By: Abby White
Conversion Rate Optimization (CRO) is the process of raising the percentage of visitors that take a given action. The action is most commonly measured by purchases, but leads and user engagement can be used as well. Our clients want people to become buyers, and they have to figure out what causes people to leave without purchasing. This can all be done by modifying and designing some changes to a webpage to register behavior.
CRO and SEO (Search Engine Optimization) at Impakt are generally mentioned together. The more total visitors a website gets, the better CRO is for them. Our specialization in eCommerce leads us to prefer to offer CRO to clients over SEO, and I think this writeup will explain why we think that CRO as a service offers more bang for a clients’ buck. Let’s dive in.
We will outline our process exactly here, so that aspiring agencies or clients that can not yet afford our services can grow their business. We will also provide a case study by which you can see some of the more simple tests that we execute.
Once the client has signed their contract, the Impakt web team immediately logs into the client’s website and installs Hotjar. Hotjar allows Impakt to watch recordings of people visiting the client’s website. We come up with a hypothesis as to why users are not converting based on these recordings. Impakt also collects user surveys to ask them why they are happy/unhappy with their experience. We get this information going as fast as we can, because it allows us to have a larger dataset to start with in the second week.
Once a hypothesis is established, Impakt installs Google Optimize on the client website. Google Optimize allows Impakt to run tests to see if our hypothesis is correct. These tests can be A/B tests, meaning the original version of a page will be served 50% of the time and the variation that we create will be served the other half (this is the most common type that we use). A multivariate test means Impakt can create up to 16 different combinations of a web page and discover what the very best combination is. In order to measure these test types, we integrate Google Optimize into Google Analytics, so that conversion and revenue goals can be set up easily using enhanced ecommerce.
Google Optimize is capable of serving different versions of websites depending on the following factors:
Your next question is probably: what are the variations? What do you actually change in these tests? If that is what you are thinking, the best way to explain this is to go through one of our case studies.
Sidenote: Social Media campaigns are extremely effective with UTM CRO because you can serve pages tailored to what you are targeting. A unique UTM can be created for each ad and each target will receive a different variation.
A motorsports online retail company came to Impakt with about 10,000 unique visitors every month. Their desktop conversion rate was about 3% and their mobile conversion rate was under 1%. Their product is extremely complex to order, since it is customizable. Here is the result of just one week of conversion rate optimization:
First Week Summary
3 experiments were conducted to overcome objections observed and recorded. The two main focuses were to increase the conversion rate of the product pages and increase the checkout completion rate. Our hypothesized objections were that people were just growing instead of purchasing and that shipping could be too expensive.
The first experiment added the promo code to the page. The line used was:
“This Week Only: Use Promo Code 10OFF to get 10% off your graphics kit!”
If the customer did not act “this week” then they would miss out on the 10% deal. The vagueness of the timeline meant the user didn’t know when the deal actually ended. It could have been today, or it could have been that weekend. Switching out the promo code every week proved helpful, especially since the text is vague enough to alter.
Since we wanted the codes to be effective long term, we had to truly disable the codes after a week. If a fence-sitter was not convinced when they saw the 10OFF code, then they returned and failed to use it a week later, they were much more likely to convert when they saw the deal again! We had to maintain honesty with the code timings, so we had to disable 10OFF and replace it with a new code the next week for select users.
The second experiment added a logical explanation for the shipping: “[client name] does not mark up shipping fees. You are seeing live prices straight from the carrier.” This message was added right under the shipping price on the checkout page. Impakt tested the simple explanation because most of the users surveyed said that their biggest gripe about [client name] was shipping fees. The client defended themselves to us, saying that they do not mark up the shipping. That sounded reasonable to us, and as you will see, it did to their customers as well.
This experiment produced more conclusive results. Out of about 2000 visitors, 1000 were given the original checkout experience, and 1000 were given the variation Impakt created. Out of these 2000 individuals, users that were given the variant were 1.8% more likely to convert. This difference went up as the sample size got larger. The recommended phrase to publish was, “[client name] does not mark up shipping fees. You are seeing live prices straight from the carrier.” This was the combination of words that proved to be most effective.
Week One Suggestion
An idea we had to increase the mobile conversion rate was to display a subtle progress bar as the user customized their graphic. We attached a mockup to make sure the communication of what we wanted was clear:
Since the user had to scroll down a good bit to finish customizing their product, the hope was that seeing the progress would help them continue. It was doubtful that it would hurt anything, but that’s why we test first! With CRO, you can’t make any assumptions.
This post got their team talking internally. Once they saw it, they started coming up with their own hypothesis to test! Not only did they start to understand the value of CRO, they started having internal meetings to make this a priority. To this day, they still bring me new ideas that are even better than ours. This client has an excellent method of internal communication, and has been extremely receptive to our test ideas.
Fortune 500 companies all use CRO on their website. Why? Because data matters more than gut feelings. Best Buy does not make changes to their website without testing them first (not even tiny changes). Until now, the client has made website changes based on instinct. Impakt CRO allows the client to mitigate risk by guaranteeing that all changes to their website have been tested and proven to be more effective than what they already have.
If a business is working with us, they are usually already successful. For example, if their revenue is only at $100,000 per month, but their conversion rate on their e-commerce website is 1%, and if we increase the conversion rate by just 1%, we have essentially doubled their revenue! If we can only increase their conversion rate by a mere .25%, we just made them 25k/month. Think of all the good they can do for their community and employees with that money.
Every website is a leaky bucket in some way. CRO plugs those leaks and pushes more people into converting. CRO is very much like the sales process Impakt Account Executives use. We identify the most common objections to clients signing on with us, and we find proven ways to deal with those objections. We employ the same tactics for CRO as we do for these outbound sales meetings. Sometimes we test a CRO hypothesis and it doesn’t work the way we expect. We then return with a new hypothesis based on our findings, and it works. The key is to have an open mind, patience, and to assume nothing.