Five steps to champion a data product strategy

Here are the five key considerations that data leaders must think about when looking to get the best out of their data product strategy

The difficult truth facing data and IT leaders is that a significant percentage of data projects never make it to completion or deliver against their desired goals. However, there is a new approach seeking to improve these outcomes: the data product strategy.

Treating data like a product gives more structure to the ownership, processes, and technology needed to provide the organisation with access to clean, curated, continuously-updated data. So, the data product becomes a consumption-ready set of high-quality, trustworthy, accessible data that can be applied to solve genuine business challenges. In short, it’s the best version of data available to service a defined purpose and achieve a desired outcome for the business.

For CDOs that want to implement a data product strategy, there are five key steps that will help them help them make the process a success.

1. Start with your why

Before you start developing the data product strategy, define the objectives that you want it to achieve within the business. I’d recommend starting small – identify a specific aim that reflects the business’s priorities, such as controlling costs or managing risks. Then align with the relevant stakeholders across the business – leadership, line of business (LOB), and IT – that will need to work together to fulfil this objective.

If you’re unsure where to start, I would recommend starting with your customer data. Customer data products can increase visibility into your customer journey, in turn creating opportunities to improve the customer experience, targeting, and conversion rates.

2. Answer the tough questions

Once you’ve defined your objective, you need to determine if you have the capabilities to implement a data product strategy. That means gaining an understanding of:

  • Where the data lives
  • How it is integrated across systems and departments
  • If the data is accurate and complete
  • How often it is updated

This will not only help determine the quality of your data, but what budget resources you will need to build a high-quality data product. When it comes to the technology, you will need tools capable of mastering your data, enriching it with external data sets, and integrating it across your systems and departments, as well as having analytical capabilities.

Building the team and assigning their responsibilities is also key – particularly for identifying potential resource gaps early so you can plan how to fill them. The role of the data product manager is becoming increasingly prominent to manage the design, build and cross-functional development of a data platform.

3. Determine the use case

Now it’s time to get stuck in. The LOB leaders will have challenges which are preventing them from reaching their objectives – but may not realise how data could help. Defining a use case based around such a problem is the best way to censure the data product tangibly influences a business outcome.

To determine the requirements, you’ll need to explore:

  • Why is a data product needed?
  • How will the data product be used?
  • Who will use it?
  • How will it be consumed?
  • What data should to include?

The LOB leaders might not have all the answers, but that’s your superpower. Help them connect the dots and show how a data product could help them

4. Get buy-in from the business

To move forward with your use case, you’ll need leadership buy-in. This is key to ensuring your project has the required funding, resources, and support. Share your use case and the roadmap to deliver it and make it clear how you will show value in the short term, before evolving the data product to increase its value. It’s also important to outline how you will measure success. Be prepared with KPIs that align everyone on the goals and – critically – what success looks like.

5. Develop a minimum viable data product

The first step once your strategy is signed off is to develop a minimum viable data product (MVDP). Start small so you can release quickly, before iterating and delivering further capabilities. Each release of your data product should offer a little more value. This will help drive adoption, as well as showing returns which will help you secure any additional funding or resources required.

Success will of course also depend on your LOB partners understanding how to use the data product as part of their existing working processes. It is rare that adding a new process will be widely and successfully adopted. Instead of changing habits, show how the data products enhance their existing processes. And then take on their feedback to ensure future releases are even more valuable to the end users.

Treat data as the asset it is

No one questions the value of data. But few data leaders will confidently say their organisation is successfully using it as the asset it is. That’s why data products present such a massive opportunity. Solving business problems with data transforms its role within an organisation from passively to actively delivering value.

This all starts with the right data product strategy. Starting small, proving its value, then iterating and growing to support more areas of the business. These five steps will put data leaders well on its way to implementing a successful approach.

Suki Dhuphar is head of EMEA at Tamr.


Five steps to build a business case for data and analytics governanceHere are five steps for data leaders to take towards building a strong business use case for data and analytics governance.

How to make progress on managing unstructured dataLong-time Forrester analyst and Elastic CIO discuss strategies for realising value from data with search-powered technology.