About the Guide
Day 1: Intro to Growth
Day 2: Customers
Day 3: Data
Day 4: Metrics
Day 5: Analytics
Day 6: Analysis
Day 7: Growth Priorities
Day 8: Funnels
Day 9: Psychology
Day 10: Conversion Rate Optimization
Day 11: Copywriting
Day 12: Landing Pages
Day 13: Acquiring Customers
Day 14: Pricing
Day 15: Trials and Plans
Day 16: Onboarding
Day 17: Retaining Customers
Day 18: Upgrading Customers
Day 19: Referral
Day 20: Keep Learning
Day 10: Conversion Rate Optimization
Conversion Rate Optimization (CRO) involves experimenting with changes to your website and product to achieve higher rates of conversion.
CRO allows you to get more out of what you already have. It’s one of the main tools of growth hacking.
CRO isn’t magic, though. It requires a lot of hard work to see gains and the process includes far more failures than successes.
You'll want to follow the Question-> Hypothesis-> Test-> Analyze process of the scientific method for your experiments. This simple structure helps minimize bias and makes experiments less subjective.
CRO is all about working with friction. The more work that person has to do to buy your product, the more friction. Reducing friction is meant to increase your numbers through a step and increasing it is meant to increase quality through a step.
To start prioritizing areas for CRO, you want to look at your analytics to see what steps in your key funnels have the highest numbers of people dropping out.
It can involve adding things that are missing such as information that addresses potential customer fears. It can also involve taking away things such as unnecessary fields in forms.
There are a number of tools that make conducting CRO easier. They vary from full on testing suites like Optimizely and VWO to general analytics software for watching the effect of experiments through the whole funnel.
Look at the conversion rates of each step of your "getting customers" funnel. Find a step with a steep drop off in conversion rate. Look at the page for that step and think of three things that could be causing friction.
Why people do what they do
Your analytics software will never tell you why someone converts. This is where talking to your potential and current customers comes in. Talking to them you can lead you to the real reasons why they want or don’t want your product.
Use your customer research to figure out what their fears are and what they want out of your service. Then incorporate that into your writing and include visual elements that increase their trust.
Reasons people won’t buy from you
You will never convert 100% of potential customers to the next step. General conversion percentages depend on each step potential customers take in your funnel.
Some of the many reasons they won't convert:
- Many people simply aren’t in your target market.
- If you’re targeting small businesses, you can turn off those who arrive at your site with enterprise needs. Sometimes people are just curious about a tool and have no immediate need.
- Other times, they’ll be performing research and your product will only be one of many they evaluate.
- You have to expect some people to be very interested up until your pricing page because your pricing is out of their range. Sometimes it’s even too low and causes concern for the potential buyer that your product won’t live up to their needs.
- Maybe that potential customer has a feature in mind that they can’t live without and you don’t have it.
When working on conversion, you will run experiments. The most common form is the simple A/B test where a control and a variable of an element are tested.
The more items you test (such as a headline, image, CTA button, etc.) the more complicated your test becomes. These are called multivariate tests where multiple page items are tested at once.
We always recommend you start with A/B testing and keep things simple in the beginning.
There are many tools for A/B testing; the most commonly used among startups are VWO and Optimizely. Both these tools are built to involve as little developer time as possible.
You need to keep an eye on the downstream effects of your test, even if it results in significant gains. It’s easily forgotten when it comes to A/B testing is that it’s rarely an isolated effect.
The changes can be something as small as sign up button copy or as big as a completely different layout.
When starting out, you'll want to get a feel for how the software you’re using and A/B testing overall works. This way, when you test something big, you’ll have all the kinks worked out in your process.
Another name for A/B testing that’s not used as often is “split-testing”.
Look at your homepage. Make a list of 5 things you think should be tested.
While running tests on the color of your signup button is tempting, it’s more beneficial to run bigger tests when you start out. An example is your headline. Making changes here is much more likely to result in big gains than changing button colors.
Be a scientist
Run each campaign as an experiment - come up with a hypothesis and test that hypothesis. If your experiment fails to lift your conversion rate, move on to a different experiment.
If your hypothesis proves correct, are there any other places you can use this new knowledge?
The process for conducting an experiment is:
- Ask a question - “Why is my bounce rate so high?” or “Why aren’t more people converting from the homepage to our signup page?”
- Construct a hypothesis - “I think changing the headline will lower my bounce rate” or “I think that changing the sign up button text to something more actionable will increase our conversion rate”
- Test your hypothesis
- Analyze your data and draw a conclusion
Which study would you trust more - one that interviewed 10 or one that interviewed 100? You’re going to trust the second more because of the larger number of people interviewed.
The results of the second study were more statistically significant because more people were asked. Statistical significance means that the result of your test is likely not random.
Each test requires a minimum number of people and actions. If your numbers are too low, you won’t have statistical significance and you won't be able to trust your results.
To determine statistical significance, use a significance calulator.
Tests that are most likely to result in big gains are also most likely to take more setup time than smaller tests.