With so many options for data analysis and metrics to look at alongside the readily available tools to perform A/B testing on almost every element on the sales pages that lot of people do it wrongly.
With so many options for traffic analysis and metrics to look at alongside readily available tools to perform A/B testing on almost every element and every line on the sales pages that lot of people overdo it. To add fuel to the fire, many blog share those eye-catching results about how sales grew by more 1000% with just changing the color of the call to action button. I am sure you may have seen those and chances are you may have at some point of time, opted for a different colored button in your design or at least may have thought about it.
The focus with data driven approach is to make the most out of each visitor that lands on your site or sales page. The aim is to make the most of out of the design elements it may improve sales numbers but what’s more important is to improve the profitability. At times too much of data can cause you more harm than it can help.
It’s not only important to use the right kind of data but also use the same in the right context. Lets see when you can safely ignore such data.
1. High Cost Of Data
There is always a cost attached to the data that you may not always realize. Yes there are free analytics tool like Google Analytics but you still need time to create goals and experiments and then implement those goals and experiment on your site. You may not always be able to implement those yourself and may need to hire people to get those done. Even if you are developer, you may be better using the time somewhere else.
The above process may need to be done for each experiment and though we don’t realize, there is an inherent cost attached to make the most use of those free tools.
Apart from setting up the goals and experiments, you may not be expert at analyzing those data or it could so happen that you may have too many experiments running to make any conclusive decision on the data available and may need to hire data mining expert to give you the right picture of the data.
The ongoing process of experimentation can add lot of cost that you may not realize. Apart from experimentation costs, there are so many other analytics tools that are not free to use to gather data and the cost to those analytics tools may be lot more than the actual benefit you may end up after the experimentation with sales.
2. Too Long to experiment
At times it becomes too lengthy an experiment when you have too little to experiment with. If you are not focusing on good sales volume and experimenting with little volume of data, it may end up taking too long for the experiment to run and fetch the data that can be conclusive.
As an example if there are only 5 to 10 sales per week visitors on a site, don’t focus on making more sales from the same traffic number, instead try to focus on bringing more traffic and users to your site and once you have good number of traffic and sales from your site per day, you can work on experimenting and increasing the sales. Yes effort of bringing the traffic may not be converting well for now but then it would make much more sense to be testing once you can experiment in very short span of time and conclude really fast to make the most out of your time and effort.
As an example if you are making 5 sales from 500 visitors, instead of experimenting with A/B testing and making 10 sales from the same 500 visitors, it can be lot easier to just bring 1000 visitors to your site and make 10 sales as of now. Once you have good traffic numbers, you can always make those experiments to see what works more for your kind of users and traffic.
3. Data That Doesn’t Matters
Yes you have heard it right. There is lot of data that actually does not matter and at times we are just stuck with such data. Data mining is a big field of study and just knowing that Red button generates more clicks than Green button for some blogger does not mean you should be making all the buttons as red on your sales page.
Actually just knowing part of the metrics can be lot horrifying.
The main aim of the sales page is not to make users click on call to action button but what matters is the action that user needs to take once they clicked the button. Did more users complete the call to action for the red button or green button is more important metrics. Red button may be lot more catchy to the eye and so instead of reading what you offer, people just clicked on the button without even realizing what they are doing and then just closed the window when asked for payment. Green button on the other hand was almost hidden (not literally) but people found the button when they read what is being offered and so made more purchases once they clicked on it. So focus should be to use the right data.
4. Tests That Make You Uncomfortable Implementing
You know something always work but then are not comfortable implementing it. It is better to not experiment with such elements or else you may be tempted to be implementing it and may not feel good about it.
The classic example can be popups or exit popups which are annoying for everyone but if you test a popup window or an exit popup message, you may see lot more call to action but once you have the data, what if you are not comfortable implementing it on your site?
5. Tests That You Don’t Care About
A classic case is when there is complete makeover to site’s design. You don’t want to be opting to the new design because it may just fetch you better numbers but it is something that you feel your users will like it. You don’t feel like running a side-by-side test of old and new design.
In a complete site makeover there is very little you can conclude with the test because there are lot of things that can go into making the old site perform better than the new one and then you don’t care because you want to be making sure your new design is what you want. The new design can always be experimented further to optimize and perform better. Being always data driven may not work and at times you have to work on your gut feelings and ignore the data.
Conclusion
According to me experimenting with data and metrics is something that is for established businesses who have maxed out the options to grow traffic or at least are on such levels where the effort to grow further is more costlier than experimenting and making the most of the users that they have build over time. Are you working on building your blog or just trying to experiment too much and going nowhere? Share your views in comments below.
Readers who read this article also read …
