WHAT Data Products can Data Science build?

aka a Data Product taxonomy

Sep 01, 2024

Data Products in Data Science are tools or applications that use data to solve a specific business need or problem, not raw datasets or architecture components.
There are 4 classes of Data Products: (1) dashboards (for at-a-glance metrics), (2) answers (for deep data insights), (3) deployments (for automated data-driven algorithmic processes), and (4) custom-built apps (for unique business needs).
Building and maintaining each class of Data Product requires different combinations of skills (e.g. Data Engineers, Data Scientists, Machine Learning Engineers) and technology infrastructure (e.g. data, computing and/or solutions architectures).
The choice of Data Product(s) to implement should be driven by a business's specific needs and challenges, not by trends. Data Products should evolve as the business grows and changes.

We’ve discussed WHO builds Data Products and WHERE these Data Products are built and maintained in a Data Science Shop. Now, it’s time to explore WHAT kinds of Data Products are there.

Unless you are a Data Scientist, it can be a bit mysterious what Data Science can do for you (or for your business). They key thing to remember about Data Science is that its purpose is to solve a need or problem in your business. How does it do that? Through the Data Products it builds for you.

So, what is a Data Product?

First, let’s clear up some confusion: when we talk about Data Products in Data Science, we're not referring to pieces of a data architecture or raw datasets. Instead, a Data Product is a solution that uses data to solve a specific business need or problem. Think of a Data Product as a tool or application that takes your raw data and turns it into something useful and actionable for your business.

For example, a Data Product might be a dashboard that shows real-time sales performance across all your stores, an algorithm that predicts which customers are likely to churn, or a custom app that optimizes delivery routes for your fleet of trucks. These are all solutions that leverage your data to help you make better decisions or improve your operations.

But you might still be a bit fuzzy about the form of Data Products are available to you in a Data Science Shop. Let's dive into the four classes of Data Products that can transform how you use data in your business operations and decision-making processes.

Four classes of Data Products delivered by a Data Science Shop

Now that we've got a handle on what Data Products are, let's explore the different types of Data Products that a Data Science Shop can create for your business.

In my years of experience, I've found that most Data Products fall into four main classes. Each one has its own strengths and is suited for different situations. Understanding these categories will help you get a clearer picture of how Data Science can work for you, and what kinds of solutions might be most valuable for your business.

But, first, take a look at this short animation to gain intuition about the Data Products that can be built in a Data Science Shop:

1. DASHBOARDS: your business at a glance

Let's start with dashboards. Imagine having all your crucial business metrics at your disposal and updated in real-time. That's the power of dashboards.

The beauty of this class of Data Products is that they take complex data and present it in an easy-to-understand format, often using charts, graphs, and other visuals. It's like having a constantly updated snapshot of your business's health and performance right at your fingertips.

Dashboards are perfect for businesses that:

need to move away from manual, time-consuming reporting processes
want real-time updates on key performance indicators
seek to empower employees with readily accessible data

To build effective dashboards, your Data Science Shop will need:

Data Engineers to set up data pipelines and ensure data quality
Data Analysts to select meaningful visualizations and interpret trends
a robust data architecture to store and process information
a solutions architecture to host and deliver the dashboard to users

2. ANSWERS: digging deeper into your data

Now, let's talk about answers. When we say “answers” in the Data Science world, we're not just talking about simple yes-or-no responses. Think of an answer as a deep dive into your data to address a specific, often complex, business question.

This class of Data Products applies advanced scientific methods to answer specific business questions. Answers can tackle four main types of questions, each giving you a different perspective on your business:

explanations: say your customer satisfaction scores suddenly dropped last month. An explanation-type answer might reveal that the 8% drop in satisfaction happened among users that connected with your new automated customer support system, but no change in satisfaction happened when a human was on the line.
scenarios: you might ask, “What would happen to our profits if we raised prices by 10%?” This type of answer helps you play out “what-if” situations before making big decisions.
projections: these are all about peering into the future. An example might be forecasting your sales for the next quarter based on current trends and historical data.
causes: this is about understanding the drivers behind certain behaviors or outcomes. For example, you might want to know what factors most influence customer loyalty in your business. Is it price? Quality? Customer service? An experiment or a causal analysis can help you pinpoint these key factors.

