How to create a Data Strategy

Introduction

These are my thoughts on how to create a Data Strategy. It’s based on my experiences and principles. I hope that you find it interesting and useful.

The starting point for any Data Strategy is your organisation’s Business Strategy. They work in partnership, but the Business Strategy is the “source of truth” and the Data Strategy will show how your use of data will support it. If you’ve not got a defined Business Strategy, then you may wish to work that out first.

A couple of important starting points.

• A Data Strategy is never created in isolation. Internal and external factors always influence it. Understand what they are.

• The format is flexible. If you need a formal document that is fine, but it can equally be a slide deck that describes the strategy.

There are three stages to creating the strategy

1.      Discovery Phase

2.     Development of the Strategy

3.     Building the content

Discovery Phase

The discovery process identifies what the Data Strategy should address and what information is available to support it.

The following inputs are essential.

1.      The Business Strategy

This includes business goals, success criteria, or ambitions. It lays out what the organisation aims to achieve and how.  This includes targets, resources and propositions.

Here is a useful link to Business Strategy information for those interested.

 What is Business Strategy? Definition, Components & Examples Explained (consulterce.com)

2.      Technology Strategy

This will detail the established technical environments that must be considered and catered for. For example, if the organisation is Microsoft 365 based or uses AWS cloud or even has both.

It offers a base of the technical options that the data strategy can be developed from.

3.      Data Source Information

A Data Dictionary provides this information. Otherwise conduct a thorough audit of your existing data platforms, the data that exists within each and understand their future.  Determine if any are being replaced as part of this process or if there are any other initiatives underway.

You must be ruthless to avoid letting unnecessary legacy systems linger.

4.      Data Governance Policy

You should have a policy that addresses the following

• Data Quality

• Data Ownership

• Data Security

• Data Awareness

It operates as an input but must also be updated as part of the process.

5.      Investment levels and Risk appetite

Budget is always a factor. You must clearly articulate the investment and the benefits case. Quantify the ability to being able to make better decisions as a result of this strategy.

Equally, you need to make an accurate assessment of the risks in following this strategy. There may be a need to push for a more aggressive strategy, whilst other parties may ask for a safer path. This needs to be agreed objectively.

You should ensure that you have an understanding of your budget options and the risk appetite of the organisation.

6.      External factors

It is very rare these days that any company is able to operate independently. There will be suppliers or customers that will share data with you and these must be considered within your strategy. In many cases, these will be constraints that it must work within.

These may also be opportunities that can be exploited. There may be external datasets or processes that can enhance your offering.

The rapid evolution of AI is something that a modern data strategy must consider, but always from a data perspective.

Strategy Development

Once the discovery phase is completed and you understand all the background factors, you can develop the strategy.

The principal document will be the business strategy, the other documents will operate as filters and levers, narrowing and honing your options.

Here are the parts that you now need to consider. 

1.      Purpose

Convert key points from the business strategy into specific data-related considerations.

This is the core of the strategy and will require time and skill to ensure that the data aspects are accurately represented.

For example

• If the business vision is to deliver a low-cost solution, the data strategy should focus on measuring cost.

• If the business vision is to deliver a brilliant customer experience, the data strategy should focus on knowing everything about a customer.

Try to hone this into specific measurable objectives. It will take a number of iterations.

2.      Timescales

It is vital to articulate timescales within the strategy. Boards will always be looking for quick wins and an immediate return on investment. No Board is interested in a back-loaded 5 year strategy.

Balance short-term (up to one year) wins and long-term (no longer than three years) goals.

Short term wins keeps the focus tight and works within an organisation’s annual budgeting cycle.

Long term goals should be no longer than 3 years. Beyond that horizon, there are too many unknowns and too much potential change. You will never get there.

3.      Budget

You will sell the vision, but the Board will always want to know what the investment looks like.

For the Data Strategy, you should not provide detailed figures (avoid this until you have actually done the subsequent design/specification work) but must be able to provide a high-level figure linked to a return on investment.

A list of tangible benefits with outline financial information is essential. I guarantee that if you don’t have this level of information, you will proceed no further.

