Data questions answered: How can we build more confidence in our data strategy?

Building a data strategy can be a scary exercise. It lays out a multi-year vision for how to improve data usage in an organisation. It typically also asks for large pieces of tech and process change, as well as a budget appropriate to the scale of that change. It’s a technical leader going out on a limb to lay out their vision of how they can help the business. 

Building confidence in your data strategy and having your stakeholders and team have confidence in this is a critical step on the path to delivery and getting full buy-in.

This is where data strategies can sometimes go wrong. I’ve seen them end up as insular, inward-facing and often quite verbose presentations. It’s never the intention, but it’s easy to get excited about a strategy that addresses all the technology, process and infrastructure changes. What can get lost is the impact on business value and strategic objectives. 

Depending on the stage of your data strategy, there are several ways you can build confidence in your approach and ensure it is the right strategy for your organisation.

The stages of a data strategy plan

1. The planning stage - focus on the return on investment

During the planning and strategising stage, our main focus should be stakeholder engagement, meeting stakeholder needs and delighting them. A data strategy that has the following is inherently starting from a good perspective:

• A clear line of sight to the business problems that it needs to solve, and can clearly articulate how it will solve those problems
• An understanding of which problems are solved best by which tools, rather than trying to force everything to fit e.g. an AI solution
• A view on technology, process and capability changes required to deliver these solutions.

Understanding and articulating how data is aligned to delivering key strategic business outcomes is the critical first step. Whether that’s deepening relationships with existing customers and serving them better, reducing costs in the supply chain, or optimising net interest income, ensuring your priorities in the data ecosystem are aligned to the right business outcomes will be key.

What does success look like?

During this stage, project definition is critical - using the outcomes we just defined to map into specific pieces of work, which are then supported by the infrastructure, capability and governance requirements needed to deliver the project. Also critical is being able to clearly articulate the definition of success. 

The definition of success should take the structure of a set of KPIs that result in a project being successful. For example, we might need to improve our fraud detection recall from 80% to 90% while maintaining current performance, or we need to cut demand forecasting error rates from 30% to 24%. These should be able to map into a rough estimate on the business value deliverable - by reducing error rates or fraud, we should be able to highlight a clear return for the business on the outlay. 

A clear set of KPIs should also make it very simple and objective for a business stakeholder to decide if a project will meet their needs.

2. The building stage - focus on the short term, build for the long term

One of the key parts of any data strategy is the definition of it. Building a data strategy that stakeholders will have confidence in relies heavily on having incremental blocks of delivery. 

Instead of embarking on a multi-year, major top-to-bottom transformation approach, it’s important to look at what value can be delivered in the short term, and focus on what opportunities can be delivered tactically. 

While modern data platforms and ML/AI environments are important for businesses to be capable of rapidly prototyping and delivering new data solutions, it is still possible to prototype and potentially productionise solutions. There will be a higher cost and technical debt for the sake of building an ML or AI solution without the core platforms.  However, it means you can deliver into production 3-12 months faster, rather than waiting on the platforms, and show return on investment faster.

Articulate the benefit

It’s important to think about “What’s in it for me” for every consumer of the strategy. This could be the data team itself creating a clear plan for how they will work in the end state. It could also be business stakeholders who will have various business challenges solved.  

Making sure the data strategy caters to a range of perceptions is critical for its success. Ensure a senior stakeholder is visibly championing the data strategy, with a particular focus on the tangible benefits it will deliver.

Think about the operating model

It is natural to have a desire to build a data empire, where all things data directly report into one person in the central data function. The thinking is that this makes it easier to exercise complete control over technical standards, processes, and governance.

In reality, business teams that have had a data capability internally in the past may find this very discouraging and disengaging. My advice is to think about the right operating model that allows teams to move at speed, while still having the necessary guardrails to operate safely and in a governed fashion. 

Some elements of a corporate data ecosystem do require full ownership and accountability. However, it's also reasonable to let some elements sit within the direct business teams and drive a stronger culture of self-service where appropriate.

Stakeholders will, typically, want to know there is a capability that can support them when required, but that they aren’t wholly beholden to (in their eyes) a technology function having to make a prioritisation call across multiple competing workloads.

Communication, communication, communication

Good project delivery correlates strongly with good communication. Both internally within the delivery team and externally to business stakeholders, being clear on progress, setbacks, and successes is critical for keeping people engaged.

Strong communication helps to drive enthusiasm and demand for data services and can include:

• Roadshowing data successes to help improve data literacy (understanding data and how it can provide insight and drive the right outcomes) 
• Encouraging business users to create new opportunities for exploiting data to deliver value 
• Onboarding users into new governance and project lifecycle frameworks, such as data lifecycles.

Don’t forget the cultural elements of your data strategy

Technology change in isolation doesn’t help anyone. As new solutions are being developed, the most important measure is successful adoption rates - tech, untouched, is not going to drive a business outcome. It’s important to work out what the right level of depth is dependent on the stakeholder - some business stakeholders have a passion for ML & AI and will want to understand some of the details, but other stakeholders will only want to see the results.

Creating a general data-literate culture in a business relies on hearts and minds, as well as capable technology platforms, and clear operating processes, and governance.

