Master the Essentials of Conversion Optimization

Master the Essentials of
Conversion Optimization

9 Conversion Lessons
1. Mindset of an Optimizer
2. Conversion Research
3. Google Analytics for CRO
4. Mouse Tracking and Heatmaps
5. Using Website Surveys
6. User Testing
7. From Data to Test Hypotheses
8. Getting A/B Testing Right
9. Learning from Test Results
“If you can’t describe what you’re doing as a process, you don’t know what you’re
doing.”
-W. Edwards Deming

Mindset of an Optimizer
What It Means:
You need to get in the correct mindset to be successful. Conversion
optimization is the means by which we achieve growth it isn’t an end goal in
itself. Using best guesses and assumptions isn’t an effective way to approach
optimization you need to utilize a process where you identify where the problem
is and what caused it.
The Process:
1. Set Goals 2. Set up measurement and gather data 3. Analyze data
4. Turn data into insights 5. Turn insights into prioritized hypotheses
6. Test your hypotheses 7. Get data from tests 8. Back to data analysis

Conversion Research
What It Means:
We need to do research to figure out where and what the problem is. The data
collected needs to be relevant and lead to hypotheses. Peep Laja says, “we
don’t just need more data, we need better data.” He has a six step process
called ResearchXL.
ResearchXL:
1. Technical Analysis 2. Heuristic Analysis 3. Web Analytics Analysis
4. Mouse Tracking Analysis 5. Qualitative Research/Surveys
6. User Testing 7. Sum-up

Google Analytics for CRO
What It Means:
Approach analytics with a specific problem and way to measure it. Learn to
analyze the data and draw insights about your website. Some information is
useless so thinking through what you want to track is important.
KPI’s:
Key Performance Indicators are desired actions that your customers can take.
They are things like email sign up or purchase made. The raw data you get
from analytics are metrics that help you track your KPI’s.
Funnels:
Set up funnels so you can track the journey your customers took to a desire
action. For example if they visited the product page before purchasing you want
to know.

Mouse Tracking and Heatmaps
What It Means:
Everything done with the mousepad can be tracked. There are many services
that do this in different ways so review them and think carefully about what
you’re trying to learn.
Heat Maps:
A graphical representation of where users let there mouse hover. It is designed
to give insight into where they are looking but isn’t completely accurate.
Click Maps:
Shows where users click on your site. Set up link attribution in Google Analytics
to see which links were followed.

Mouse Tracking and Heatmaps
Attention Maps:
These show what users are viewing on their screen and takes into account
different screen sizes and scrolling. Allows you to see if consumers are paying
attention to your key information.
Scroll Maps:
Shows how people view your site in relation to scrolling. This allows you to
track how far down the page users view so you can place all the important
information in areas that are being looked at.
User Session Replays:
A video recording of people using your site. There is no audio but it can provide
insight into how people are navigating.

Using Website Surveys
What It Means:
Surveying customers to collect qualitative data about your site.
How to Survey Correctly:
First, it is important to have a good design. Follow best practices survey making
and ask customers about data that gives insight to a problem. Have questions
seek to discover what is keeping each consumer from completing the desired
action on each page (ie. make a purchase, sign up for email). Only present the
survey to “good users” or people on your site who seem to have the intention to
make a transaction. Finally, make sure you have enough surveys to have
statistically relevant information. It depends on your site but at least 200 is a
good rule of thumb.

User Testing
What It Means:
Observing and taking notes on volunteers as they use your website for the first
time.
Steps:
1. Create a set of actions each user needs to accomplish. Give one specific
task, one broadly defined, and one that forces them to complete a funnel.
2. Don’t ask for opinions or lead testers to a conclusion. Simply have them
use your site and comment on their unique experiences.
3. Recruit people who have never used your site in the past so you can learn
how usable it is to a new user.

From Data to Test Hypotheses
What It Means:
Take all the data you’ve gathered so far and draw intelligent insights from it.
Take each issue you find and place it into one of the following five groups.
Test:
These are obvious opportunities to increase growth that can use further testing
to identify what will work.
Instrument:
These are areas where better or more in depth analytics reporting is needed.
Hypothesize:
When something isn’t working but there isn’t a clear cause as to why. Come up
with intelligent hypotheses based on data and develop tests to gain more info.

From Data to Test Hypotheses
Just Do It:
Clear problems with obvious solutions. Just go in and make the changes
quickly.
Investigate:
There is a general problem but the cause isn’t clear. Do more research,
possibly across different devices, to determine what the issue is.
Prioritize:
Formulate some system for ranking which problems are most important and
should be addressed first.

Getting A/B Testing Right
What It Means:
There is no perfect answer to how to design your site. In order to be successful
you need to iterate and optimize through testing cycles and back up your tests
with data. A/B testing allows you to evaluate one decision against another to
see which works better.
Statistical Significance:
None of the A/B testing will work if you don’t ensure statistical significance. This
means your sample size needs to be large (350 conversions per variation) and
you must have at least a 95% confidence interval. Finally, the test should span
a duration that includes all relevant business cycles (weekends, weekdays,
blog posts, etc.)

Learning From Test Results
What It Means:
There is a good chance that the variation you run will lose to the original or
provide inconclusive data. That is okay as long as you learn from it and
continue to iterate. Make sure your hypotheses are sound and continue to drill
down on a solution until you find something that works. The goal of testing is to
collect data so failures are only failures if you don’t learn from them.
“Test is really data collection. Personally, I think the winner/loser vcabulary
perhaps induces risk adversity.”
-Matt Gershoff

Conclusion
Conversion optimization isn’t a set process that works everytime. Every site is
different and a solution that works for one might fail miserably for another. The
key to optimization is iteration. Dispel your preconceived notions of what will
work and test to discover what actually performs best. Follow these intsructions
and keep in mind…
“If you can’t describe what you’re doing as a process, you don’t know what
you’re doing.”
-W. Edwards Deming

Master the Essentials of Conversion Optimization

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Master the Essentials of Conversion Optimization

Similar to Master the Essentials of Conversion Optimization (20)

Recently uploaded

Recently uploaded (20)

Master the Essentials of Conversion Optimization