First principles of data products
The basics of data products, and it's not about tools and applications.
TL;DR:
A data product is insights that empower people to make decisions and take action to improve their life or business.
Contextually relevant Insights must answer:
What? & Why?
So What?
Now What?
7 types of data products: Data Feeds, APIs, Alerts, Dashboards, Reports, Predictions, Data Tools.
The next post on the newsletter is the answer to the question: How to data product?
Context
We live in a data hungry world. The demand for skills in data has grown exponentially over the past 20 years. We are going through several hype cycles in the data world in parallel, and it is expected to continue as far as we can see in the future.
Whenever I have experienced such shifts in the market, I have observed a few trends:
Demand-supply gap: high demand and low supply of talent
Exaggerated expectations: Most underestimate effort and overestimate impact
Encounter with reality: Significant churn due to improper expectations
Our focus must be to reduce the churn by clearly grasping the basic concepts. My goal is to state the concepts with simplicity.
Many people at this point in time in 2023 are new to the concept of a data product. There are many who even say that data products are a novelty. A new concept, a new paradigm.
Not to burst anybody’s bubble here, but data products has existed in the digital economy for a long time. What is described as a novelty is the way in which we collect, process, deliver, and consume data.
Understanding data products is foundational to a number of data roles like data analysts, data engineers, data scientists, machine learning engineers, analytics engineers and whatever new title the market comes up with.
One of the recent additions in the list is the role of Data Product Manager. In order to understand data product management we need to understand data products.
Folks have already started searching for the basics and asking questions in conversations. While there are ebbs and flows in historical interest on data products, in the past couple of years the interest is trending higher.
what is a data product
A white paper from Tableau says the following:
A data product is an application or tool that uses data to help businesses improve their decisions and processes.
That’s a decent definition with some room for iteration. Let’s double-click and work backwards.
For-profit businesses exist to build profitable products, no surprises there.
But building those products requires tools, processes, people, raw materials or code, time, and effort—all of which come at a cost.
To turn a profit, businesses must market and distribute their offerings, driving sales and revenue. And that means constantly making decisions about every aspect of product development and distribution.
Remember: Not taking a decision is still a decision!
In the digital economy, data informs and improves these decisions, adding value along the way.
Based on the expected outcomes and jobs of the customer, we can define data product as:
A data product consists of data and insights that empower people to make decisions and take action to improve their life or business.
Data becomes insights when processed with the right software tools and applications. In a digital context, these tools are part of the data product ecosystem.
But for paying customers, the tool doesn't matter - it's dispensable. What they want is valuable data and meaningful insights that help them solve problems and reach their goals.
Contextually relevant Insights must answer:
What? & Why?
So What?
Now What?
Data products must deliver real economic value.
Tangible value can be measured by savings in time, effort, and costs. Successful data products increase productivity, throughput, revenues, and profit margins.
Bottomline: A data product that does not deliver insights that leads to economic value is useless and will not be profitable.
Types of data product
Data products come in many shapes and sizes. Here are the main forms:
Data Feeds: Datasets which include metadata, reference data, enrichment data to improve the data inventory.
APIs: Programmatic access to data and interactions.
Alerts: Automated notifications on unusual values based on defined thresholds.
Dashboards: A view into the latest vital signs of a business or project.
Reports: Summary of past events and trends from relevant datasets for a time period.
Predictions: Algorithmic categorization or forecasting based on historical data.
Data Platforms: Software to capture, process, profile, visualize, model, and distribute data products.
How to data product
This is the setup of my newsletter!
Building data products is tough. The innovations keep coming and the demand-supply gap keeps expanding, making it even harder to succeed.
This is my fifth time reliving the same pattern. I've struggled through steep learning curves in Big Data Engineering, Data Science, Machine Learning, and Deep Learning since 2009.
This time, I'm focused on Data Products and Data Product Management. But here's the good news: I've been playing at the intersection of data and product management for 14 years now!
Data since 2009, Product since 2014.
So even though the climb is steep, I've been training for this moment.
The next post on this newsletter is to share my method to the madness.
If you are enthusiastic about data products and data product management, here is how you can subscribe:
Hope you enjoyed this piece. Let me know your thoughts or questions in the comments.
Keep at it!
Deb.
I like the simple definition and the list of data product types. Just wondering if "Data Platforms" (tools) should fall under this category as those tools often enable one to build data products like reports and dashboards.