1. Forming the analytical questions
Hi, in this video we will learn tips and techniques to help us translate business questions into well-formed analytical questions.
2. Transitioning to the analytical world
This translation is a process where we are moving from the business world to the analytical world. That is, into a language more familiar and actionable to the data experts. Our goal is to form one or more analytical questions that follow the SMART methodology, meaning they are, specific, measurable, actionable, relevant and time-bound. This will allow the analytics team to select and deliver analytical solutions that solve the business problem.
3. Steps to forming the analytical questions
In order to form analytical questions that exhibit those characteristics, we can follow a number of steps:
Extract the key information from the business question.
Break down the business question into smaller, more specific analytical questions that can be answered using data.
Review and refine the analytical question to make it as specific and focused as possible.
Review relevance to make sure the analytical question formed can address the business question.
4. Consider the following scenario
Now, let’s go through these steps in more detail by considering a scenario from the online marketplace for short-term rentals “CozySpace.com”. The marketing team has raised a business question, and the analytics team has collected more information to understand the problem, with the resulting business question: “How effective was our campaign in December measured by the booking rate for the group of frequent travelers in the US?”
5. Extract the key information
Let’s start by extracting the key information and factors that the business question is asking for
From this question, we know that our goal is to measure the effectiveness of the campaign,
the specific metric we want to analyze is the booking rate.
We are focusing on the US market and the group of frequent travelers.
The timeframe of the analysis is in December.
6. Break down the business question
Next, in this critical step, we need to break down the question into analytical questions by considering the information we extracted. This is a brainstorming process where we should ask what we need to know from the data to answer the business question. In many cases, one analytical question might not be enough and we need to form multiple analytical questions.
In our scenario, the analytical questions can be the following:
"What was the booking rate for frequent travelers in the US one month before the campaign?" This question establishes the baseline for the current booking rate to help measure the effectiveness of the campaign.
The second question can be: “Is there a difference in booking rate during the campaign compared to our baseline?” This will help provide quantitative evidence of the campaign's impact on the booking rate.
7. Refine the analytical questions
The next step is to make the question as specific and focused as possible, or otherwise SMART. For example, instead of asking a broad question like "What factors affect sales?", you could ask a more focused question like "What is the relationship between marketing spend and sales, based on our past data?"
Once we have formed the analytical questions it is important to consider what data is required to answer them. Due to data limitations, we might not be able to proceed with certain analytical questions. In our scenario, we need to use data from our bookings platform with the granularity of market, date, and customer groups.
We can refine the second question as: “Is there a statistically significant increase in the booking rate during the campaign compared to our baseline based on the bookings data?”. This question is measurable and more specific.
8. Confirm relevance with business question
Great! Now as a last step, it is useful to make a final review and make sure the analytical question or questions formed can address the business question. This way we can be confident that we are on the right path to solving the business problem. Looking at our scenario, the analytical questions are focused on testing the hypothesis that the campaign caused an increase in the booking rate.
9. Let's practice!
Now, let’s check your understanding!