Mapping the Consumer Journey and Testing Hypotheses

Digital Marketing - Study Notes:

The role of web analytics

Web analytics tools can track the user or consumer journey from user origin and profile, through to on-site activity, to goals achieved and the exit point. This allows analysts to measure and understand user interactions and any valuable actions on the website using data to inspire hypotheses and insights into why these behaviors are occurring. It also allows analysts to try to recreate the conditions to drive more valuable actions or reduce the instances of poorer performing interactions.

Gaining a return from the metrics in programs such as Google Analytics is an ongoing process of monitoring, analysis, and testing enhancements from data-driven hypotheses.

Testing

Testing allows analysts to optimize campaign performance by altering single or multiple variables at a time in a controlled way. When data insights point towards an opportunity to improve performance, analysts can create a data-driven hypothesis to test it.

Testing involves a control and a test. The control is the original set of variables; the ‘test’ or ‘challenger’ is the new set of variables. When testing two or more hypotheses, in a 50/50 test, one option will be worse than the other. Testing reduces the risk of choosing the poorer performing option for the duration of the test and once complete. The results allow for a more informed decision on whether or not to proceed with the hypothesis.

Steps

Here are some suggested steps for testing hypotheses:

1. Choose one variable to test based on your insights or data.
2. Identify the goal for the test.
3. Create your control (existing set of variables) and test (new set of variables).
4. Set your test split (for example, 50/50).
5. Set your sample size, length of test, and channel or campaign to test.
6. Begin testing.

False results

A by-product of testing is the possibility of false positives or false negatives.

A false result is an error in the data where the result wrongly points to an outcome or condition. In testing, this is known as a type one error (false positive – where the data shows a positive result for the test when the true result was a negative), or a type two error (false negative –where the data shows a negative result when a positive result was the true outcome of the test).

So, how can analysts test for false positives? An analyst can:

• Use common sense when interpreting results.
• Trust but verify.
• Review historical performance and look out for anomalies in the trends.