This week of my course with ConversionXL’s CRO minidegree was a good one.
It started off with User Research and how to gather information from the end-user to build a better experience for them on your website.
I attempted my own user research – albeit in a very limited way – with my own website, inglesconversacional.co.
Having already created my sales funnel, I wanted to get “fresh” eyes on the website to get feedback.
I enlisted the help of some of my wife’s family, who are Colombian and have shown interest in learning English.
Using Zoom, I shared my website with them and asked them for their initial reactions. Then I asked them to go through the funnel – giving me their thoughts as they did.
I know my first attempt wasn’t perfect but it did reveal some things about the site that could be improved, and clarified.
The big change was the colour used. It can be described as Pepto Pink – the colour of Pepto Bismol.
It was the first thing that each person commented on. And even though they were women, they hated the colour.
So I changed the colour of the website.
The next course was Google Tag Manager for Beginners, by Chris Mercer.
I’ve said it before and I’ll say it again…
Chris Mercer is an excellent presenter. He maintains a quick pace but isn’t too fast. He is enthusiastic about his topic and makes it really interesting.
I was extremely excited to go through this course. It was clear from Mercer’s course on Google Analytics that GTM plays a huge role in collecting data for your website.
After going through the material, I was not disappointed.
The power of GTM is astounding and is definitely something I’d like to continue to learn. Although, with all of the other information to absorb, I don’t know when I’d do it.
Using the information provided in this course, I have replaced the plugins being used to pass website info to GA with the GTM.
I’ve also replaced the Facebook plugin with GTM and will continue to learn how to apply GTM to capture even more Facebook data. That is one area I will definitely spend time on.
Facebook has a plugin which seems to do this more easily but I’d like to reduce the use of plugins if at all possible.
The other thing I have planned to do is to install GTM on MedicosExpertos.co and set up some tags for measuring the interaction with my pages.
The top 5 pages have very high bounce rates, and I want to see if people are actually doing anything on the page, or if they’re just landing and leaving.
The pages are very specific so it might be a case of them coming to the site … getting what they want … and leaving.
The next course on creating Quick and Rigorous Personas was not as interesting, or relevant to me. I’m never going to invest the money or time to do what was covered in this course.
The presenters are obviously experts in their fields but the presentation was a little dry.
Finally, I covered Heuristic Analysis Frameworks For Conversion Optimization Audits.
I found this course really interesting. I love seeing how to break down an existing website, make changes, and how they can be improved.
It was also stunning to me to see how bad the websites of big companies are. The checklists provided to take you through the 7-levels of conversion was excellent. And the method used to objectively score a website on its conversion-ability was excellent, and something I’ll come back to – not just for my own sites, but if I begin working for clients too.
As always, my notes and key takeaways from each course are below…
Why everyone should be doing more user research
- Helps build better experiences for the user.
- A lot of websites don’t take into account how people with colour blindness will see your site.
- Almost 75% of people score “poor” in their skills to use technology.
- User research focuses on understanding peoples’ goals, needs, and beliefs.
- Asks meaningful questions, and thinking critically about the answers.
- Takes no biases, beliefs, or assumptions for granted.
- Reduces risk, and saves money by eliminating subpar options from early decisions.
- Identifies opportunities to delight customers.
User research is not…
- Usability testing
- Always right
- Buyer personas are not always the end-user personas – especially in B2B.
- Market research is often done on a larger scale than user research.
How different research techniques will help you answer different questions
- Use the 5W+1H to develop questions
- The less you know about a situation, the more open-ended questions you should use.
Never begin with a method in mind. Always start by stating the problem, and what you want to learn. Then choose the method or methods that will help you achieve that goal.
Question to answer:
What do people feel most challenging about learning English?
3 Research Questions
- What methods have you used to learn English?
- What is the most challenging aspect of learning English?
- What one thing do you think would help you most in learning English.
Listening to what people say
Being good at performing customer interviews requires that you be good at having a conversation – especially listening.
Asking questions is a key skill to practice
- Ask questions with both the + and the –
- Eg. Were you successful or unsuccessful at finding the help button?
- Keep your language neutral
- Avoid asking for more than one thing – eg. What are the things you like and dislike?
- Avoid jargon
- Don’t assume your user can predict future behaviour
Errors made when surveying
- Coverage error – not able to survey a representative sample of the population you are studying.
