Conversion Rate Optimization: A 12 step process to success and ideas of what to test

success means increasing your conversions

Below is VKI Studios' 12 step process to conversion rate optimization testing, and some ideas on what can be tested.

Information gathering- To begin the process you need to start gathering and understanding a portfolio of background information to help set the context for the test. Some of this information will likely include:
- Understand your site and page's goal(s)
- Know your visitors/customers' Demo and psychographics
- Understand your site/product's unique selling proposition
- Review Analytics Platform: Google Analytics
- Interviewed customer service reps
- Reviewed customer feedback surveys
Analytics review- Once you have collected and processed the background information you will next want to review the analytics information that has been collected so that you can:
- Agree upon Key Performance Indicators (KPIs)
  - Unique visitors, bounce rates, conversion rates, revenue
- Establish baseline data from analytics
- Look for problems and opportunities
  - Product pages that have high bounce rates
  - Low click through rates from landing pages to product pages
  - Low browse to cart ratio
Create personas- With all of this information in hand it is now time to start creating personas that will represent your different visitors and customers. These personas will help you better understand how to communicate persuasively with your visitors.
Formulate performance hypothesis- Now that you have created your personas and gone through both the analytics and information gathering steps it is time to start hypothesizing why your pages are performing at the levels they are, and what could be done to improve their performance.
1. Why is performance poor and how do we improve it?
2. Test hypothesis with quick/informal user test
3. Develop shortlist of pages to test
  1. Test pages need to get a significant amount of traffic
  2. Test pages also need to offer a significant number of success events.
    - If a sale or lead isn't generated often enough, make the success event a step further up the conversion funnel
Select testing methodology- Decisions, Decisions
1. Select a testing methodology and test that will generate momentum
  1. Start with an A/B test vs. a Multivariate test
  2. Create a test that will be able to have 100 successes/combination in a 1 month period
2. Define successful outcome
Develop test pages- It is now time to bring your performance hypothesis to life.
1. Focus on big things first and subtleties later
2. Potential items to test
  - Headlines, subheads, calls to action, images…
3. Create alternatives to be tested
  - Both text and visual elements
Set up test- get the code on the page and test your test.
1. Add code to test page and conversion page
2. Set up test in application administration area
3. Test implementation
Launch test- hit the launch button and sit back and relax. Don't get too excited and start jumping to conclusions when you haven't collected a meaning amount of data.
Analyze results- Once the test has collected enough data to provide statistically significant results it is time to dig in and see what worked and what didn't and to try and figure out why.
Run follow up test- if you need to be extremely confident that you have the winner then you will want to run a follow up test. This will put the original vs. the winning combination in an A/B test.
Implement the winner- Implement the winner and remove the test code from your pages.
Start planning next test- Now that you are done it is time to start all over again.