Data Roadmaps
Choosing The Right Initiatives
Many companies struggle to go beyond simple reporting and generate real, actionable value. So how can you ensure that your data initiatives make a tangible impact? How do you involve business stakeholders in this journey, excite them about data’s potential, and set realistic expectations regarding outcomes and timelines? More importantly, which data projects should you pursue to maximise business value? Here’s a step-by-step guide to achieving meaningful data transformation in your business.
Step 1: Understanding the "Why" of Data
Before embarking on a data initiative, you must clearly define why your company needs data. While most businesses will say it is to make better decisions, data can provide five distinct advantages:
1. Holistic View
Integrating data from various systems, channels, and tools creates a comprehensive picture of your business operations. This clarity allows for better decision-making across departments such as finance, marketing, procurement, and operations.
2. Agility
Recognising early signs of change enables proactive responses. Initially, data initiatives focus on reacting faster to change, but over time, they evolve toward anticipating and predicting trends.
3. Customer Closeness
Data insights reveal customer demographics, preferred channels, lifetime value potential, and risk of churn. Understanding these factors allows businesses to enhance customer experience and retention.
4. Streamlining & Cost Reduction
Data helps identify inefficiencies, automate processes, and allocate resources more effectively. This leads to smarter spending and improved ROI.
5. Automation & AI
Data-powered automation simplifies decision-making, whether through AI-driven recommendations or automated workflows.
Actionable Step
A common problem is that different stakeholders have different priorities and often whoever shouts loudest gets the most attention. A practical exercise is to have stakeholders distribute a total of 100% across these five benefits based on their priorities, and then add them up amongst the group. This later allows you match your data ideas to the priorities of the business, and ensures that you don’t get tied to one particular benefit. (For example having to find a cost reduction with every initiative.)
Step 2: Background Knowledge
The best data initiatives are those that are embraced across the organisation, not just mandated from the top. Idea generation should involve:
- Management for strategic alignment
- Heads of Function for operational insights
- End-users who will work with the data daily
We have found that in order for people to embrace change they need to feel ownership of that change. If data is just a mandate from management, it won’t get the buy-in it needs to be successful. Encouraging input from all levels fosters a sense of ownership, which is key for adoption. However, for meaningful contributions, stakeholders should have a basic understanding of data pipelines and their potential applications.
Useful knowledge to frame your discussions:
- How does a data pipeline work?. What are the different stages/tools?
- Some basics around cost-implications – ie Stakeholders often want ‘real-time’ data but may not actually need to leverage that data at such a frequency.
- An idea of typical timeframes for projects – or at least which aspects take longer and which are easier.
A little bit of knowledge can go a long way to align stakeholder expectations and later in democratising data.
Step 3: Idea Generation
With a strong understanding of what is important to the business and how data actually works, you can begin formulating ideas! There are no ‘wrong’ answers at this stage and it is important to encourage contribution. The point of the exercise is to get people excited about data, map out different areas/ processes where it might be useful & uncover areas for integration. Try and foster conversations around data touchpoints, sources and where data can be activated.
A broad ideation phase can be daunting for data teams, as not every idea will be feasible. This is an opportunity to educate stakeholders about what is possible and guide discussions toward realistic and impactful initiatives.
Expert help is only a call away. We are always happy to give advice, offer an impartial opinion and put you on the right track. Book a call with a member of our friendly team today.
Step 4: Establishing Priorities for Further Review
Once ideas have been generated, they must be rigorously analysed for feasibility and impact. But before you go into each and every idea, you should first:
1. Refine the Idea
A strong data initiative should focus on driving decisions and actions rather than just identifying problems. For instance, predicting customer churn is valuable, but it must be accompanied by clear interventions, such as targeted retention campaigns. How does the data ladder up to a decision?
2. Popularity Among Teams
While some ideas may be more complex, those that generate excitement among employees can improve adoption and long-term success. While these may be bigger ideas for tomorrow, you can start working towards them today and creating data products that teams will enjoy using should always be a priority.
3. Alignment with Business Priorities
Refer back to the business priorities exercise. For example, if customer closeness is the main goal, then customer analytics projects should take precedence.
4. Business Impact and ROI
Early-stage assessments can help estimate potential returns and set KPIs to measure success.
5. Implementation Complexity
Certain initiatives may require foundational work before they can be executed. Understanding dependencies helps in structuring a realistic roadmap.
After an initial review, you will of course need to conduct a deeper feasibility review, but a top-level look is useful so as to provide feedback from the session quickly.
Step 5: Feasibility Review
You have engaged your stakeholders, gathered ideas, handpicked the most relevant ones and now comes the feasibility review which is crucial in ensuring that all data requirements can be delivered. So how should you go about it?