- Sampling error – you are not measuring everyone in the target population – only a portion of them.
- Measurement error – the way you choose to measure something leads to inaccurate data.
- Non-response error – some groups within your survey complete the survey less than others.
Watching what people do
You’re going to learn from each one of the tests you do and improve as you go along.To do that, you ideally should see things shift. So will user satisfaction go up? Does it take less time as we make improvements to the interface for this person to get to their goal? Whatever matters for your goals as an organization,you should be tracking these and hopefully seeing meaningful differences as you’re making changes to your tests,to whatever you’re testing.
When doing user testing, keep these things in mind:
- Build realistic scenarios and tasks
- Keep test time to less than 30-minutes.
- Don’t give users clues or hints
- Avoid selling
- Keep things neutral and inoffensive.
What areas would I like to test, and improve?
With my new project, Inglesconversacional.co, I would want to test the onboarding process once someone becomes a new member. Getting them involved quickly, I believe, is crucial to keeping them on as a member.
How to conduct research with the right people
Who you need to recruit is dependent on your goals.
E.g. Improving onboarding process needs new customers. Improving shopping cart conversion requires people who would buy your product.
Finding people to participate in the testing is the biggest challenge of most researchers.
An important thing to remember when recruiting participants in your research is to make them feel valued and important. They are giving their time to help you.
Yes, you can and should compensate them. If they do it for free, they may provide different feedback than if they are compensated. By compensating it, you make it a time-for-money relationship and will likely get more realistic feedback.
Quantitative testing requires a lot of people to participate and therefore is very expensive. Only conduct quantitative testing when you have a clearly defined goal in mind.
When doing usability testing, 5 people will uncover about 85% of the problems.
How to synthesize your research
Data does NOT equal Insights
Data needs to be tied to hypotheses.
Insights often come down to data organization and logistics. You need some way to process data and information and then make sense of it.
Use a synthesized network to look for patterns and meaning in the data.
Use sticky notes to find emergent patterns. Each separate idea gets its own sticky note. Colour code the information. (Could use trello with labels)
Synthesize data to support the riskiest decisions.
Creating memorable, useful research deliverables
Creating the artifacts gives different people in the organization the info they need to know.
You want to distill the information to give them only the most salient points they need.
Rather than use the word “persona”, a better choice might be “archetype”.
Personas create a shared vision of their end-users, which allows people to focus on a smaller number of user needs.
Personas focus you on the most important points.
Google Tag Manager For Beginners
Your Intro to Google Tag Manager Basics
What is a tag manager?
- To help you get visibility into what’s happening on your site.
- E.g. Did a user interact with a video?
- How far down the page did they scroll?
- Take purchase info and forward it to FB, GA, etc.
- Becomes a control panel for all pixels and tracking code.
GTM vs GA?
- GA collects, stores, and reports info.
- Tag manager replaces GA for collecting data. GTM data then goes into GA to be stored, and then reported on. It can also send data to FAcebook, and other third party platforms.
What can GTM do?
- If you have a script that needs to be in the head or the body, you can use GTM to install the script on your site.
- It’s NOT good for split testing scripts like Optimizely.
- “Container” is similar to “Property” in GA.
Getting to know GTM
A tag is a script that sends data back to the platform (GA, FB, etc)
One tag could give conversion info. Another gives info that someone was on a particular page.
Getting to know GTM: Triggers
What’s a trigger?
Getting to know GTM: Variables
- A variable is Info that GTM needs to do its job
Getting to know GTM: Data Layer
- A data layer is like a virtual filing cabinet where details are temporarily stored for GTM to use.
- When GTM loads, a data layer is created. Once another page loads, a completely new data layer is created. All the old data from the previous page is dumped.
- Two parts to a data layer:
Value = play
Key = video name
Value = all about widgets
Key = product name
Value = widget
What kind of data layers would I want for ME.co and IC.co?
For IC.co, I would want to know:
- Logged in or not
- Scroll depth
- Watched a video? If so, how long?
- Clicked a button? Which one?
For ME.co, I would want to know:
- Scroll depth
- Clicked a button? Which one?
This would help me determine why they’re bouncing off my top pages with such regularity.
Getting to know GTM: Organization
At the Account level, people you add can be one of two roles: 1) Admin or 2) User
Admins can add and DELETE containers, so vendors, or employees should ONLY be Users.
