Recently, I came across a paper by Todd Little entitled, ‘Schedule Estimation and Uncertainty Surrounding the Cone of Uncertainty’ and was intrigued by the findings. As an expert in agile software development and a member of the board of directors of the Agile Alliance, Mr. Little provides extensive mathematical analysis of projects completed while working at Landmark Graphics, and concludes that even seasoned project leaders provide estimates which typically fall significantly below actual results. He attributes this to a Cone of Uncertainty – a variety of unknown factors at the start of the project, including resource availability, unanticipated requirements and corporate cultures.
What he learned from his analysis of these projects, is that over time doubling estimates is more accurate – and more importantly, that uncertainly decreases and the estimates become more accurate as these factors became known. “Agile and Iterative development approaches acknowledge the presence of uncertainty and adapt to situations, rather than try to control situations to meet the plan.”
So, how does this apply to the implementation of store lifecycle management systems?
I can tell you after almost a decade working in the industry that doubling estimates, while mathematically may prove more accurate, would not generally be welcomed by clients constrained by a fixed budget. But there is a different way to approach this issue. By combining a hybrid-agile approach with a strong client-vendor partnership during contract scoping, clients can ensure transparency of hours spent throughout the project. This enables them to easily identify and acknowledge any uncertainty – the unknown factors that Mr. Little identifies — and then pinpoint what is in and out of scope to ensure the project stays on track and avoids cost overruns.
Investing time up front to detail each deliverable, each individual’s role and any assumptions can help to reduce the factors involved in the Cone of Uncertainty considerably. However, this is not a silver bullet. In the two examples below, I will demonstrate what can go wrong when the agile feedback loop is not met and what can go right when clients actively partner with their vendors.
Manage Scope Aggressively
In this first example, you can see what happens when scope is not managed actively.
A large retail client of ours undertook a significant Store Lifecycle Management implementation which lasted a year. Due to the culture of the organization, detailed deliverables and managing-to-metrics including periodic check-ins were not a valued part of the process. The culture was more focused on flexibility with little concern for the actual scope.
Well, you can guess what happened. Not surprising the client faced significant project overruns – 90 percent of which could be attributed to a lack of commitment to the process. While issues were documented and Tango pushed for change management – and asked for more attention to requests made to Tango on deliverables – the client’s lack of commitment to partnering on the unknowns and refusal to address change control was the number one contributing factor to project overruns. In this case, the outcome was roughly double what the original Phase 1 estimate had been.
Continuous Feedback Loop
Part of my role at Tango entails the development of a Cost Model for the offerings I lead. As most readers know, a cost model tool helps estimate work, which ideally feeds directly into a statement of work (contract). For this particular initiative, I developed a simple set of tools which scoped a project’s hours by module and by deliverable.
The benefit of providing the granular level of transparency — e.g. design vs. build hours for each integration, each Conference Room Pilot, each design document, etc. — to clients at contract signing and then through bi-weekly hours reporting throughout the project, helped identify where the ‘unknowns’ were within the “cone” and allowed us to either offset or proactively define alternate mitigation strategies. During scoping, we stepped through the model actively with the client adjusting based on their feedback to ensure alignment on hours and deliverables.
Layering in the hybrid-agile methodology provided near real-time visibility into where the hours were being allocated. It also demonstrated the results to the client, in the form of configured systems or documentation, allowing for better resource and cost management. We experience considerable success with this methodology, and the client quickly realized their resource gap and decided to pre-pay 1,000 additional hours to allow for expected future overruns due to their resource limitations.
We would not have been able to point them to the drivers of the cost discrepancy had we not initially partnered in scoping, where we aligned on hours/cost per minor component and continuously provided burn rate analysis tied back to the original scope. It helped that the client also cared about Change Management controls! But in the end the partnership and effective management paid off.
Cone Control
These examples illustrate that a proven methodology combined with a focus on client partnership delivers the edge on estimating and controlling the unexpected costs associated with SLM implementations. This combination is key in an environment where many IT leaders come from a large ERP background and try to approach store lifecycle management in the same way as other large packaged software implementations. The transparency and scoping facilitates the initial understanding in the learning process of managing large IWMS implementations, and helps clients actively mitigate the cone of uncertainty.
