We covered a lot of ground this week in my website optimization minidegree from CXL. The most relevant material was the course on Landing Page Optimization.
It’s relevant because I’m building a new website at InglesConversacional.co – a website to help Spanish-speakers to learn English.
The pandemic set off a chain of events that lead me to this site.
First, my consulting work took a hit.
Then with little work to do I decided to spend more time on learning Spanish.
While surfing YouTube for Spanish learning videos I noticed that a lot of people were watching these videos. And I mean A LOT.
To the tune of hundreds of thousands or millions per month.
And many of these people were the same as me. They were stuck inside with nothing more to do.
So the creators of these free videos were making tens of thousands of dollars on ad revenue in a pandemic. So I decided to put my TubeBuddy subscription to work and started looking for opportunities in the English-learning market.
Needless to say there was some. And my new site was born.
So back to this week’s material.
I went through the course on Landing Page Optimization by Michael Aagaard.
Michael does an excellent job of breaking down the flow of a LP. The wireframes of different layouts was helpful too.
The most useful exercise was the one where he gets you to walk backwards through the process from checkout to the headline.
This made sure my home page was complete and answered, at least, the big questions someone might have when landing on my site.
I also used the CXL.com website for inspiration – specifically the pricing page. Keeping it very simple so as not to create any friction on the way to the checkout page.
For the copywriting part of the page, I went back to Momoko’s course for help.
I don’t have any customers of my own right now, so I scoured YouTube comments, and comments on Facebook for other courses as my voice-of-customer research.
In the past, I’ve tended to be long-winded when writing copy so I worked hard to be more succinct while still covering the important points. Only time will tell how well I do.
Below are my notes on this week’s material.
Landing Page Optimization
Quantitative LPO research using Google Analytics
One of the first examples of why GA is important when designing LP’s is mobile vs desktop traffic.
ME.co gets most of its traffic on mobile and therefore it’s ridiculous to design big pretty hero sections for desktop.
- Bounce rate is an important metric for landing pages because you want people to do more than just read the page and leave. You want them to visit a landing page, or other page in the funnel.
** This video seemed to suppose that I was familiar with the reports feature of GA. I am no.
- Start with device type. Optimize and build for the device that gets the most traffic.
- Then analyze where the traffic is coming from.
- See what the exit page is compared to the second page. If the last page and the 2nd page are the same, you have a problem.
- Look at gender/age data to see who is converting. If your main audience is male men aged 24-34 but they aren’t converting, you need to examine why not.
- Look at device usage and browser usage. Knowing which browser people are using can tell you how to test your landing page design.
- Another good report to look at is the Google Ads Campaign report. Look for instances where the ad spend is greater than the revenue generated. This can show you where potential problems exist, such as an ad pointing to a 404 page.
- Calculate your test capacity – how do you calculate how much traffic you need? Using the AB+ Test Calculator by CXL
Qualitative LPO research – part 1
Using the Qual LPO worksheet, do an analysis of the ad/email/source and the landing page.
Do it early in the process when you aren’t familiar with the ads, and landing page.
This is a good place to start forming initial hypotheses for testing.
Take screenshots to annotate for use as you go through the steps.
Then go into GA to confirm your questions about the ad and lp. E.g. Are people leaving the landing page at a high rate?
Rather than guessing at what people think about the company, go and talk to the people who talk directly to the customer.
Qualitative LPO research – part 2
Find a weak spot in the funnel and then use session recording to verify the problem.
Session recording is very useful because you’re observing REAL users go through your page.
You should have thousands of impressions when using heat maps, scroll maps, and click maps. Be careful of small sample sizes when drawing conclusions.
** Use this on ME top 5 landing pages to find out why they’re bouncing.
- Be specific in your questions
- Don’t give too many options (2-5)
- Find weak spots in the funnel and implement polls to better understand why users are having trouble/leaving.
- Get at least 100 replies before drawing conclusions
- Don’t hit people with a poll the second they land on the page – trigger on delay of 10-15 seconds (rule of thumb).
- Try using different polls on different devices.
- 5 second test. Makes sure people can tell what your page is about.
- in-person usability testing can be “overkill” for landing page optimization.
- Be careful not to hire “pro” usability testers.
- Use real testers from real target audience/demographics
- Give them relevant scenarios.
- Don’t lead the witness.
Landing page copywriting part 1 – How to write copy for landing pages
Landing page copy should do the following:
- Shorten the journey from click to conversion
- Follow up on promises made in the ad source (consistency)
- Speaks to user motivation and addresses barriers
- Answers important questions and creates clarity
- Creates a clear path to the conversion goal.