Containers should include all domains used for one entity.
- Infusionsoft.app (payment processor for maindomain.com)
Hierarchy for naming — Source/Platform – Type of Tag (Event) – Thing being measured
Help you organize tags by vendor or source/platform (eg Facebook, GA, etc)
Getting to know GTM: Preview Mode
Preview mode allows you to preview your site before making any changes live.
Can share the preview with anyone by sending them the preview link.
Changes made to GTM will not be reflected in Preview mode until you refresh the Preview mode.
Getting to know GTM: Workflow
- Start with workspaces
- Make your changes
- Preview and test
- Publish the changes
- Once published, you’ll end up in “Versions” tab where you can see a history of all of your implemented changes/additions.
- Workspaces are temporary, protected places for you to go in and make changes.
- After making changes, you must update the default workspace. This merges changes made in other workspaces with the default workspace.
** Workspaces are really only helpful if you have multiple people working on the same container.
Getting Started: Creating your first tag
After creating the tag, make sure to…
- Enter preview mode
- Confirm that the tag is firing
- Confirm things are working in GA live
- Then publish the changes.
Getting Started: Scripts and Pixels
Followed the instructions in the recommended Facebook article to add the Facebook pixel to my site using GTM.
Tracking Engagement: Clicks and time
When a person clicks on a link, the link text is saved and passed through. You can use this to create other tags…
“If the link text includes ‘monthly’, fire this retargeting tag to Facebook.”
Clicks on links that are not hyperlinks such as share buttons will not show for the “All Links” trigger. To record clicks for these types of buttons or links, you need to set up a trigger for “All Events”.
With “All Events” ALL CLICKS of the mouse are recorded.
Using a Timer trigger you can create an Event of “Hyper” engaged people by saying…
- If someone clicks a link after 20 seconds on the page, tag them as HYPER.
Tracking Engagement: Scroll
Scroll events can mess up your bounce rate unless you set the “Non-interaction Hit” to “True”.
Tracking Engagement: YouTube Videos
- Turn on the video variables
- Add a trigger for YouTube video
- Default = start and complete
- And choose “Progress” in time or by %.
- Time is the video time not elapsed time. So if they skip through to minute 10, it shows they’ve watched 10 minutes.
- Choose to make it an interaction hit because they physically have to click a button to play the video.
Data Layer 101: Storing Details
- Data layers are storing data with keys and values.
- No comma at the end of the DLP code (last line of script doesn’t need a comma)
- No semicolon at the end of the script
Pro Tip: After entering DLP key and value, enter a “event” with value that describes what’s happening.
Data Layer 101: Reading Details
- Create a GTM variable called “Data Layer Variable”
- DLV Name = key (case sensitive)
- Create “Custom Event” trigger
- Event = “characterUpdate”
Click “console” in dev tools. Allows you to see the data being pushed – even on other peoples’ sites.
Tracking Ecommerce: The Basics
Create product variables
- Create a variable in GTM – URL variable -> Query … for every product variable (price, sku, order id, etc)
- Goes into the url and queries parameters (params are case-sensitive)
Push variables into data layer
- Plugins already do this
- Create a Data Layer Push Tag (Custom HTML)
- REMEMBER to enter a line to change the EVENT to “ecommerceUpdate”
Create “Report Ecommerce” Tag
- Create new tag “Google Analytics – Universal Analytics” -> Transaction
- CAN ONLY BE USED WITH STANDARD ECOMMERCE IN GA
- Trigger = all “ Event = ecommerceUpdate”
Deep Dive: Cross-Domain Tracking
Cross-domain describes a customer journey that involves more than one domain.
- Add the secondary domain to GA -> Tracking Info -> Referral Exclusion List.
- Go to your main Google Analytics Settings variable.
- Click “more settings” and set up cross-domain tracking
- Enter all domains you want tracked, including your main domain.
- Go to “Field to Set” and choose “allowLinker”. Set it to “true”.
Deep Dive: Tag Sequencing
Timing matters when tags fire. Tag sequencing is about controlling the order of tags firing.
To set the timing of the tag, go to into the tag…
- Tag -> Advanced settings -> Tag sequencing
He keeps saying there are occasions to use this but really doesn’t go into any of those in-depth.
