts, we apply the same mechanism in e-commerce, we seek for concrete results prior to taking an action on the website.
The mechanism that helps us to reach a result is basically called the A/B testing.
What is this A/B testing?
A/B Testing is basically setting up two different experiences for randomly splitted traffic for any website and comparing these two variants’ performance for the pre-defined goal or action.
Running an A/B test is really simple.
- Find a hypothesis
- Test
- Analyze & Iterate
Finding a Hypothesis:
A hypothesis in A/B testing jargon is an educated guess you crafted from
- Your Website Data (Google Analytics, Adobe Omniture,HotJar, Crazy Egg etc.)
- Directly from your customers (User Surveys, Audits, Client Reviews, Return Causes)
An hypotheses can be;
Source: Google Analytics Demo Account.
On the checkout page, people are dropping above %65-70 the reason might be the lengthy unnecessary form structure.
Test:
As we already have a hypothesis that we believe is the reason behind the data we extracted, we need to come up with a testing apiterius that should consist;
- Design for Variant A (our test group) as we will keep variant B (control group) as the website is, we don’t need two designs.
- Implement this variant to randomly splitted traffic via A/B test tool to a given traffic % (A/B Tasty, Insider, Google Optimizely, Qubit etc.)
Results and Iterate:
As our problem was already defined in the hypothesis part, we will need to track what if variant A is resulting in a better next page penetration or not.
But the critical question here is when to iterate, which means for CRO experts; Adopt it to full scale or seek for further optimization suggestions.
So, here the most important indicator is statistical significance, which basically tells you whether your test has converged to a concrete result or not.
CRO experts generally measure on Bahesian method when determining the significance but still a good amount of experts are still holding up to the z-significance test as well, that said, you can rely on any online tool available.
But, If the aim is a holistic Conversion Rate Optimization unfortunately “One swallow does not make a summer.” So, a single test would be a good start but it is just the beginning.
For a proper CRO project you need to follow these “simplified steps”
- Analyzing data and come up with a data-centric list of hypotheses to not to hurt your existing conversions.
- Prioritize hypotheses and prepare a A/B testing roadmap (It is not ideal to run so many tests at the same time, it will ruin your attribution modelling and delays you to reach to a significant results due to more variables included)
- Come up with use cases that will match with the hypothesis and map those according to funnel
- Implement complex tests to the website by CRO tools
- Test until a significant outcome (Statistical Significance)
- Report multi-layered attributed conversions for individual campaigns.
- Iterate to further results and test
Final Words
All in all, when you look at just the points above you may think you need an entire CRO department to be recruited. Indeed.
However, as 3WL with proven 15+ years of experience in total, expertise in almost all verticals, and worked with vendors, we can provide the ultimate experience to set your A/B testings and ultimately Optimize Your Conversion Rate seamlessly.
So, you can have your free audit and strategy here!