Example: Ashton Plumbing & Heating
- Value prop = $35 off any plumbing repair
- LP needs to have that value prop clearly shown.
- Design and copy need to go hand-in-hand.
- Don’t design the page before writing the copy. If you do, you risk not being able to convey all the information you need to convey.
LP Copywriting 2 – The 5 most important landing page copy elements
Here are the 5 most important landing page elements and their purpose.
- Headline – Match the message to the traffic source (ad). Capture their attention. Trigger dopamine.
- Benefits/features – Present important information. Emphasize the value proposition. Trigger dopamine.
- Credibility – Make the content trustworthy. Answer questions and address objections. Mitigate cortisol.
- Expectation manager – Ensure users know what to expect. Mitigate ambiguity. Avoid disappointment / negative reward prediction errors.
- Call-to-action – Make users click and take the next step.
3 Headline Formulas
- (Do something difficult) in (short amount of time) without (problem).
- (Do something difficult) in (short amount of time) and get (something valuable)
- Avoid (something frustrating) by (doing something difficult) in (a short amount of time) with (product x)
Features & Benefit
Features are product focused. Benefits are customer focused.
Here’s what the product does (feature). Here’s why that matters to you (benefit).
Your CTA copy needs to:
- Motivate the user to take action
- Give a clear idea of what happens when they click
- Be relevant to the conversion goal
- Start with a verb
- Focus on what they get – not what they will lose or not get (e.g. “Get this…” instead of “buy now”)
- Set realistic expectations
“When I click the button, I would like to _____”
Landing page design part 1 – the 6 most important design elements
- Expectation manager
- Use headlines or captions for videos to sell people on watching them.
Landing page design part 2 – visual hierarchy, the 6 most important aspects
- Visual hierarchy is making sure the most important information is seen first. Helps guide user visually through the landing page.
- 5 most important aspects:
- Colour and contrast
Landing page design part 3 – form design
- Your form has to make sense on its own – independent of the rest of the LP
- What is the form about (headline)
- Why should I fill it out
- What will happen after I fill it out. (expectation management)
- Top align the labels – makes it easier to scan the form
- Short forms can use the inset labels
- Forms have to work on all devices
For Brand New Businesses: Customer Development
Forget CRO if you’re a new business. Instead, focus on customer development – figure out which product to build and for whom.
Create a customer avatar, and find people who fit that profile. Then ask them questions to help you determine how to help them.
The Research XL Model
If you just go by gut feelings, you’ll screw things up. You need data to be able to make the right decisions about optimizing your website.
“What can I do differently based on the data?”
Step 1: Heuristic Analysis
Heuristic analysis is an experience-based assessment of the website.
When optimizing for conversions, it’s best to begin by understanding the user’s experience on your website.
Each page should be assessed for…
Step 2: Technical Analysis
The most persuasive website won’t make money if it doesn’t work on the device and browser that the user is using.
Check to see if there are lower than expected conversions on a specific browser or device.
Step 3: Digital Analytics
Google Analytics or something similar helps understand where the flow is stuck – where people are leaving.
What types of behavior equates to more conversions.
Step 4: Qualitative Research
Consists of surveys and user testing, as well as surveying new customers.
Step 5: User testing
Recruit people who represent your target audience when possible – but anyone is better than no one. Give them a task to complete and note how easy or difficult it is. Make the task as specific as possible.
Step 6: Mouse tracking analysis
Mouse tracking allows you to track clicks, and to a limited degree, how users are moving around a page.
What to do after research?
At this point, you have a list of problems or issues. The next task is to rank the issues on a scale of 1-5.
5 = severe issue costing the company a lot of money
1= minor usability issue
Two criteria are more important than others:
- Ease of implementation
- Opportunity score. How close to the money is it?
Measuring the effectiveness of a testing program
- Testing velocity. How many tests are you running per month or per year? The more tests you can run, the better.
- What % of tests that provide a win?
- What impact did your successful experiment have?
When doing a walkthrough, we need to answer these questions:
- Does the site work with every major browser?
- Does the site work with every device?
- What’s the UX like with every device?
HA is an expert based analysis that uses experience-based techniques for problem solving, learning, and discovery. Its results are not guaranteed to be optimal.
There are 4 main frameworks to choose from when doing HA:
- The 7 Levels of Conversion by Web Arts.
- Invesp Conversion Framework
- MarketingExperiments Methodology Heuristic Approach
- LIFT Framework
- Assess each page for clarity.
- Evaluate page relevancy. Does the page relate to what the visitor expected to see?
