08 Sep 2020
Data Insights: defining the questions that matter
Glenn Exton, head of data and analytics, considers the importance of asking the right questions and how to identify them in order to yield insights.
Unless we define the questions well, how can we provide answers that matter?
Too often, we leap to answers without first pausing to examine our questions. We get excited about solutions without considering whether we are addressing real or relevant challenges or priorities.
Data has the potential to improve the way that we solve problems. However, the insights generated from data are only as good as the questions we ask. We need to define the questions that matter the most – those that, when answered via the right data, will make a transformative difference.
Change at an unprecedented rate
The variety and complexity of the challenges facing the financial services industry and the customers we serve are changing at an unprecedented rate – whether it be a pandemic, climate change, economic uncertainties, cybercrime or changes in ways of working. At the same time, stakeholders, investors, regulators and customers expect us to take evidence-based, ethical and measured decisions that, if not made right, could significantly diminish the trust and confidence that we have built over many years. Not only do we need innovative solutions; we also need innovation in how we define the problems we encounter and subsequently develop our response.
So, how can we get started? We do this by defining the areas that matter the most to our customers and stakeholders. In a macro view, we might cluster questions around climate or strategic partnerships; at a department-wide level, the focus could be on operational excellence or cost to serve; while a team leader may want to look at service delivery, or the health and well-being of the team.
The key step is figuring out which areas you want to focus on and the most important questions that, when answered, will make the biggest difference.
Forming the team and establishing mindset
Upon defining the areas of focus, you need to identify a group of people who have knowledge about the area and the typical problems being faced, and those who have a fundamental understanding about the data within. This cohort of specialists should have a ‘champion’ who leads the team in identifying the questions that need to be answered.
The key step is figuring out which areas you want to focus on and the most important questions that, when answered, will make the biggest difference
In the formation of the questions, the cohort need to take a holistic view and explore questions that relate to situations that have already happened (descriptive), something that may happen in the future (predictive) or something that the cohort are attempting to influence (prescriptive). Keeping an open mind is critical; this is not the time to be protective or reserved. Things that matter the most are most likely the areas we must tackle first and head-on. An example of this is below:
- Descriptive: What market segmentation and channel mix yielded the highest profit margin in the last six months? What caused the dip in profit margin?
- Predictive: What market segmentation and channel mix will most likely yield the highest profit margin within the next six months?
- Prescriptive: What intervention options can we take that will yield the highest profit margin over the next six months for our target customers?
Formulating the questions
This is where the fun really starts!
To kick things off, the cohort should write up the who, whom, what, why, where, when, will, ‘can we’, ‘do we’ to help unlock creative and critical thinking. During the early stages, keep the questions open to encourage exploration. Closed questions are a great way to secure confirmation of the hypothesis that is seeking to be tested.
At a micro level, let’s assume that the cohort selects customer experience as one of their focus areas. As a result of a 20-minute session, the cohort arrives at the following questions that mattered the most to them:
- Where are customers experiencing the most frustration with our services?
- Who will be most affected if we introduce this new service?
- Why are customers leaving at certain points along the customer journey?
- At what points along the customer journey are customers leaving the most?
- When will customers most likely recommend our services to others?
- In what scenario will customers engage the most with our service offering?
- Will customers respond to suggested next-stage recommendations?
Remember, it’s also important to keep a look out for questions that don’t matter. These are questions people ask when there is little chance that they will learn something from the answer. They include leading questions and statements disguised as questions. And while they can sound harmless, they can also prevent us from obtaining a better understanding of the situation around us.
Keep questions basic
We tend to learn more when we ask basic questions. These are the questions that matter. They’re short, simple, and meant to extract information.
The more detail that’s included in a question, the more we increase the complexity of the response that’s needed, and we may not uncover what we were actually looking for.
Ways to generate questions
Applying Google Ventures’ ‘Crazy Eights’ technique from its Design Sprint methodology is a fun way to generate questions that matter the most.
The cohort is given eight minutes to produce eight questions that matter. Upon reviewing and discussing the questions, the cohort is encouraged to repeat the exercise, producing another eight questions. At this point, the cohort can find it tougher to derive new questions. This is where some of the most important questions could be generated. Keep the questions basic!
For the cohort that are seeking to further diverge their thinking, invite people outside of the core area to participate in a Crazy Eight session.
Discovering new questions through data exploration
There are times when you may not even realise that you could have asked a question in the first place.
This is where data and analytical labs come into the frame. Labs are being established in organisations where they have access to a myriad of data to undertake experiments, look for patterns or anomalies and discover new insights. We are all on a journey exploring what is possible within the data discipline.
Creating your top 10 list of questions
The last stage is for the cohort to create a top list of questions that, when answered with the right data sets, will have the most transformational impact on the area of focus.
To help define these questions, I suggest the following approach:
Recap the entire process: why the domain was selected; the rationale behind the defining questions that matter the most first; what descriptive, predictive and prescriptive analytics mean; and discuss each question being posed
Give each cohort member five ‘dots’ and place them over the questions that matter the most to them. This will provide a first bank of top questions that matter the most.
To help shake things up a bit, include people who are not from the core area in the process and have them explain their point of view.
To bring a more inclusive approach to the process and subject to the area being investigated, it may be very helpful to apply ‘the wisdom of the crowd’. The coach is then encouraged to open up the voting to a broader audience, giving them one vote to a question that matters the most to them.
Finally, be wary of spending too much time defining questions that don’t matter – you need to be laser focused on investing time in questions that do matter to your organisation, customers and key stakeholders.
What’s in store this week for currencies and FX markets?
What’s in store this week for currencies and FX markets?
Our economists share their views on the key economic trends to watch in the month ahead.