In the previous piece, we uncovered the different characteristics of a well-designed data product. In this second blog, we will illustrate a real case scenario that tells the story of a data analyst, “Fred”, trying to navigate the complexities of his data landscape using our data product platform.
Once upon a time when people used to go to the office for work, Fred bumped into Tatiana, a fellow data architect, in the elevator and asked for her help with some problem that has been keeping him awake at night and he goes:
“I started a new project with one of the biggest car dealerships in the city and they want to be more customer-centric. To help them do that, we need to build a 360° view of their customers so that they can deliver personalised services. As you know, such a holistic view of customers’ data means all hands on deck. From customers’ basic contact information, present and past purchases, interaction history with offline or online channels to any recent requests or claims. We need to pull all the data on the interactions they had with the business touch points across the entire customer journey.
However, I have been reaching out to the marketing teams, sales, anyone that might have the data but in vain. No one seems to know where it is and what it looks like. Each system is so siloed and isolated that not even the tech people have an end-to-end view on the data flowing around the organisation. Why can’t data be more like a self-service vending machine?”
Tatiana says: “It sure can be. I know that getting the data together can sometimes turn out to be even harder than the analysis itself. Not anymore. We started using a data product platform that basically gives you the ability to work across data silos in a self-service mode.
In your case for example, with a simple self-serviceable search functionality, you can easily find what you need on your own and in a matter of seconds. In the platform, there is a repository that captures all the data details from the siloed systems at the enterprise-scale. This is one of the core functionalities of a data catalogue. There, you can search for all the available data sets and their descriptions, that are organised in a way that makes them discoverable and understandable.
Once you select the data you need, you might want to have a closer look at the description attached to it. Why? Simply because that description brings the necessary context, data stewards, business architects like me and business analysts like you need to address and understand the data so that it can be used in confidence and consumed to create value. These questions can easily be answered: data origin, lineage, frequency of updates, synchronisation, what happens to it across the multiple data systems. All the information you need is at hand.
What you can also do is take a few samples from your data sets and run some tests to check the quality and trustworthiness of your data and identify key performance metrics to unlock the most value out of your data.
Now that you have chosen your data sets, made sure that they tell the right story and that they are exactly what you need, you are ready to use them in your analytics, data science or operational projects. But clearly you would also like to ensure that your data assets are consistent and do not get misused. That’s possible thanks to the embedded governance. Who can access data? For what purpose? At which confidentiality level? Everything is kept under lock and key to protect your data.
To sum it all up, you can now discover what data is available, understand what it looks like, where it is and whether it is valuable for you to feed it into your project or analysis in a simpler, faster, and safer manner.”
As the elevator approaches its destination, Fred looks at Tatiana and says: “Power to the people, now I can finally be in control of my data projects!”