For business leaders, getting from the vision to a new operational target model is challenging and fraught with risks. The same McKinsey study revealed that 70% of these change programmes had failed to achieve their target impact. Previous surveys have revealed the common failure points when mapped against both scale and time (see Figure 1).
With data and analytics still at an early stage of their adoption, any role they are expected to play in transformations creates a double risk – firstly, where being used as a change agent in their own right if value fails to be delivered and, secondly, as a potential scapegoat for a failed transformation. While the current mood around D&A is highly positive, it is important to recognise that the function does not have ownership of any specific business process. As a result, it is very exposed if and when failures occur.
Understanding root causes of failure
1. Technology adoption curve
As a transformative technology, data and analytics are both enablers and disrupters. To practitioners, they appear to be logical, evidence-based and objective. To the areas of the business being asked to adopt them, however, they can appear challenging, unreasonable and threatening. This is especially true for any process area which has no history of evidence-based decision-making or data-driven activity.
When developing the roadmap for a data and analytics-oriented transformation, it is important to bear in mind the technology adoption curve (see Figure 2). First posited by Everett Rogers in his book, “Diffusion of Innovations”, in 1962, this model has been repeatedly validated in academic studies and proven to be robust. While the specific distribution of segments can vary depending on the nature of the technology being adopted and the incumbent culture, the underlying bell curve persists.
DataIQ Leaders are, by a process of self-selection, undoubtedly in the first two quintiles as either innovators or early adopters. Yet the individuals and teams they will be engaging with across the organisation are more likely to span the early majority and late majority segments. For the first of these, rational persuasion is usually appropriate with clear use cases developed to explain why new data sources, methods of processing or distributing data, analytical tools or techniques have become necessary.
The second group is more challenging because their adoption is based more around a sense of having no other choice. Often, this is an accurate assessment, since it is these types of roles (middle management, manual processes, “gut feel” decision-making) which are due to be transformed.
In discussing the problem of failing transformations at a DataIQ Leaders event in Summer 2018, this group was identified as creating a significant risk – as colleagues in this segment are unlikely to have visibility of the full transformational scope and may not be the first to experience value from it, they are likely to voice concerns and dissatisfaction which can spread quickly unless challenged.
Introducing new practitioners from outside of the organisation or even its traditional industry sector is likely to exacerbate the fear of the unknown which is triggered. Incumbents can quickly start to feel that they are one step behind the curve, are struggling to keep up and even that they are not valued as highly as new arrivals.
Solutions:
- Provide detailed delivery timescales for data and analytics roll-out
- Sustain internal communications programmes across the transformational journey
- Build local ownership of specific new data-driven processes or tools
2. Expectations and hype
Transformations are usually announced from the top-down, wrapped in a vision of how the business is going to overcome current obstacles and achieve significant new goals. Very occasionally, a transformation may start bottom-up, with a successful business challenged to rethink its model in the face of market disruption or significant regulatory change. Data and analytics may be the catalyst for this through identifying areas of opportunity or cost.
The first type of launch is the natural inclination of any organisation with a strong leadership model, especially one that has external investors and stakeholders who expect constant adaptation. By painting an optimistic picture of where the company will be in three years’ time, the intention is to build sufficient internal energy to sustain the change management programme across that period.
Often, this is accompanied by an initial “Hollywood project” to grab attention and demonstrate the new way in which data and analytics is intended to be used. The focus on this can be important as it will provide “show, not tell” evidence that the new strategy is working. Equally, it is high risk because a) the project may not deliver the expected value and b) it will typically only benefit one part of the business if it does succeed.
While creating an atmosphere of expectation can be beneficial, it may also simply precede a slump when the next phase of implementation turns out to be more complex, time-consuming and less impactful. As one DataIQ Leader pointed out, “after the quick wins, you are cutting into the old wood and incurring costs against cost-savings. Many of these elements of a data-driven transformation are foundational, not revenue-driving.”
When the classic hype cycle is mapped against the technology adoption curve, it is notable how closely the “trough of disillusionment” fits into what is routinely described as “the chasm” – the gap when early adopters start to struggle with a new technology and may not be able to explain its benefits to the early majority (see Figure 3).
Two approaches have been adopted by different DataIQ Leaders to mitigate this challenge. The first is to initiate data and analytics-driven change “under the radar” – running smaller-scale projects which provide proofs of concept for scaling up and win political support from stakeholders across the organisation.
The second is to adopt fully-realised project management techniques which give visibility to everything that is involved and impacted by the transformation. A high-resolution roadmap of this sort would show the relationships between all the components and their inter-dependency. By doing this, the actions needed to make something effective when it lands, such as the data, compute power, platform, doctrine and processes involved, can be identified and any points of failure called out. This is then used as a through-life manual until the completion of the transformation. Most importantly, each aspect of the roadmap should be assigned to a specified report who is accountable for its delivery and measurement.
Solutions:
- Identify second-wave initiatives that can land relatively soon after the quick wins
- Ensure responsibility has been assigned for every component of the roadmap
- Communicate successes and have open, no-blame debriefs on any failures