Embarking on the journey too early risks losing buy-in and enthusiasm. Incremental change, by delivering outcomes for a team or department, allows us to evolve our data strategy. By evolving the data strategy and its supporting pillars, we help to continually mature and validate our plans with real world experience, and therefore, continually deliver more value.

Data maturity assessments - make sure the target state fits your needs

One thing often lost is that data strategies aren’t one size fits all. Not every business needs to be scored a 5 on every pillar of its data capabilities - you don’t necessarily need to have cutting-edge capabilities in every domain before you start to see what each business team needs and start to deliver business value. 

In a similar fashion, it's not necessary to govern all data across an organisation to the highest standard. Instead, focus on the critical data sets first (e.g. legally/regulatory sensitive, such as PII data), then commercially critical data sets, such as ones that support key ML/AI use cases. 

It’s worth conducting maturity assessments (even light-touch ones) with business stakeholders. This will work out where they are and where they need to be. Furthermore, it will help build confidence through both an engagement visibility approach and ensure the strategy meets their needs and delivers the right outcomes. 

3. The execution stage - focus on feedback and building trust 

Once we’re out of the planning and definition stages, it becomes all about the ability to execute. Budgets will inevitably tighten, despite the best intentions, scope will shift, and potentially new disruptive technologies will come along and change what needs to be done - in a fashion similar to the launch of Generative AI tools in 2022. 

Multi-year plans will struggle to remain intact, and will need to show the ability to change and adapt to shifting sands while still retaining the overall goal and objectives.

Demonstrating incremental value, in an Agile fashion, is critical for unlocking further budget and being able to deliver on the bigger objectives. Suggesting major tech change, such as a data platform migration or building a significant suite of ML and AI capabilities, is much more palatable than investing upwards of £500k or more before any returns are deliverable. Similarly, waiting 12-18 months before business value is returned will not inspire any confidence from a business stakeholder.

Establish feedback loops and accountability

A good data strategy isn’t built in isolation, nor is it built once. Confidence in the data strategy also comes from its ability to be course-corrected while in-flight. Ideally, there should be no major changes, but adaptability is crucial. Being able to incorporate feedback from stakeholders and showing how their evolving needs and requirements are catered to will build confidence.

Establish clear feedback loops - both formal (e.g. steering committees, roadmaps, etc) and informal (coffee and progress catch-ups). These will ensure that stakeholders aren’t feeling pains that they either don’t feel comfortable voicing, or don’t find themselves having the opportunity to air. 

Convey progress and challenges clearly and simply - whether KPIs, delivery timelines, or cost overruns and resourcing challenges. Clarity will be key for continued buy-in.

Communication is particularly important for the first round of the data strategy during the initial outreach phase and the “ramp up” stage. It will help all the parties involved in the conversation build familiarity.

In my past data migration programmes, we did an initial engagement across several business departments to understand their needs, data fluency and use cases. As we moved into a more factory-like model of migration, we re-engaged with them to ensure their needs at a more granular level were being met. This also meant that our proposed changes were being delivered across all pillars of people, process and technology rather than purely shiny new tools which don’t end up being adopted. 

Where we see velocity decreasing, it’s critical to show accountability for resolving these issues and ensuring benefits are fully realised.

Share and showcase your successes

It’s critical to continually share successes across the organisation and convey the real, tangible wins and developments delivered. Sharing wins helps to drum up demand and confidence and makes the concept of data more real. 

Explaining how certain business outcomes have been achieved and showing a little “behind the curtain” will help other business teams understand how they can use data better as well. 

Creating an advocate community, especially with business contributors who need to use the data or see its outputs, is crucial for enduring success.

Embed data ethics and build data trust

In addition to making sure we can manage our data, it is critical to spend time - especially for applications that may have a material impact on staff or customers - understanding the following:

• Bias that may exist in the data collected
• Where the data has originated from (e.g. lineage)
• How it has been transformed and prepared
• What consent has been collected regarding the use of the data

Building awareness and processes to review the ongoing ethical use of data and being able to audit a model for its bias and fairness, for example, is vital in order to help drive sustainable, successful business adoption.

Resource, capability and skill development

It’s important that data doesn’t appear like an ivory tower for business teams; instead, it presents as an exciting new area of capability that people across the business can choose to develop and upskill. 

While concepts like the citizen data scientist haven’t necessarily yet come to fruition, and may not, providing more people across the business with the skills to interrogate data is critical. Enabling them to do this and build the dashboards and reports they need to get insight out of data will help to encourage its adoption and create more business value from data.  It will also help them identify ML and AI opportunities better, further increasing returns!

The considerations in this blog will help ensure that, at every stage of your data strategy lifecycle, you're building real confidence with your stakeholders - delivering measurable success in the short term while laying the foundations for longer-term transformation. 

A confident data strategy doesn’t need to be perfect from day one; it must be flexible, transparent, and grounded in business value. The strategies that succeed are those that evolve, adapt, and continue to earn support through delivery, engagement, and outcomes that matter.

Is your data function, infrastructure, and governance aligned with your business goals?

Find out with our Data Maturity Assessment.

Follow me on LinkedIn for more data insights.