The Definitive, No-BS Guide to Increasing Conversions (PHASE lll)



heuristicanalysisWPStep 7: Heuristic Analysis

Up until this point, we have been working without a clean set of data.

However, while we were working through the previous steps, a beautiful, clean set of data was building and is now ready for exploration!

Online Customer Surveys

Before changing any code, read over all of the returned customer surveys to get a feel for the overall response. Write-up a short report to document the common trends and the actions you plan to take.  See this blog on how to categorise survey responses.


Categorise responses into six to eight buckets based on trends that are presented in the results. Breaking the responses into categories will make it easier to perform analysis and identify specific issues that are having the greatest impact.


I recommend doing this type of research in one to two-hour stints to ensure that each response receives the same amount of effort. I am big fan of the Pomodoro Technique, especially for focus-heavy tasks. I have a huge list of tools that I use to keep my productivity at a rate that works for me. Vitamin R 2 is one of those – to see the rest, check-out this post.


As you read through the returned surveys, answer the following questions:

  1. Are there any common trends?
  2. Do I really understand the problem and does this page or site solve/answer that problem?
  3. Am I focusing enough on the end-benefit of the product or service to the user?
  4. Can I get some testimonials and/or case studies to target specific objections to the sales?
  5. What is currently working really well on the site (and should be left alone)?
  6. What is not working?
  7. How much comparative shopping are people doing? Should I create a page comparing the site’s products to competitor products and include a compelling case for purchasing from this site?
  8. What are the main sources of friction associated with completing the purchase and how can I overcome them?
  9. What information is missing from the site that needs to be there?
  10. How does this market segment like to buy?
  11. Is there anything market specific that I might have missed?
  12. What voice and language is the market using?
  13. Are there any insights as-to the emotional state of the visitors?
  14. What are the common characteristics of this market segment?
  15. How do the respondents describe the problem, the solution and the desired benefits?


After thoroughly reviewing the surveys, generate a word cloud to add to your research report. The word cloud clarifies common language and confirms research results. Use the report to build personas and hypothesize about what to test and why.


Building buyer personas is an in-depth process that requires looking at a variety of factors and weighing the importance of each to determine what motivates your ideal buyer. While building a persona is not covered in this guide, here are a few resources I highly recommend:


Live Chat Logs

Similar to the customer survey responses, review all of the replies looking for commonalities and answers to the questions listed below.


Use the identified commonalities to categorise the responses and identify trends. Add words that are often repeated in the responses to build a word cloud using Word It Out. The word cloud will confirm your identified trends and simplify your analysis.


It helps to categorise the chats by code. Use categories that define what is important and should be tested.


Reading through the chat logs will answer the following questions:

  1. What are the top pre-sales questions? (this is super important!)
  2. Which pages are generating questions?
  3. What are the most common questions being asked? Are they usability issues?
  4. Which products and services are people asking questions about?
  5. What are the main questions, doubts and hesitations?
  6. Which of the live chat answers have resulted in the visitor taking further action?
  7. Which page is getting the most live chat engagements and which page is getting the least live chat engagements?
  8. How is the live chat engaged traffic affecting the conversion rate of an individual page? (Google Analytics report)

Add results to the analysis report. Rank hypotheses against all of the others in the report. The goal is to identify which hypotheses are most likely to have the greatest impact on increasing conversions.


Note: Don’t forget the formula for writing a hypothesis that matters: “We believe that doing [A] for people [B] will make outcome [C] happen. We’ll know this when we see data [D] and feedback [E]


For Example:

We believe that creating two to three referral programs for highly active users of Acme Inc will increase the MAU by 10% per month. We’ll know this when we see that MAU’s have increased by 10% per month as a result of referrals and a higher use of referral codes.


Web & Exit Surveys

As with the survey responses and chat logs, collect all of the information, review it and categorise it into codes.


The web and exit surveys reveal:

  1. Did the site match their needs? (Are we attracting the wrong traffic or offering the wrong product for that market?)
  2. Is information missing from the site?
  3. What is the intent of the user? Is it to buy, look around or get more information? (establish user intent)
  4. What is holding the user back from making a purchase? (sources of friction)
  5. Was the user able to do everything they wanted on the site today? (sources of friction, usability issues and missing information)

Add the results to the research report and rank each based on the impact it will have on increasing conversions to determine which findings will be tested and implemented.


NOTE: When writing this type of survey, follow a yes/no question with an open text box that allows the user to explain their response. It is also a great idea to give something away free at the end as a surprise thank you gift. It is best to make the gift a surprise so that respondents are not tempted to breeze through responses just to receive the gift.

A great article from the Hotjar team on building surveys can be found here


User Testing

Google Analytics will tell you which pages are having issues; but when you try to determine the cause it may be clouded by your opinion of the site.


This is why it’s important to do user testing to identify bottlenecks.


During user testing, you will observe how real users interact with the website. Pay particular attention to the way they navigate through the site and individual experiences to identify common patterns.


User testing should be conducted at the start of the optimization project, when a re-design is complete and/or when a critical aspect of the website design is complete.


The best results come from testing with an actual target audience; however, if this is not possible, use a tool like


During user testing, ask:

  1. What is the website trying to accomplish? (Is the value proposition strong enough?)
  2. Shop around for a <product/Service> you like? (Are there any tedious tasks or is shopping difficult? Can it be made easier?)
  3. Find <product type> <product name> in <parameters>, under <price range>. (How do the users use the filtering and search features to find something specific?)
  4. Find 3 <product category>, evaluate the selections and pick the best one for you. (How hard is it for users to compare products?)
  5. Find <specific product/service name>. (How do the users find the product or service?)
  6. Start and complete the checkout process (or lead process). (Is there any friction? Can it be optimized?)

After the test is complete, ask the user:

  1. How was your overall experience?
  2. What was the best part of the experience?
  3. What was the worst part?
  4. What would you change?

Remember that testers are not buyers; and if you’re not using a target audience you will only get a rough idea of how your actual audience will interact with the site.


For Example, technical users like all the information in front of them on one screen as opposed to doing multiple searches across different screens. This usually goes against all design best practices; however, it works for that market.


When user testing is complete, watch all of tests again and take notes on how the users interact with the site. Look for areas to optimize for usability and areas that could be streamlined to increase conversions. Record recommended changes on the Research / Analysis Sheet.

See Phase IV: More Analysis here.



The Definitive No-BS
Guide to Increasing Conversions

How to Kill It in the Conversion Game

100% Privacy. No BS. no games, no spam, When you signup.

we’ll keep you posted with a few emails per week

Joshua Bretag is a data-driven growth marketer that manages over 6 figures in advertising spend per day. He has a passion for data driven marketing and analytics that has taken him to working large multi-channel marketing campaigns.

Next ArticleThe Definitive, No-BS Guide to Increasing Conversions (PHASE IV)