L.E.K. Consulting has seen a surge in interest from our client organisations in becoming more data-driven and having the ‘right data’ — data that is integrated and trusted as the foundation for that approach. With headlines increasingly dominated by AI, executives want to know how they can leverage their data to improve their business performance and access the power of generative AI (GenAI) and machine learning (ML).
Data creates value through its application across analytics, data science and business intelligence, and ultimately to underpin data-driven decision-making, which increasingly underpins customer experience and operational excellence. Obtaining the right data and then using it to its full potential requires a targeted and well-defined set of commercial use cases focused on the critical issues your business faces. Use cases can range from simply understanding business performance trends by customer group to deploying advanced AI models to predict future customer behaviour or implementing GenAI to streamline routine tasks. Only when these use cases are underpinned by the right data can the full value of the data be realised.
Senior decision-makers within global organisations increasingly recognise that improving the coverage and quality of the data used within their companies can help in gaining competitive advantage and making their business more valuable. Across all sectors, having access to the right data can help ensure that the business proposition is optimised to meet customer requirements and that the operating model is streamlined to deliver that proposition effectively. However, frequently businesses face barriers including data silos and a lack of required skillsets and tools (including AI).
In this Executive Insights, we outline the steps that companies can take to rapidly unlock value from the data they already have and how they can improve their datasets, including the critical success factors in achieving this goal.
Progressing towards a higher level of data development
L.E.K. has identified four levels of progression for companies that are working towards a higher level of data development. Inevitably, the current and appropriate target levels of data development may vary across your business — e.g. by business unit, function or process.
Nascent: There is limited data validation or verification (very manual where it does exist), or not even a consensus on how relevant data should be defined. This can stem from the business having legacy in-house systems and processes that were not primarily designed with data collection in mind. A lack of ownership and awareness about the importance of data quality is also likely. This disjointed array of data assets can lead executives to distrust the underlying data due to quality or completeness issues.
Foundational: Some basic datasets with broad applications have been identified and validated, and there is a common data structure (or taxonomy) against which some business definitions can be consistently applied. However, at this stage, not all relevant data is being captured, and datasets can remain within different silos rather than effectively linked. Validation will be manual and costly, and senior management also remains unclear as to what other data should be collected, why and who is responsible.
Developing: Business logic is now being more consistently applied, and senior management is showing ambition in identifying and collecting new data to fill any gaps. Such companies also employ clean, up-to-date, commonly used datasets which have a long time series and are accessible to most obvious user groups. Initial projects for moving to a product-based approach to provision of data and insight may be underway, supported by multi-skilled teams with dedicated data, analytics and AI roles.
Pioneering: All relevant data has been captured and structured at the appropriate level. Such companies will have integrated end-to-end datasets that are easily accessible across the business, curated thematically by departmental need. There will be clear alignment and delivery progress on the data management and architecture required, including the data fabric (i.e. more centralised) or data mesh (i.e. more decentralised) to reinforce value streams. The company will use automated processes, leveraging the power of AI to ingest, validate, clean, and transform both internal and external data so that a consistent, long time series of data can be established to support decision-making.
Not all businesses will benefit from becoming pioneers in data development. The level of development just needs to be fit for purpose, i.e. to enable the commercial use cases that the business can valuably pursue. Aligning on a pace that makes it possible to achieve a certain level of maturity is a key element of an effective data strategy.
Resetting the approach to data
Time and resources are often wasted, and opportunities missed, due to a lack of alignment across the business on where to begin in collaborating on trusted data and insights. A key success factor is to facilitate a process of data strategy development which will ensure business goals and priorities are front of mind when aligning on priority use cases.
Identifying the right data for a business and what to do with it requires thinking more holistically and more commercially about the data the organisation has access to and uses. This typically requires changing mindsets and addressing long-held misconceptions and legacy choices. Also, the right data will depend on what level of data development has currently been reached: nascent, foundational, developing or pioneering.
L.E.K. has observed four common ‘data resets’ that often are required, almost irrespective of the present stage of data development.
Take business ownership of data: In too many organisations, IT functions are seen as primarily (or even solely) responsible for data, largely because data tends to be stored in IT systems. In reality, business processes generate data, and the entire business should benefit from quick and easy access to this data. Senior business ownership of data focuses attention on what is most valuable to your business and results in data that is both more relevant and more accessible. The fastest way to begin building this awareness is by deploying a small, multi-skilled data management team to unlock business-critical use cases.
Focus on quality, not quantity: Countless businesses fall into the trap of chasing data just because it is available. This can lead to organisations being overwhelmed by the three Vs of so-called big data: volume (the amount of data being collected), velocity (the rate at which data is received) and variety (the range of formats data arrives in). The focus, however, should be on data coverage and data quality. Even with the power of AI to parse, analyse and interpret data, the garbage-in-garbage-out principle still applies.
Define clearly how datasets work within the organisation: A wide range of companies (often under guidance from their advisors) define data quality too narrowly, without taking into consideration what variables may be missing from datasets, how datasets are linked and whether current taxonomies are commercially advantageous. Careful attention to these elements forms the basis from which the benefits of a professionalised data management service can be gradually scaled across products and markets.
Prioritise delivery of commercial benefit from data platforms: Data platforms can be essential tools, but their development all too often sees overruns in both time frame and budget, with the additional risk of ill-defined outputs and unclear business benefits. Clarity and commercial oversight are required to avoid platforms that host unwieldy data or software engineering projects with a low return on investment (ROI).
In summary, we believe that businesses need to focus on obtaining, improving and managing the relevant data that will not only best support the next stage of data development but also, critically, unlock new commercial opportunities and lessons about the skills and tools needed to make this approach successful.
Five characteristics of the right data
Getting the right data is a crucial foundation for most businesses and helps unlock value. Plus, it helps accelerate and ease decision-making around data migration, e.g. by reducing business user frustration and confusion.
Any assessment of data must take place within the context of the business need (particularly use cases), but we typically find that there are five characteristics of the right data that together drive commercial benefit (see Figure 1).