Each of these types of answers - explanations, scenarios, projections, and causes - gives you powerful insights that can shape your strategy and help you make smarter, data-driven decisions.

Answers are ideal for businesses that:

require quantifiable, science-backed insights for decision-making
need to understand complex cause-and-effect relationships in their data
want to understand why something happened, explore "what-if" scenarios, or make data-driven projections

To generate meaningful answers, your Data Science Shop will require:

Data Scientists skilled in advanced statistics and machine learning
Data Engineers to prepare and provide access to relevant data
a comprehensive data architecture to store historical and current data

3. DEPLOYMENTS: putting data to work automatically

Now, let's talk about deployments. Think of a deployment as putting your data to work automatically, around the clock. It's like having a tireless, super-smart Data Scientists who's constantly crunching numbers and delivering insights.

The beauty of this class of Data Products is that they work silently in the background, feeding valuable insights and actions directly into your business processes. For example, imagine you're a retail business owner. Instead of manually forecasting inventory needs each week, a deployment could automatically predict what you'll need to stock, updating in real-time as sales happen. With deployments, you're essentially baking data intelligence right into the day-to-day operations of your business, helping you make smarter decisions faster and more consistently.

Deployments are crucial for businesses that:

need to reduce human intervention in data-intensive processes
require frequent updates to key business metrics or forecasts
want to embed advanced Data Science directly into their operations

To create effective deployments, your Data Science Shop should have:

Data Engineers to build robust data pipelines
Data Scientists to design and calibrate advanced algorithms
Machine Learning Engineers to scale and optimize these algorithms
both data and computing architectures to support these automated processes

4. CUSTOM-BUILT APPS: tailor-made solutions for unique needs

Finally, let's talk about custom-built apps. Imagine having a suit tailored just for you instead of buying one off the rack - that's what a custom-built app is like for your business. It's a unique application designed specifically to address your company's particular needs or challenges.

Maybe you've outgrown off-the-shelf software, or perhaps your industry has specific requirements that standard tools just can't meet. That's when a custom-built app comes in handy. For instance, a logistics company might need an app that combines route optimization, real-time traffic data, and customer preferences all in one place. Or a healthcare provider might require an app that securely manages patient data while integrating with various medical devices. These apps are built from the ground up with your business in mind, incorporating advanced Data Science, machine learning, AI, or whatever other capabilities you need.

They're often more flexible and scalable than pre-made solutions, growing and adapting as your business does. While they require more time and resources to develop, custom-built apps can give you a significant edge by solving problems in a way that's uniquely tailored to your business.

Custom-built apps are essential for businesses that:

have outgrown existing services due to data scale or the complexity of the tasks to fulfill
require industry-specific or highly tailored solutions that do not exist off-the-shelf

Developing custom-built apps requires the full spectrum of Data Science Shop talent:

Data Engineers for data pipeline development
Data Scientists for algorithm design and prototyping
Machine Learning Engineers for scaling solutions
Project Managers to oversee the development process
Data Analysts to interpret and communicate results
all three data, computing, and solutions architectures

Choosing What's Right for You

Here's the most important thing to remember: there's no one-size-fits-all in Data Science. The tools you need depend entirely on what problems you're trying to solve in your business.

You might start with a simple dashboard and later realize you need some automated processes. Or you might jump straight into custom solutions. It all depends on your unique situation.

The key is to focus on your business needs first, then figure out which Data Products can help. It’s never a good idea to keep up with the Joneses - worry instead about what will actually make your business run better.

Growing with Data Science

As your business evolves, your data needs will too. That's normal! A good Data Science Shop can adapt and grow with you, creating new tools as you need them.

Remember, Data Science isn't about the fanciest tech or the most complex math. It's about solving your business problems and helping you make better decisions.

Did we spark your interest? Then also read:

WHO are the people that make Data Science possible? to learn about the specialized technical roles that work in tandem on a Data Science Shop
WHERE are Data Products built (and maintained)? to learn about the technology backbone of Data Products
What is Data Science? to learn more about the practical way to understand Data Science: focus on its outputs (Data Products)
What is the Data Science Shop? to learn more about the roadmap for the operation of Data Science in a business