ENSURE YOUR SAMPLE SIZE IS LARGE ENOUGH
Larger sample sizes make for more accurate data. Larger sample sizes also represent your target more efficiently.
MAKE SURE YOUR TESTING IS LONG ENOUGH
The longer you test, the more data you can collect. If testing is cut short, your data may be flawed and may not represent your true test market.
USE APPROPRIATE ATTRIBUTION MODELING
Attribution modeling makes it easier for you to see who gets the credit for the sale made. Here are some common attribution models...
First Interaction
All credit goes to first action customer took to convert.
Last Interaction
All credit goes to last action customer took before converting.
Linear
Equal credit given to all interactions.
Time Decay
More credit is given to interactions closer to when conversion occurred.
Position Based
First and last interactions are given most credit, while the rest get equal amounts of credit.
MAKE SURE CONCLUSIONS ARE UNBIASED
Sometimes it is hard to be unbiased in your conclusions, but you have to remember that correlation does not equal causation. Just because two things may seem related doesn't mean it is, there may be an outside factor that is causing that relationship.
easy to read and follow