Peep talked about how proper user research is the foundation to a testing program and so I’ve really focused on this section and understanding it in depth.
Talk to Customer Support — talking to the people that are picking up the phone with customers that have questions before they buy or after they buy will bring invaluable insight. Also, be sure to include the people that are emailing and chatting with customers. Ask them the top 10–20 questions they get asked. Also, ask them what they tell the customers that alleviates these concerns. This can really help inform your copy that will then alleviate concerns before they happen and also reduce the load on customer service.
Web & Exit Surveys — Customer surveys ask questions of people who have bought from your site, and web surveys ask questions from people while they’re on your site. The primary purpose of this is the same: identify sources of friction. Customer surveys ask about it after they experienced it, and web surveys ask them about it while they’re experiencing it.
These surveys are pop up boxes that appear based on certain rules. You can trigger polls in a few different ways.
- Time spent on page (10+ seconds)
- Above average engagement
- Upon exit
Visitors do hate these, but you can’t improve the user experience for the user unless you find out what bugs them about it in the first place. Not all visitors will see these though, it’s well worth it (incredibly cost effective way to gain insight about the user experience vs. other methods) and there’s no other good way to get the same kind of data, it’s temporarily up, and it’s become common across sites (I see them on Home Depot all the time).
There’s no perfect question to ask. The root of what we’re trying to find out are these two things:
- Why did they come to the site? Does our site match their needs? If not, are we attracting the wrong traffic, or are we underutilizing an opportunity here.
- What are the sources of friction? This is more specific than “why they didn’t buy” although understanding that is our main objective, but we have many goals to understand the big picture.
Ideas for questions:
- What’s the purpose of your visit today? (establish user intent)
- Why are you here today? (establish user intent)
- Were you able to find the information you were looking for? (identify missing information on the site, best asked on product pages)
- What made you not complete the purchase today? (Identifying sources of friction — ask this only during checkout pages, and beware that some people are still considering completing the purchase)
- Is there anything holding you back from completing a purchase? Y/N (ask for explanation and identify sources of friction)
- Do you have any questions you haven’t been able to find answers? Y/N (Identify sources of friction and missing information on the site)
- Were you able to complete your tasks on the website today? Y/N, and when No is selected, ask “Why not” — identifying sources of questions.
Make sure you frame your questions with urgency. Instead of saying: “What is holding you back from purchasing?”, say “What is holding you back from purchasing right now?”
Also make sure you frame the questions related to the goal for the page. For example, if it is an ecommerce site, you should poll the product page and identify the sources of friction specific to that page? Example: “What is keeping you from adding the product to the cart today?” Also make sure to only run one poll on the site at the time. It would get annoying real fast to go to an ecommerce site, get a poll pop on the home page, then the category page, then the product page page, then the checkout page… you get the point. That would drive your customers nuts. Run one poll with one question at a time to gather the data you need.
You can expect to get 2–4% response rate and run the poll until you get to 200 responses. If you have a low traffic site, something is better than nothing. If you are seeing the response commonly that it is “too expensive,” it really has to do with either being a wrong target audience or you’ve done a poor job of communicating the value proposition.
You are going to get trolls that give you stupid answers, so don’t mind them. People want to be comedians or trolls and it’s the internet, which has them on it.
The best ways to get more people to respond:
- Ask a single free format question (the more you ask, the less answers you get — I don’t recommend asking more than 1 question at a time).
- Ask a simple yes/no question, and ask for an explanation once they’ve answered it.
The second option works the best for CXL around 70% of the time. The theory is that it is easy to answer a yes/no question and then asking a follow-up open ended question will be more likely to get an answer. This is a psychological trigger called “commitment” from Cialdini, which is a trigger that people have to complete the path they started, and answer the question. An overwhelming amount will answer yes or no, and write a short comment because of this.
Some ways of wording the survey will work better than others, so you will need to experiment with it. In an example that they gave, there was the same amount of impressions for the survey pop-up, but one survey got 182 responses and one got 517 responses.
- Losing — (182 responses) “Is there anything holding you back from making a booking? Y/N”
- Winning — (517 responses) “Do you have any questions you haven’t been able to find answers to? Y/N”
In another experiment they asked the following three questions:
- Why didn’t you complete the purchase today?
- Is there anything holding you back from making a purchase today?
- Do you have any questions you haven’t been able to find answers to?
The second question performed overwhelmingly better than the others. Make sure you’re experimenting even with the questions that you ask. Every site and every audience is different.
The three tools recommended to try out are Qualaroo, HotJar, and Webengage.