4.      KPI’s

Define some Key Performance Indicators that will be used as the base measurements throughout the development and implementation of the strategy. These should be directly linked to the business strategy and your goals.

• If the strategy is about low cost, then measure the costs.

• If the strategy is about customer experience, then measure feedback scores or repeat business.

 Building the Content

A Data Strategy should contain

• Vision

• Objectives

• Data Components/Design

• Benefits and Costs

• Data Security

• Impacts on People and Processes

• Roadmap

1.      Vision

This is a simple description directly linked to the Business Strategy and should show directly how the data will support the strategy.

It doesn’t have to be complicated but should paint a clear and compelling picture as to why this Data Strategy is essential to the organisation.

Typically it can act as an Executive Summary

2.      Objectives

The strategy should clearly explain its purpose and objectives in a simple and concise manner. It needs to address:

• The current data landscape and identify any gaps.

• The intended use of the data and the expected outcomes.

• The benefits this will bring to the organization.

While avoiding technical details, the explanation should be accurate and comprehensive.

Create a simple diagram that shows the end-to-end picture. It needs to be simple to explain and show how data moves throughout the organisation.

It needs to cover the full data estate, even as it then proceeds to focus on specific items.

 3.      Benefits and Costs

Lay out the tangible benefits that will derive from implementing this strategy. Make sure that you link them back to the Business Strategy and show how they will be derived.

Provide an outline of the costs and where they will arise.

Demonstrate that this is a good investment and show how it will continue to be a good investment into the future.

4.      Design and Components

This is where the detail of the strategy will be found.

Describe what the individual components of the strategy will comprise.

The strategy always needs to be embedded in the present and it should always be built on the existing capability.

It should

• Refer to the existing data architecture.  

• Consider its strengths and weaknesses.

• Identify where it is not supporting the existing needs.

• Identify where improvement can provide significant business advantage.

A future target environment must be defined

• This will consider components that that need to be replaced because they are nearing end of life.

• There may be parts that can be decommissioned because the business no longer requires them.

• There will be parts that require investment because of new business initiatives.

Consider the tools that you are going to use to implement the strategy. These will principally come from the Technical Strategy, but you will deciding how they are going to operate and how they will support the data and its manipulations/presentation.

Remember that this is called the Data Strategy because it is all about the data. Do not get confused with technology. You should always approach this from the logical data perspective. Always think from the position of your fundamental data principles.

For example, the business strategy may require that you move to real-time data insight.

• How will that affect the data, the storage, the management and the governance.

• There will be a conversation about how that data will be transported, bandwidth requirements, hardware etc etc.

5.      Data Security and Privacy

As part of the privacy by design principles, consider both the data protection and data security aspects within your strategy.

This requires its own specific focus within the strategy.

• What level of data security do you require?

• What are the factors that will require this?

• These include your public presence, the sensitivity of the data and the impact and consequences if you were attacked.

• Define the impact that this change is going to have on your current data protection processes and policies.

It is essential to reaffirm this within the data strategy.

6.      People and Processes 

Understand the impact on people across the organisation. If your organisation is extremely data literate, then this may be a light touch. If not, then this will require an appropriate investment to get buy-in.

Describe the overall affect that this will have on the overall data culture within your organisation. There may be overlap with the governance processes and change management processes.

This must be communicated correctly. Everyone should understand the data strategy, its purpose and its benefits.

Consider additional training/awareness for anyone affected by this. Promoting an excellent data culture within the organisation is never a bad thing.

7.      Road Map 

Explicitly lay out the short term wins and the long term goals. As stated above, these should be one year and three year timeframes. They give excellent balance and give future room for manoeuvre depending on the needs of the organisation.

The goals should be matched to the Business Strategy and described in terms that will be recognisable to the broader audience.

For example, you may be creating a new reporting system and you can describe some of the specific reports that will be produced.

Or you could be committing to improve the level of data quality against specific items.

Finally

Things that the Data Strategy is not

•               A project plan

•               A technical data architecture

•               A technical specification

•               A progress report

•               A historical document

These things come later and they all use the Data Strategy as the guiding star.

You may wish to push some detailed information into appendixes. This is entirely a decision for you and your audience.