- Assess incentives to take action.
- Evaluate sources of friction on key pages.
- Pay attention to distracting elements.
- Understand buying phases. Are visitors rushed into a commitment too soon?
After going through the site, take your notes and go to GA to see if you can find data to confirm or contrast your data.
If no data exists, your job is to figure out how to get that data.
“I don’t know, but I’ll find out.”
Usability is about making a website easy to use.
Survey Design Theory
- Surveys allow you to…
- Understand attitudes towards our brand
- Stay aware of competitors
- Evaluate new products
- Understand our customers
- Quantitative questions are easy to grade using scales
- Qualitative questions are graded using a “zero-sum” grid – look for keywords to group answers.
- Well-written survey questions are:
- Not leading in any way
- Able to directly address the desired info
- Larger sample size minimizes errors in the data.
- Nothing less than 200 respondents on an A/B test should be considered.
- Not using intuitive scales (1-5)
- Mixing questions of behaviour with questions of attitude
- Questions that don’t communicate or don’t relate to your target audience.
- Making the survey too long
- Not keeping a neutral learning curve. As the survey goes on, people get a sense for what you’re looking for (or think they do), and begin to answer in ways that align with their perceived goals.
- Customer surveys
- Existing customers are always biased one way or the other towards your company.
- Keep customer surveys to 10-15 questions.
- Need at least 100 responses for the data to be valid.
- Don’t assume product knowledge on the part of your customer.
- Cognitive biases
- Reading the room. When the researcher tells the client what they believe the client wants to hear.
- Order bias. Items listed higher in the survey get a better score. To combat this, randomize the order of the questions.
- Biases on the user end
- Write summaries in Lehmann’s terms – no jargon
- Debrief the customer ahead of time.
- Selective perception. Customers pay attention to the questions that they agree with.
- Using the voice of the customer
- Speak with a few customers by phone or in person to get a sense of their dialects, and verbiage.
Customer Surveys via Email
The goal is to understand the friction they experienced during the journey to becoming a customer.
Should be sent out within the first week of becoming a customer.
Response rate will vary depending on your relationship with them, and whether or not you include an incentive.
(gift cards are a good choice for incentives)
Send survey of 8-10 questions.
What to ask
- What’s the purpose of this survey? (improve conversions)
- Only ask open-ended questions unless segmenting answers like gender, or age.
- Use the customer’s language from their answers to your questions in your copy.
- “What made you buy (sign up)?”. Avoid asking “Why?” because it forces them to defend their decision.
- “What doubts & hesitations did you have?”
- “What questions did you have but couldn’t find an answer to?”
- “What is the one thing that nearly stopped you from buying?”
- “How many websites did you visit before making a decision? Which ones?”
- Aim to get 200-250 responses
Talk To Customer Support
People who deal with the customer are excellent resources for finding out what the most common questions are, what the top concerns are, and what people want.
Ask them how they respond to questions and concerns. These responses can be used in your web copy.
Web & Exit Surveys
Show survey to people who have been on the page for at least 10 seconds, or who are about to leave the page (desktop).
Run one poll at a time – 1) Product page 2) Shopping cart
Expect 1-2% response rate.
On ME.co, could ask people if they found the info they’re looking for on the top 3 pages.
Live Chat Transcripts
- Make sure to save support chat transcripts. Can be used to help formulate copy, and determining what’s holding people back from buying.
- Pre-sales questions are golden for incorporating into website copy.
User testing is about gaining insight into the user’s mind, and about measuring usability.
It is NOT the same as usability evaluation.
Main benefit = Identifying bottlenecks
Jakob Nielsen determined that usability optimization lead to an 87% increase in conversions.
A/B testing is different in that people don’t know they’re part of a test. It’s done with live visitors on a live site.
A/B testing requires hundreds or thousands of participants. User testing can be done with very few – 10 people or less.
It’s a MUST DO for any new CRO project.
The minimum number of people needed to conduct user testing is 5. Maximum is 15.
10 people will find 95% of your problems. 15 will find them all.
Remember that testers are not buyers, and they have a different mindset when going through the website.
Only make changes when there is a clear trend among testers, and put their challenges into context. Don’t be too quick to run out and make changes just because 2 out of 10 people had trouble with something.
Mouse clicking software helps identify…
- Where people click and where they don’t.
- How far down they scroll on any given page (scroll maps)
Always start collecting mouse tracking data as early in a project as you can since it takes so much time to collect data.
- Click maps tell you if CTA’s are being clicked on.
- They tell you if people think something’s clickable and it isn’t.