Wrap Up & Resources
Tags = what you want to do
Triggers = when you want them to happen
Variables = information you want passed to GA
Fast and Rigorous User Personas
What are User Personas
A user persona is a model which gives an approximate description of the customer and provides insight into their motivation and goals.
Info needs to be relevant and actionable.
Fast & Rigorous Framework
3 steps to make a persona:
- Collect data
- Identify groups
- Build the archetype
- Helps cross functional communication
You can build personas on numbers not collected through surveys.
Collect Quantitative Data
Data is the foundation of a good persona. They must be…
Setting up the survey
Go-to tool is Qualtrics (expensive).
Use a tool that allows open-ended questions.
Validating Data & Next Steps
Run a small version of the study before scaling it up to all people. Make sure the data provided is good.
Share the data with colleagues to get their input.
Once you have all of your data, simplify the data. Two ways to do that:
- Factor analysis
- Clustering analysis
Simplify your data
- Each row should be a response (one unique participant)
- Each column is a variable (answer to a question)
Exploratory Factor Analysis (EFA)
Each question is tapping into one of a few key underlying factors.
You don’t need to define the underlying factors beforehand.
The factor analysis quantifies the influence each factor has on each question.
You can assign a score for each respondent for each factor based on their answers to the questions:
16 questions -> 3 factors (value, expert opinion, value/safety/economy)
Respondent 1 scores: factor #1 (91) factor #2 (38) factor #3 (48)
How does the cluster analysis work?
Put your respondents into clusters/groups.
- Start with the cluster analysis.
- Layer on demographic info (age, gender, etc)
- Add an image that would be a physical representation of the persona.
- Include quotes from open-ended questions to help you build empathy for the persona.
Heuristic Analysis Frameworks For Conversion Optimization Audits
Introduction to heuristic evaluation
There are multiple methods for analysing websites – each with their own pros and cons.
- Voice of customer does not necessarily reveal the truth – people don’t always say what they mean or they believe.
- Data doesn’t tell you why something happens. Example: Optimizing landing page causes overall conversion rate to decrease – you’ve changed the overall experience.
You should always use both methods – Qualitative and Quantitative to understand the what and why.
SOR – stimulus, organism, reaction
Stimulus = website
Organism = User (brain)
Reaction = results
7 levels of Conversion
- Relevance – Is that the right page for me / my problem?
- Trust – Can I trust this company / vendor?
- Orientation – Where do I have to click / how do I find the right product?
- Stimulus – Why should I buy / click here?
- Security – Is it safe to do that here?
- Convenience – How easy will everything be?
- Confirmation – Did I do the right thing?
Websites have a “body language” that works on a subconscious level.
To improve relevance, always analyze where traffic is coming from. And then sync the website experience with expectations from traffic sources.
- User personas to think about desired values and emotions (Qual)
- Make an A/B test to find out what works (Quant)
When analyzing a website, ask…
- What emotions do you see represented? Do they emotions align with what was promised in the ad or source of traffic?
Trust and Orientation
Trust is a feeling, and can be evaluated in milliseconds.
Elements that increase trust:
- Information architecture (navigation)
- Social proof
- Trust seals (logos)
- Testimonials from celebrities or people of authority
Orientation is about helping to make it easier for people to make a choice and find what they’re looking for.
Help people overcome the paradox of choice.
Stimulating Users To Buy
You competitors are just a click away
On the product page, is there any reason why someone would choose to buy from you over someone else?
Users need a reason to stop their search and buy.
Security and Convenience
For ecommerce, you can provide security by…
- Explaining return policy on checkout page
- Answer questions about returns, such as “how much does it cost?”
- Provide testimonials below the checkout form to reinforce that they made the right decision.
Anticipate inner dialogue and questions they may be asking about what you want them to do.
Example: Why do you need my address in this form?
Ways to increase the perception of “EASY”…
- Clear hierarchy
- Light colours
- White space
- Big buttons
- Show good reasons for making the buying decision
- Use micro-feedback on pages and elements (e.g. inline validation on forms)
- Present good reasons for buying on the thank you page
- Use fun, and interactions to increase fun and give feedback
When talking to gatekeepers – who may not be open to your work – say “I’ve made an analysis of your website”, instead of “Your website sucks”.
Separate your personal opinions from your findings.
Decision makers love benchmarking.
— you are here, and you’re competitors are here