Project portfolio management is not a new concept. As technology executives, we all do it; and chances are that we are good at it based on traditional norms. This article is to redefine the art of the possible for project portfolio executives during digital transformation, and offer them a few new techniques.
Prior to digital transformation, the primary goal of portfolio executives was to maximize the delivery of technology outputs within budget and schedule. This IT-centric mandate emphasized output over outcome, and risk over return. Furthermore, the traditional IT financial framework was essentially a cost-recovery model that wasn’t suitable for portfolio executives to articulate how to maximize business outcomes on technology investments. Consequently, portfolio management was marginalized to a bureaucratic overhead and a nice-to-have extension of the program management function.
In the digital age, a robust portfolio management function has become a strategic necessity:
- Pace of business change has accelerated: the window of digital opportunities is short-lived; and, without considering the opportunity cost of delayed projects, i.e., cost-of-delay, investment mix decisions are distorted. Case in point: a recent WSJ article explains how JPMC is challenged with mobile payments due to quicker competitors;
- Iterative development is replacing plan-driven project execution: Cost, risk and return trade-offs are often unknown during initiation and continuously change during execution;
- Product-centric IT operations offer new opportunities to optimize operational trade-offs, i.e., resource allocation, work prioritization, scheduling; and to implement advanced risk mitigation techniques at a portfolio level, i.e., pooling, diversification and hedging.
The size of the opportunity
Many enterprises are yet to upgrade their project portfolio management capabilities as part of their digital transformation efforts. Consequently, there lies a significant unclaimed value in technology portfolios. Consider the following scenarios:
- Work done but no longer needed, or not aligned with the portfolio objectives;
- Work incomplete and left to age in team backlogs;
- Non-mandatory work performed under mandatory programs;
- Systemic scope inflation, and schedule slippage;
- Systemic underestimation of work size;
- Systemic capacity and demand mismatch;
- Unmanaged cross-team dependencies.
Most IT portfolio executives agree that project portfolio performance can be improved. Some even think that a 20 to 30 percent improvement in return should be plausible. The challenge, they say, is how to get there.
We have been working on this problem for a while now. We recently developed a robust technique to quantify the unclaimed value in various kinds of technology project portfolios in a repeatable and verifiable way. Our research indicates that most of this value is associated with dormant output, cost of delay and system friction like a supply and demand imbalance. For a detailed description of these terms, please refer to “The complete economics of running IT as a business.” Based on the business value framework introduced in that article, we estimate:
- More than one-fifth of technology outputs frequently become dormant prior to deployment;
- In a high pace digital economy, cost-of-delay can account for up to 40 percent of the potential business outcomes; and,
- Predictive modeling of resource capacity and project demand can yield a double-digit productivity gain.
During our research, we identified a number of management techniques and decision algorithms to get to the source of the portfolio value. The following techniques are worth elaborating here:
Investment mix optimization
Based on numerous empirical studies, we concluded that the relationship between technology investments and business outcomes is non-linear. When demand significantly exceeds capacity, more work remains in progress longer. When there is too much capacity, low value work creeps in. In between these two cases, there is an optimum point where the marginal portfolio throughput peaks. Although intuitive, this finding invalidates a common practice, that is as long as teams work on the highest priority item in their queue, the overall productivity of a portfolio is maximized. (See here for a detailed discussion on this finding.)
The position of the peak marginal portfolio throughput is a vital insight for portfolio executives when:
- Articulating the true enterprise impact of a portfolio budget cut/increase;
- Addressing projects that are likely to underperform before they hit a trouble spot;
- Articulating the cost of systemic scope increases.
Projects that are completed ahead of schedule won’t create additional value, while delays often reduce business outcomes, i.e., cost-of-delay. Therefore, the most effective scheduling algorithms are those that exploit the slack time of each task to minimize required team size by balancing early and late activities without jeopardizing committed milestones.
It is important to note that many traditionally run portfolios report on their average delay at program and project level, which is often misleading. Because, true optimization opportunities are often hidden at a lower level, such as activity or task; and they are realized by accelerating delayed tasks at the expense of early tasks. Consequently, managing the combination of the standard deviation and the average of portfolio delays at a granular level is a more effective way of optimizing schedules.
