Remember science fairs? The excitement that comes from standing in front of a tri-fold board and explaining the meaning of words you don’t even know to strangers who walk by is…exhilarating, right?
Having a ‘method’ to the madness that is the startup/my life has become an important task. That doesn’t mean I shy away from crazy and irrational decisions…but it DOES mean I very often (but not always) analyze the data prior to making a choice (side note – be open to always collecting new data points!).
The more you practice a procedure, the more it becomes subconscious to your daily challenges and the way you look at new problems.
1. Ask a Question
How do I get more customers to my site?
How do I reduce the amount of abandoned shopping carts?
What will increase the average price of an order?
Why aren’t my sales increasing month to month?
Would anyone buy my product idea?
2. Form a Hypothesis
Example: I can increase sales from abandoned shopping carts by sending a follow up email a day after a shopping cart has been left abandoned, asking if the customer is still interested in purchasing the products.
3. Setup the Experiment
There are many ways to experiment and run tests. The most important thing is to have some organization to it so you can measure your results and data.
Some things to consider:
Create both independent and dependent variables.
Create at least 2 cases (i.e. end a follow up email to half the customers and have a pop-up screen asking to finish checkout for the other half of customers).
Have a control group! This will often serve as your baseline (it might be what you are currently doing with no change).
4. Collect Data
This is important. It is not useful to make changes and then just ‘hope’ it helped. Collect many pieces of data – even if it’s collecting in manually. Make sure to collect it for an appropriate amount of time. One day of testing might not be enough depending on the size of market.
5. Analyze the Results
Take the raw data and put some meaning behind it! This is the best way to make data informed decisions. It’s okay for an experiment to have results that contradict your hypothesis. This happens all the time. This just means you need to run more experiments that test additional variables.
5.1 Once you have results, repeat the process!
Whether it is for the same problem or to solve a new one, always be testing for ways to improve your business.