There are eight key areas of review that each of your analytics projects must go through:
Data Availability And Quality, Technical Feasibility, Resource Assessment, Regulatory And Compliance Considerations, Risk Assessment, Cost-Benefit Analysis, Activation & Alternative Solutions
Data Availability And Quality
The first step in many of our projects is a Data Dictionary. A data dictionary captures and defines KPIs across all business functions. It outlines the business and technical definitions for each metric along with other useful information. This should provide the blueprint as to which sources of data are needed for the data initiative, how complete that data is, the accuracy and consistency of the data, and any timing issues in terms of data availability.
To ensure the quality and completeness of the Data Dictionary and the underlying data, we conduct spot checks as part of our QA process. These spot checks typically involve:
Sampling key metrics from the data dictionary and validating them against raw source tables to ensure the logic accurately reflects the intended business definition.
Cross-referencing metrics across functions (e.g. Finance vs Product vs Marketing) to check for alignment and identify discrepancies in definitions or calculations.
Checking for nulls and anomalies in core columns and joining keys, particularly for high-impact metrics, to confirm the underlying data is reliable and complete.
Verifying timestamp consistency for time-series metrics to ensure proper alignment between event timestamps, ingestion times, and reporting schedules.
Confirming the presence of all expected dimensions and fact fields in the raw data, based on the metadata outlined in the dictionary.
These checks help to proactively catch issues before downstream analysis or dashboarding work begins, saving time and improving trust in the outputs.
Technical Feasibility
When considering a new idea or initiative, it is crucial to evaluate it against your existing technology stack. This helps you understand whether your current tools and infrastructure can support the desired outcome, or whether introducing something new is justified. Compatibility, ease of integration, and existing team expertise are key factors here. Even if a tool looks promising on paper, it may not mesh well with your current data architecture or operational workflows, and the cost (in time, complexity, or money) of forcing that integration can outweigh the benefits.
If you are implementing a new technology, or using an existing tool in a way that it was not originally intended for, it is worth doing some early-stage due diligence. This can include reviewing product forums, FAQs, documentation, and real-world use cases to identify potential limitations or known issues. Vendors can often oversell capabilities, implying that everything is seamless and scalable. It is easy to get caught up in a polished demo or a sales pitch that glosses over the reality of setup complexity, incomplete feature sets, or painful edge cases. To mitigate this risk, it is a good idea to carve out some time to play around with the tool yourself. Create test environments, replicate simple versions of your use case, and evaluate how intuitive, flexible, and robust the tool feels in practice.
Another major factor when assessing technical feasibility is scalability and long-term viability. Your data volumes, user base, or operational complexity are likely to grow over time, can the tool or architecture handle that growth efficiently? You should examine limits on data storage, processing throughput, or user concurrency, as well as how those limits affect pricing. Tools that work beautifully on small datasets can become unstable or cost-prohibitive at scale. Additionally, it is wise to think ahead: if you outgrow the tool or the vendor changes direction, how painful would it be to migrate away? Tools that lock you into proprietary formats or have limited data export options can create long-term friction and technical debt.
Taking the time to explore these considerations at the outset helps reduce the risk of surprises later, when making changes is more expensive and time-consuming. In short, balance ambition with realism—test, validate, and plan for the future.
Resource Assessment
Before kicking off any data project, it is vital to ask: who is going to work on this and how long will it take? Assessing internal capabilities is a crucial first step, do you have the technical expertise in-house to deliver the project, or will you need to invest in training or hiring? Even if the right skills are present, team availability often becomes the limiting factor, especially when analytics teams are already balancing multiple priorities. These competing demands can lead to delays or even cause projects to be paused indefinitely, so it is important to have a clear picture of what’s realistically achievable within your current resourcing model.
Estimating timelines can be particularly challenging if the project touches on a new area, unfamiliar data sources, or introduces new tools. That is why we always recommend breaking the work into smaller, manageable chunks, each with a clear and tangible deliverable. This modular approach not only helps reduce risk and complexity but also creates natural milestones for review and reflection. If priorities shift and work is paused, having those defined chunks means you are never restarting from scratch, you can pick up where you left off. It is a practical way to build momentum while staying flexible in the face of inevitable changes in focus.
Regulatory And Compliance Considerations
Depending on your industry, there might be specific regulation and standards in place that you need to adhere to and may affect your way of working. As well as ensuring that the project adheres to relevant data protection and privacy regulations, you should also consider any wider ethical issues. Many companies try and get as much information as possible on their customers and sometimes use blurry lines or fuzzy logic to justify their actions. It is important to safeguard the company from any potential fines or damage to their reputation and doing this at the feasibility phase is important.
