How do you use data to solve a problem?

For those of you who know me, you’ll know that I spend a great deal of time focusing on our teams mission to democratize data. For background, I’ve written and spoken in a number of venues about what a data democracy means and how to execute this vision. The core concepts are that data should be easy to access and easy to use by the typical subject matter experts. Subject matter experts are folks from a particular field (industry, geo, product, account etc), but not necessarily data experts.

My team and I tackle any and all barriers to data access and usability. We’ve had a lot of successes with this. But as those major hurdle are knocked over, I’ve started to notice another less obvious hurdle; the ability to break down a business problem into a data problem.

I have spoken to countless folks who understand the data available and know how to use our data applications, but get stuck when trying to form a solution plan. In trying to help them, I began thinking about how I typically handle these problems. This is when I had an epiphany. The foundation for everything that I need to know, I learned in elementary school!

How to break down a business problem with data

I know this sounds somewhat silly. But, when thinking through the steps that I take to solve a business problem, I realized that I do employ a strategy. The backbone of that strategy is based on the principals of solving a word problem. Yes, that’s right. Does anyone else remember staring at those first complex word problems as a kid and not quite knowing where to start? I do! However, when my teacher provided us with strategies to break down the problem into less intimidating, actionable steps, everything became rather doable. The steps: circle the question, highlight the important information and cross out unnecessary information. Do these steps and all of a sudden the problem is simplified and much less scary. What a relief! By employing the same basic strategy, we too can feel that sense of calm when working on a business problem.

Word Problems to Data Problems

Before tackling the business problems, let’s review the word problem strategy. Outlined below is your checklist to breaking down a word problem.

Checklist for breaking down a word problem

Read the problem
Highlight the questions you need to answer
Underline essential information
Cross everything out you don’t need to know
Pick a strategy
Solve the problem
Show your work

An Example: Sally’s Snacks

Shown below, I’ve picked a very straight forward and typical word problem to illustrate the process.

Solving the problem

Start by reading the problem. Move on to identifying the parts of the problem. Highlight the question to answer: “How many apples does she need?”. Underline essential information: “4 friends, 3 apples for each of them”. And then cross out everything we don’t need to know. From there it is pretty straight forward. To solve this problem, we need to use multiplication. If we have 4 friends and they each need 3 apples (4x3) then we need 12 apples total. Simple! This is the type of problem that we are now able to solve without a checklist, but when word problems were new to us, these strategies helped tremendously.

Business Problems

To solve a business problem, we want to take what we learned above and add a few additional steps.

Timeline and Scoping - We need to treat timeline and scoping information as essential information. With grade school word problems the timeline was typically immediate and the scope was to solve in full. In business, it’s much more fluid. You may have a shorter deadline which can force you to get a minimum viable product (MVP) out the door as soon as possible. Or this might be a longer term effort, which will allow you to build out a more involved solution.

Data Object Definition - Data object definition is something to be taken incredibly seriously in business problem decomposition. In word problems you are typically dealing with straight forward, whole objects; like apples in our example above. However, in business problems, objects definitions are typically more complex. They require an explicit definition that usually includes a number of objects, properties and relationship considerations. They often involve some data manipulation to construct the object in question.

Two Way Communication - In word problems, the communication is typically one way. The student reads the question being given and they communicate the answer once they have a solution. With business problems, we want to communicate early and often. The chances for misunderstanding a portion of the ask are high, so it’s important that we communicate with the stakeholders throughout the solution process to ensure a successful outcome.

We’ll now create the business problem decomposition checklist by including the extra building blocks detailed above.

Checklist for breaking down a business problem

* New steps are highlighted in bold

Read the problem
Highlight the questions you need to answer
Underline essential information, including timing information to help scope the problem
Box all objects and define them explicitly
Cross everything out you don’t need to know
Pick a strategy
Communicate
Solve the problem
Show your work

An Example: Test Drivers Campaign Success Measurement

To illustrate how to break down a business problem with the checklist above, we’ll walk through an example. Note, that this is NOT a real scenario or a real email. However exaggerated, it is reflective of the type of intimidating questions that we can receive in the business space. The loose object definition, expectations and emotions can make a problem seem intimidating and stressful. The hope is that when approaching these requests systematically, they can be a lot easier to face.

Solving the Problem

Below, you can see that we started with the same steps as our word problem. Read the ask and identify the parts of the problem. The addition here is that we have also included timeline and data objects as parts of the problem.

For timeline and scoping, we see two word identifiers: escalated and expedited. In these situations, the idea is to keep a cool head but understand that it’s important to try to get a minimum viable product out the door as soon as you can.

For data object definition, we also have two objects identified. I’ll dive into one example of object definition. The first object is “test drivers campaign”. While it has been referenced as if it were a simple object, do not be deceived, this not a standalone data object and it requires some construction. In this case the “test drivers campaign” is a logic definition that includes a combination of user actions, user attributes and time constraints. As an example, the definition may resemble something like this: All page view events where the URL = littlemissdata.com/driving and the campaignCode= testDriver within the past two months for users in Europe. We will have to gain agreement on the object definition from our stakeholders and perform some data manipulation to construct the data set.

As we started to indicate above, we are going to communicate, communicate and communicate some more. It is important that you are crystal clear with your understanding of the problem space and how you plan to solve it.

Once we have a solution, we are going to deliver it in an appropriate manner. In this case, it is likely going to involve some sort of report. Ideally, it will also come with a presentation to get everyone on the same page about the findings.

Thank You

There you have it! By leveraging some the foundation of what we learned in grade school, we are able to more systematically solve some intimidating business problems using data. Thank you for following along. Please comment or find me on twitter to let me know what you think. I’d love to hear about any areas you all think I missed, or any scenarios what you have been able to use these strategies!