Typically, projects are assigned a priority during the approval or initiation phases, and any decomposition of their scope inherits the priority of the parent project, e.g., epics, features and stories. During digital transformation, prioritization decisions require precise, granular and timely insights:
- Expected business outcomes need to be dynamically updated based on customer feedback and account for cost-of-delay;
- Decomposed scope should be prioritized independent of its parent, e.g., a story of a high priority feature should not be automatically considered as high priority. Our studies confirm that broad-stroke prioritization decisions considerably inhibit portfolio returns;
- Advanced prioritization schemas should be carefully evaluated, e.g., agile project portfolios that leverage the SAFE methodology are encouraged to implement WSJF (Weighted-Shortest-Job-First) prioritization algorithm. Our studies confirm that this algorithm performs well when job dependencies within a portfolio are negligible, but the algorithm becomes quickly counterproductive when the number of dependencies increases.
Demand pooling, supply diversification, risk hedging, contingency allocation and predictive analytics are a few management practices successfully utilized in other industries like manufacturing, retail and finance to address similar challenges. If thoughtfully embedded into portfolio resource planning and management decisions, these advanced practices can yield significant reductions in bench-time and on-demand use of premium-priced resources. In our empirical studies, we were able to consistently achieve a double-digit productivity improvement through the use of similar practices.
Effective forecasting is a game changer
During digital transformation, project portfolios face a higher degree of uncertainty because of the accelerated pace of business change, perishable nature of digital opportunities, increased and accelerated customer feedback loops embedded in the iterative development, and increased cross-team dependencies. Fortunately, a significant portion of this uncertainty is derived from deterministic patterns of IT operating environment, customer behavior and market trends; and therefore, it is predictable.
In fact, knowing ahead of time when a project, activity and task would be completed, gives portfolio managers a strategic advantage in planning and executing portfolio optimization activities. Specifically, effective forecasting will enable portfolio executives to implement multi-sprint resource plans, manage cost-of-delay, minimize dormant output, and continuously maintain the portfolio at its peak throughput.
When I talk to technology executives about the benefits of effective forecasting, I often get two questions:
Do I have enough data to feed in?
Most IT organizations have already implemented a project portfolio management tool, e.g., CA PPM, Planview, or HPE PPM for waterfall projects, and an application lifecycle management (ALM) tool, e.g., Jira, IBM Rational, CA Agile Central, for agile projects. These tools collect a wealth of information that can help identify detailed operating patterns and customer behaviors. Regardless how these tools are put in use, their data, if harvested well, is sufficient to enable forecasting.
Aren’t my current forecasting capabilities already sufficient?
It depends; there is a simple test to get to the bottom of this question quickly:
If so, your current forecasting capabilities are sufficiently effective.
One of the most important and influential economic theories dealing with finance and investment, modern portfolio theory (MPT) was developed by Harry Markowitz and published under the title “Portfolio Selection” in the Journal of Finance in 1952. The theory is based on Markowitz’s hypothesis that it is possible for investors to design an optimal portfolio to maximize returns by taking on a quantifiable amount of risk. MPT states that the risk for individual stock returns has two components:
- The systematic risk – e.g., interest rates, recessions and wars – is the uncertainty inherent to the entire market and cannot be diversified away.
- The unsystematic risk – e.g., mismanagement or decline in company operations – is specific to individual stocks and can be diversified away as the number of stocks in the portfolio increases.
During digital transformation, project portfolios resemble stock portfolios, because projects are increasingly exposed to system-wide factors – operating model, vendor ecosystem, customer behavior, business and technology trends to name a few. Consequently, project systematic risks are significant and distinct from the project unsystematic risks, e.g., project leadership performance, team issues, etc.
Traditionally, IT organizations have handled project risks at the program and project level. For example, individual project plans are padded with covert schedule and cost buffers due to no tolerance for failure; operational constraints and trade-offs trigger long debates among managers and frequently require executive intervention; mitigation actions create workarounds without permanent solutions.
Our empirical tests reveal that there is a significantly better way of managing project risks during digital transformation by applying the key principles of MPT to technology project portfolios:
- Many of the systematic risks are predictable and can be eliminated.
Portfolio executives can address the root-causes of systematic risk factors like operating inefficiencies, outdated policies, and unproductive interactions.
- The unsystematic risks can be diversified, pooled or hedged at the portfolio level. Some tasks delay and others move smoothly; some solutions end up unexpectedly complex and others simple. Some teams perform better than others. Hence, contingencies should be sized according to a portfolio level variance rather than the likelihood of a failure at any project or program.
- Program and project managers should continue managing the unsystematic risks of their activities within the risk-tolerance limits of their portfolio.
By influencing the IT leadership to adopt MPT principles and to implement the advanced techniques discussed in this article, portfolio executives can visibly improve the business outcomes enabled by their project portfolio. An effective forecasting capability is a game changer in a digital world; and hence, it serves as an excellent starting point to articulate the case for change for the portfolio management function to business and IT executive stakeholders.
This article is published as part of the IDG Contributor Network. Want to Join?