Risk Assessment
Outside of regulatory considerations, it is important to evaluate broader risk factors that could impact the success or stability of your data initiative. This is especially critical when you’re planning to change core elements of your data infrastructure, analytics tooling, or foundational models. These components often have far-reaching dependencies, and even small changes can create unexpected knock-on effects. Before proceeding, take time to map out all of the connected elements, such as upstream data sources, pipelines, models, dashboards, and reporting processes. This exercise helps surface the full scope of impact and gives you a clearer picture of what’s truly at risk if something goes wrong.
Beyond identifying what could break, you should also pinpoint key points of failure within your current pipeline, steps that are either fragile, highly manual, or business-critical. These should be treated with extra caution, and any changes to them should be subject to more rigorous testing and validation. Implementing rollback plans, automated alerts, and version control can help mitigate potential issues. At the same time, it is worth recognising that inaction can also be a form of risk. Failing to modernise outdated systems, improve fragile processes, or migrate away from legacy infrastructure can leave you exposed to higher long-term costs, technical debt, or missed opportunities.
Cost-Benefit Analysis
This is a clear area in which your data team will need the input of Subject Matter Experts in order to understand the true impact to your business, but it is still difficult. Analytics typically underpins success and informs decision-making. So it is the end decision that gets all the headlines whilst the data can be forgotten about. Management often has an expectation that each and every data process or model needs to generate a significant return on investment in order for the expense to be justified, but it might take weeks or months to really see the impact in surfacing those insights. Benefits typically boil down to:
Cost-Savings: Reducing manual work, using data to find wastage, improved resource utilisation.
Revenue Generation: Increasing opportunities, conversion, sales, retention and lifetime value.
Time Savings: Efficiency gains, reducing manual work, faster processes.
We would typically advise that while it might be tempting to create big numbers on impact in order to sell a data project into your business, it is always better to err on the side of caution rather than promising something that is not delivered. While the benefits and ROI can often be fuzzy to understand, costs are normally straight forward:
Tools: Costs associated with hardware, software, licences etc. As we spoke about before, it is important to think about scalability when mapping out potential costs. Also consider maintenance and support costs.
People: The cost of people working on your data initiative, freelancers, agencies etc.
Activation
A critical part of any data initiative is considering how end users will actually adopt and engage with the solution. It is one thing to build a technically sound dashboard or report, it is another for it to become embedded in day-to-day decision-making. Start by looking at the tools, apps, and workflows your teams are already using. Can your solution integrate seamlessly into those environments, or does it require people to learn something new or switch contexts? The more friction there is, the less likely it is that people will use it regularly. Even if stakeholders ask for a dashboard, that does not necessarily mean it is the most effective format, especially if current dashboard usage is low. In these cases, consider alternatives like automated reports, alerts, or in-tool data summaries that surface insights directly where work happens.
Your Feasibility Review should go beyond simply asking “is this technically possible?” and instead explore what is the best way to bring the idea to life in a sustainable, usable format. Activation is often one of the biggest failure points in data projects, not because the insights are not useful, but because teams underestimate what it takes to drive engagement. This means investing in training, internal promotion, and clear onboarding materials. It also means having champions who can reinforce the value of the solution and embed it into team rituals and decision-making processes. A well-designed solution that nobody uses still counts as a failed project, so ensure that adoption and usability are part of your planning from day one.
Alternative Solutions
Last but not least, it is important to step back and consider what alternative solutions might be available to address the same problem. Just because a particular approach has been requested, or seems like the obvious next step, does not always mean it is the most effective, efficient, or sustainable option. In some cases, the original problem can be solved with a simpler method, a different tool, or by leveraging something that already exists within the business. Exploring alternatives allows you to compare potential solutions not only on feasibility, but also on cost, speed of implementation, maintainability, and long-term value.
This kind of comparative thinking can also uncover opportunities for greater alignment or reuse. For instance, is there another team tackling a similar issue that you could collaborate with? Is there an underused dashboard, report, or model that could be refreshed or repurposed? Or perhaps the core need is more about process change or stakeholder communication than building a new data product. Being open to different paths forward allows you to focus on solving the real problem rather than just delivering a specific request.
Conclusion
This deep dive takes you through the entire process of gathering and evaluating data ideas. These recommendations can then form part of a wider data roadmap giving a clear overview to your organisation on what data has accomplished and how.
If you need help evaluating your data ideas, understanding which projects will bring the most value and what a realistic timeline may be, why not get in touch with the 173tech team today?
