Four steps in unlocking the value of data in Australian industrial organisations

| Article

Australian industrial organisations recognise data as a critical asset, but many are struggling to convert that potential to benefit their bottom line.

Faced with the dual obstacle of changing choices and complex legacy systems, organisations fail to take action or else embrace piecemeal investment in technologies without clear direction.

In this short insight, we look at how strategy, culture, and capability retooling set data leaders in this sector apart—and outline the steps needed to join them.

Australian industry provides some exemplars of how accelerating technology capabilities can capture value. In mining, for instance, the value gained from deploying advanced analytics at scale is clear, especially in the optimisation of plant processes or in predictive maintenance.

Because data are the fuel behind these opportunities, other industrial players are seeking similar capabilities. This has prompted acceleration in building data platforms, remediating source systems, and streamlining architectures.

Yet the sectorwide results are mixed. Despite strong efforts, most industrial organisations are falling short of capturing full value from their data—and they know it. In one survey, only 7 percent of Australian organisations rated themselves as ‘very effective’ at reaching their primary objectives in data and analytics, putting the ROI of these efforts in question (Exhibit 1).

Many organisations are falling short of their goals of obtaining value from data.

If this sounds familiar, your business is not alone. It isn’t surprising when we consider the simultaneous challenges industrial businesses face in delivering real impact from data.

In the face of this, is it any wonder that multitudes of hastily deployed ‘pilots’—loosely aligned (if at all) to governance and value orientation—can become the norm (Exhibit 2)?

This needn’t be the case.

A wide range of issues can prevent organisations from properly harnessing data.

Developing and obtaining up-front buy-in to a data ‘blueprint’ can underpin a data-investment approach that is strategic, measured, managed, and engineered to your organisation’s core priorities and products.

Developing the blueprint is not a simple task, but it’s a worthwhile one. If done right, it will balance value delivery and long-term capability building through:

  • A value-back data strategy. Start with a compelling vision—with executive buy-in—of how data and analytics will propel broader business strategy. This strategy must include an agreed-on business case and a two- to three-year road map of prioritised opportunities and their enabling capabilities.
  • A fit-for-purpose technology-infrastructure map. Modernising data architecture should not be a multiyear effort in which value is realised only at the end. A progressive approach should be adopted, with high-quality, ready-to-consume data becoming available over time in the format needed by your business. Reusability is a key for acceleration: often, a handful of data areas will enable most of the highest-value opportunities.
  • A robust data-governance model. Roles, processes, and tools to address data ownership, quality, security, access, and ethics (the ‘safety’ of the data world) must be put in place—and centrally understood. These elements may be rolled out progressively, focusing on those that enable the highest value first, but the full model should be clear up front.
  • A data-driven leadership culture. Building this culture requires understanding existing mindsets about data and, often, intervening. Many organisations still consider data to be an IT problem instead of a subject that should be integrated at senior levels. As in most transformations, culture is the hardest element to influence, requiring a mix of approaches, including role modeling, incentive alignment, and comprehensive change management and communication.
  • A deep-skilled, data-literate workforce. Securing tech skills is critical but not sufficient.

Many organisations have gaps in broad data literacy, reducing the potential for data-driven decision making and creating ineffective ‘internal clients’ for data stewards. Lead organisations are addressing these gaps by rolling out ‘data academies,’ with online training and informal learning tailored to existing roles and contexts.

Stop stalling. Start drafting!

Combining all the above factors can be daunting and often causes industrial (and other) organisations to stall and fail to take action.

So here are the four steps industrial organisations can take right away to avoid the ‘pilot purgatory’ trap.

Step 1: Identify the data you most need and the way you most need it

All data are important, but not all data are equal—nor are they consumed in the same way.

Examine your value chain and identify the points with the greatest potential for improvement from analytics, automation, digitisation, and so on—and map those opportunities to the data domains required to enable them. These are your priority data domains (Exhibit 3).

Identifying priority data domains is key in value creation.

Then identify how those data will be most useful to drive your technology. For example, plant optimisation may require ingesting data as rapid-stream computing parameters and passing these parameters to machines. The technical ability to deliver the data consumption your business needs should be prioritised over the value it can create.

Step 2: Empower a small team of high-performing experts focused on delivering one or two high-value opportunities

Data teams are usually structured in one of two ways:

  1. Large, centralised teams, working on ingesting and transforming data—with requirements ‘thrown over the fence’ by business, operations, or analytics teams. These teams are good at adhering to processes and standards but tend to be slow and expensive.
  2. Small ‘pirate crews’ formed by staff frustrated with a lack of easy access to high-quality, ready-to-use data. They find ways to extract raw data, save them in private repositories, and manipulate them based on their needs to generate insights (which, in some cases, are inconsistent with reports from other parts of the business).

We believe the sweet spot lies somewhere in the middle. We see leading organisations adopting agile delivery models in which ‘use squads,’ focused on delivering end-user functionality (for example, a truckload optimisation solution), work with ‘utility squads,’ focused on ensuring the required data foundations are in place (Exhibit 4).

Organisations should adopt agile delivery models that employ two different types of squads.

The advantage of bringing these groups together is that both tend to:

  • be highly integrated with the business
  • have a backlog prioritised by value
  • adopt a product mindset (rather than a project mindset)
  • rely on automation to reduce lead time

Organisations that go down this path start small with a core team focused on delivering one or two use cases with clear links to value. Early wins create excitement in your teams and help identify lessons that can be used to scale as demand grows.

Step 3: Selectively modernise your data architecture, leveraging new approaches to scale up

Some organisations are pivoting from central-enterprise data to a domain-led architecture, leading to improved time-to-market on data-driven services and products.

This requires up-front effort to design the architectural capabilities for each data domain—but in the long run, it reduces the risk of fragmentation and inefficiency and simplifies the construction of data models, accelerating the enablement of data services.

Domain-specific architectural elements can be deployed and replicated easily by leveraging ‘infrastructure as code,’ which allows for rapid scaling both within and across data domains (Exhibit 5).

Deploying domain-specific architectural elements allows for rapid scaling within and across data domains.

Organisations adopting this approach start by focusing on two or three high-priority data domains and the associated data-consumption archetypes.

Step 4: Repeat and scale

Successful organisations progressively build the muscle needed to use data effectively, approaching it as a continuous improvement journey. A critical success factor is ensuring that the approach is codified and repeatable.

The first set of use cases is extremely important for standardising the approach and helping build momentum in your organisation. These early wins attract more internal buy-in and avoid the dreaded ‘technology-first’ mindset.

Being data-driven and digitally enabled can now make or break industrial businesses.

‘Sweating’ the value of your data requires a staged approach, balancing high-value opportunities with long-term capabilities, including a data-literate workforce and a flexible architecture that supports your organisation’s objectives.

While we know (and have seen firsthand) the challenges in achieving this, we believe that a ‘value back’ data blueprint, paired with a rigorous execution anchored in agile delivery, gives industrial businesses the best chance of data success.

Explore a career with us