Quick summary: A man enters a clerk shop disguised as a normal employee and presents himself as the new freshman that is in need of training and help to know where to start. The shop assistant got the memo saying this guy would be starting today and she frowns upon the fact he is already late. After […]
A man enters a clerk shop disguised as a normal employee and presents himself as the new freshman that is in need of training and help to know where to start. The shop assistant got the memo saying this guy would be starting today and she frowns upon the fact he is already late. After that they get going on the job that involves tasks such as cleaning, answering customer requests and even selling. By the end of the day the man has utterly failed at even the most basic tasks and the shop assistant gives him a heads-up that he will definitely need to step up if he wants to keep the job. He nods and suddenly reveals himself as the managing director of the brand.
The description above is for a famous TV show where CEO’s and managing directors of large corporations accept the challenge of acting as a employee for the day in order to understand how things are going at even the lowest of their ranks. They do this because, well partly as a publicity stunt, but also because they want to see how clients interact with their brands and how employees react to the most common pressing issues of the day to day (without the pressure of the boss watching over their shoulders).
When you use software on your computer like Microsoft Word or Skype (so called Desktop software) you have to deal with software updates, restarting, installation & setup, licensing, disk space, privacy settings, permissions, etc. Obviously running a Website has none of such limitations but I would like to call on your attention something more obscure on your Website you can spy on people.
That’s right, on a Website you can track what every user does, sees or clicks – and learn from it. That is a huge advantage and one that companies like Google, Facebook, AirBnb are leveraging extensively. These companies have full time people doing experiments where they learn from user behaviour, implement changes based on hypothesis and test them out – sometimes a hundred times during the course of a single day. This is called A/B testing or multi-variable testing and they do this because they see the return on the investment and the impact it has. Can you imagine the impact just one of such improvements could have on your business?
Tools like CrazyEgg, Hotjar, Google Experiments, etc make it quick to get started doing this and can show you where users are getting distracted and falling short of reaching the goals you’ve layed out for them.
“Buy Now” or “Subscribe Now”? “Learn more” or “Read More”? What works best? Your mileage may vary from niche to niche and testing is the only sure way to find out the best call to actions and the best strategies to lead customers to a goal.
The tools for this sort of thing fall into some different categories:
1) Trackers: e.g. CrazyEgg: they show you where users are going with their mouse, where they click, scroll, etc. They usually store user sessions in a video format that you can play out.
2) A/B Testing Tools: e.g. Unbounce or Google Optimize: these tools allow you to set up one or more variants of a web page and then they redirect traffic to both of them and report back which one performed the best (with real data).
3) Analytics & Reporting, Mixpanel: these tools allow you to have access to important metrics like Bounce Rate, User sessions, etc. That can help you also understand better what is working best.
4) Live chat tools: these tools also normally track visitor behaviour and let you interact with users in real-time to answer questions or close a sale.
When doing A/B testing or multi-variant testing it is important to understand cohorts. A cohort is simply the group size you will use in order to test out hypotheses of what improves your website. Instead of relying your decisions based on gut instinct or feedback you got from one client you should aim to run experiments for a least a period of 2 weeks and a few hundred visitors. That will be your cohort – the minimum sample size before you take any conclusions/learnings/decisions about the matter.
Most of my clients get lost when I mention A/B testing. They have difficulty understanding the benefits and they surely don’t have the time to allocate to something new. Many times Web professionals also fail to explain to them that this is a very important step because they think it is too technical and they will never get it. But if you want the unfair advantage it comes with a price – you must leverage this tremendous power of spying on your clients and learning from that.