- Scroll maps show you where people stop scrolling on your page. This tells you which info on the page is of no or little interest to your visitors.
User session video replays
User sessions recordings are critical for conversion optimizers. You can see exactly how real people interact with the website.
If there’s a specific page that has a problem but you’re unclear of exactly what that is, use session recordings.
A quick and effective way to identify where your biggest leaks are.
Focused on the depth of engagement – not pages viewed.
- Manually map out the funnel or flow of a customer to purchase in GA.
- Map out traffic for each page and see where people get stuck.
- Verify that the goal in GA has been configured properly.
How to perform a ring evaluation:
- Start with a manual walkthrough of the site to map out the URL structure.
- Count the unique pageviews per layer
- Check the numbers against what you see in CONVERSIONS -> GOALS -> FUNNEL VISUALIZATION (discrepancies may mean that the funnel is set up incorrectly)
Google Analytics Health Check
- Should be run before main project work commences.
- GA Health Check is a good deliverable to sell to clients.
- Need to check that what needs to be measured is measured:
- Ecommerce & even tracking setup
! Not performing health checks can result in using bad data to make decisions.
Funnels and Goal Flows
In general, the Goal Flow report is more flexible and most accurately reflects your users’ path before completing a Destination goal.
- Goal flows show loopbacks
- If someone skips a step in the funnel, the Goal Flow fills in the missed steps – Goal Flows does not. It shows the actual path.
- The actual steps are not shown in Goal Funnel reports
- Goal Flow reports are retroactive – meaning they go back in time to show you the flow data. Funnel reports do not. They start from the creation date of the funnel.
- Goal Flows allow segmentation of your data. Funnel visualization does not.
- Goal Flows allows you to compare two date ranges for all steps. Funnel Vis allows date comparison only for overall conversion rate.
- Conversions are calculated in Goal Flows by dividing the number of completed goals by the number of sessions.
- Data in Goal Flows is not as “fresh” as for Funnel Visualizations.
Reverse Goal Path
One way to reduce friction is to set up errors as goals. Then use the Reverse Goal Path to see what steps lead to the error being made.
Key Audience Insights
It’s important to know where your visitors are coming from and to see if some areas are converting better than others.
Behavior: Visitor Type
New vs. Returning visitor conversion rate … If returning visitors convert at a much higher rate, it may be because the site is confusing and requires more time to figure out.
On average, conversion rates for visits with site search is 2x that without.
NOTE: When comparing search to no-search, make sure to filter out bounces from the audience otherwise the search numbers will look artificially good.
Most common search pages are the home page and the blog page. Also popular is the 2nd page of the session – they couldn’t find what they were looking for and decided to do a search.
Content reports identify pages that work better than others.
Look for pages with high traffic and high bounce rates (Medicos Expertos rhinoplasty recovery page).
Compare them to the “good” pages and see what one does better or worse than the other.
Screen Resolutions, Browsers & Devices
Look to see which browsers and devices results in the most conversions and highest percentage of conversions.
If you see a particular browser that sends a bunch of traffic with few conversions, you should check the site in that browser to see what’s preventing people from converting.
Another thing to check…
Look at the conversion funnel in a horizontal layout. Evaluate dimensions such as browser, device or screen resolution.
You can often find cross-browser issues where people are getting stuck on a particular browser and not others.
Unbounce found that copy is 2x more influential in conversion than design.
Two things to test are:
- Clarity (clarity trumps persuasion) and completeness
How to use Copytesting
Ask open-ended questions of users about how clear a section or a page is.
- What’s your first reaction after reading this section?
- After reading this, what’s unclear?
Get quantifiable feedback using scale:
- On a scale of 1-5, how clear is this?
- On a scale of 1-5, how much value does this give you?
Using analytics to find conversion opportunities
Getting Started: using goals to quantify outcomes
Evaluating opportunities is more about comparing a page’s stats to other pages. If you find one that is severely underperforming, you can work to improve it.
Evaluating traffic quality
Using source/medium report, you can see who is sending traffic and how they got to your site.
#1 indicator of the quality of traffic source is the amount of revenue being driven by it.
Metrics that matter (and some that don’t)
- To get the PAGE VALUE metric, you need to have Goals or Ecommerce set up, or both.
- Knowing this value is important because it tells you how much each visitor is worth to you. This allows you to set budgets for PPC campaigns.
Secondary dimensions and advanced segments
Get your segments down to 5-50% of your total traffic and it becomes easier and more effective to analyze.
Bounce rate is often misleading because you don’t know what’s going on within your page.
Event tracking helps you know what a user is doing while on the page.