Insights

Digital transformation has disrupted every sector, creating an unprecedented volume and variety of data. The challenge for business leaders is to derive value from this data to use it as a strategic asset.

The question is: where do you start?

I recently presented at Snowflake’s Data Cloud Summit and talked about how we at Infinity Works help organisations adopt data-driven decision making.

Solving business challenges with data

For the last couple of years, almost every conversation I’ve had with business leaders about their data follows one of three paths: a desire to use artificial intelligence (AI) or machine learning (ML); a want to modernise a data warehouse; or a need for a data lake. When I ask them why, it’s always for the right reasons. The answers tend to be about gaining competitive advantage, reducing churn, increasing conversions, getting a single source of the truth, or saving money. 

The problem, though, is they haven’t put any thought into how to get there. Poor quality of data, pipelines that take weeks to change, differing definitions, data all over the place (or sometimes not even knowing what data exists) and lack of skills are the common complaints that come up when looking to start a data project. 

During the last few years, we’ve helped a range of customers – including FTSE100 retailers, online gaming companies, fintechs, and e-commerce companies – solve their data problems and modernise their data estates. But in a post Covid-19 world, becoming a data-driven organisation will be a necessity rather than a desire.

This article is the first in a series of pieces we’ll share that detail the journey we take customers on when embarking on a data project. The articles will detail our unique approach to data projects, discuss real-world examples that we’ve worked on, and talk about the key learnings we’ve taken away from working in multiple sectors.

Data Strategy

We’ll start our series of articles by focussing on data strategy – something that every organisation needs before it starts to invest in its capabilities. Each data strategy will differ; some will fit on one page, some will include a detailed series of documentation, while others are simply a living whiteboard of the overall vision. However, every data strategy should contain the same common elements:

What problem are you trying to solve?

Rather than looking at what data exists in isolation, it’s important to work collaboratively across all areas of an organisation to understand the unanswered questions while keeping the strategic priorities in mind. 

What do you need to solve that problem?

It’s important to start small. It’s not about migrating all the data from a legacy system to the cloud, it’s about identifying what information is needed to be informed. Do you already have this data internally, or do you need to look externally? 

How are you going to capture, store and analyse to solve the problem?

Questions need to be asked to determine whether your data is accessible, has a common language and meaning, and if the quality is good enough. Do you have consistent ways of loading and transforming? Do you have an effective way of storing the data which is secure, scalable and cost effective? And do you have a clear path to communicate and read the data?

How are you going to govern the data?

Data can be very powerful, providing your organisation with insights that you’ve never seen before. But it’s important to remember that data can be a serious liability if it’s not managed correctly. Who is going to be responsible for it? Have you got permission to use the data? How is it going to be secured? What regulatory concerns need to be considered? Do you need to think about ethics and the honest use of the data?

How can you deliver value from the data?

Capturing data for the sake of it might seem like a good thing to do, but it’s a pointless exercise if it can’t be used to deliver value. It’s critical to think about how the data will be used in real-life scenarios. Will it be used for reporting? Will it be used to inform supply chain or sales decisions? Will it be used to inform pricing and promotions in real time? Or will it be provided to analysis and data scientists to model algorithms? 

Do you have the skills, processes and culture to support? 

This is a question which is often overlooked when embarking on a data transformation project. Data transformation has to start at the top by understanding how and why you’re using data, creating a culture of intrigue and removing unconscious bias. With the above in mind, you need to question if you have the skillsets internally. If not, can you train, hire or work with a partner to deliver what is needed? What does the local market look like for talent? Can you offer the packages to attract the right talent? Are you set up to quickly and efficiently ingest new data sources and produce new insights?

Data VMOST

As you can see, there are a lot of questions to ask, so where do you start? The best approach we’ve found to formulating a data strategy is to start with a data VMOST – an exercise which documents your vision, missions, objectives, strategies, and tactics. We work collaboratively with organisations, across two days, to define: 

• A clear Vision which describes what the future will, or shall, look like for your organisation – where you want to go and what you want to be.
• A series of Missions which are actions of changes that are concise, specific and feasible.
• Once the Missions are established, we then focus on Objectives, or goals, which can measure the success of the Missions and keep you honest to your overall Vision.
• Strategies are then defined which will help achieve the objectives set out. It also helps generate new ideas on how the Objectives, Missions and Vision could be achieved.
• Lastly, we come up with Tactics – the actions that need to be taken to fulfil the strategies.

Once all of the above steps have been undertaken and a clear documented strategy exists, we can then move onto the fun stuff – or at least what I consider to be the exciting part!

In my next article, I’ll discuss setting up teams and processes for implementation and change management, enabling an organisation to start to become data driven. The following articles will then cover operating models, data modelling and architecture, engineering, security/scalability and cost control, upskilling and culture, and finally, using data to derive tangible business results.