Case Study 1: Private Health Insurer
Case Study 2: FTSE 100 Investment Company
Case Study 3: FTSE 100 Retailer
Note: a worked example of an agile transformation cycle is also available (PDF).
Unfortunately the associated agile practices struggled to gain traction, despite competent and knowledgeable coaches having been hired. The agile adoption initiative was effectively an unlimited work-in-progress. More manageable changes - discretely achievable and clearly focused - had not been identified or prioritized. Coaching was largely reactive and lacked direction, and the progress of the transformation was difficult to assess empirically. In short the organization lacked the control surfaces that are needed for a measured and managed approach to agile change.
Team-level symptoms included unclear timeboxed goals, and planning was confused with requirements refinement. Product ownership was disengaged from the development process, since incremental delivery was neither understood nor valued. There was no agile governance and team composition was subject to frequent and unforeseen change. The relationship between Business and IT was poor.
The agile transformation pattern was implemented so that the change program could be controlled and made empirically assessable. A DSDM/Scrum "target operating model" was configured, and which expressed the organization's sponsored vision for its future agile way of working. A transformation team was then identified. It consisted of a sponsor, coaches, scrum masters, business change agents, and managers with a stake in organizational accountability and governance.
With guidance from the coaches, the transformation team identified more than eighty discrete changes that would be needed to implement the vision. These were articulated with clear acceptance criteria and prioritized on the transformation backlog. A physical task board was used to help manage the backlog and provide visibility over work in progress, and regular "standups" were held to a weekly transformation heartbeat. Throughput metrics were harvested. A date for the implementation of the operating model was then forecast and a transformation burndown was tracked. Impediments to the transformation were exposed and their resolution was facilitated.
* One of the items on the transformation backlog was to identify and train in-house coaches. As an externally validated measure of competence, they achieved relevant scrum.org accreditations (PSM, PSPO, PSD). The client was then in a position to further optimize the transformation using in-house coaching resources.
Unlike the first study, there was no high-level sponsorship for transformation at all in this case, even at an IT departmental level. Certain middle managers with long service at the company were becoming increasingly concerned that the organization would not survive unless product and service innovation was improved, and the validated learning cycle closed. Senior executives were not opposed to such change and broadly acknowledged the threat. However, there was political resistance as there had been a failed agile transformation some ten years beforehand. Senior executives preferred to delegate the responsibility for any improvement...and hence the risk of any further failures...to those middle managers.
Sponsorship for agile practice therefore came from:
This represented a transformational challenge. Without clear executive sponsorship, only local optimizations can be expected which do not empirically and clearly evidence improved product value. Benefits are limited: a bit of improved transparency here, perhaps some better accountability there. This can be worth doing in itself, on the basis that it still amounts to progress of sorts, but it is non-transformational in nature. Clear organizational sponsorship is needed for real enterprise transformation. Organizational inertia must be overcome, and if that is to be achieved, senior executives must be able to transmit a sense of urgency for change.
Although organizational sponsorship was weak, enough was in place to establish and coach two clear Scrum Teams for pensions and related services. They did not own the entire journey into service for their products, but they owned enough of the respective value streams for the delivery process to be inspected, adapted, and rendered transparent. Support was also obtained, from the department's quality assurance function, to conduct regular agile health checks on cadence. Metrics concerning the delivery of increments for potential release could then be elicited and the process of transformation subjected to a degree of governance. Empirical evidence of the likely benefits of further agile change could also be gathered.
The agile transformation pattern was implemented so that this pilot initiative could be empirically controlled. A transformation rollout team was established consisting of a middle manager as the immediate sponsor. This sponsor had a vision for the improved delivery of value at the level of pensions and related products. The team also had Scrum Masters and a senior coach and a transformation backlog of around sixty discrete changes was elicited. These were held electronically in an Excel spreadsheet and a transformation heartbeat was observed during which progress was inspected and adapted, and the transformation backlog revised and reprioritized. Agile health checks were conducted on-cadence by the department's quality assurance function.
Tangible Benefits Obtained
The client also had a pre-existing strategic directive to use agile ways-of-working wherever possible, and hence an agile coaching capability was brought in at the beginning. It should be noted however that the connection between agile practices and an improved ability to innovate was not clearly understood by client stakeholders, at least to start with. Also, there was no clear governance or assurance function, with no clear oversight of the transformational approach, or of quality concerns such as technical debt.
One requirement stipulated by the client was that there should be a pre-delivery "Discovery Phase”. During this phase, it was expected that a Product Backlog would be elicited at epic level, the essential agile ways-of-working laid out and socialised as far as possible, and an architectural baseline established which could be used by delivery teams. Development and delivery infrastructure, including continuous delivery and integration environments, and automated test harnesses, would also be instantiated. In effect, the Discovery Phase was an opportunity to put an agile transformation framework into place gradually, and with less sudden change to the client.
Since a Discovery phase had been provided for, which is unusual in transformation initiatives, the agile transformation pattern could be assembled without haste. Its various components were thus introduced and explained to the client before they had to be put into practice.