Data provided Great Britain a strategic advantage in World War II. Likewise, project data can provide your business strategic advantages today.

As a general rule, the most successful man in life is the man who has the best information.

Benjamin Disraeli (1804 – 1881), British Prime Minister

Bletchley Park

During the early days of World War II, the German war machine with its Blitzkrieg (“lightning war”) overwhelmed Europe, leaving the island nation of England standing alone to halt the threat of Nazism to the world. On the surface, the two forces were ill-matched. In June 1941, the Germans had more than 7.2 million soldiers, most of whom had battle experience. The Germans had superior numbers in armaments (tanks, artillery, and first-line military aircraft). More importantly, they had spent years preparing for war under the noses of the Allies with extensive planning and an efficient command organization.

Britain’s forces in 1941 consisted of 2.2 million men, doubling since 1939 with a mandatory draft. The bulk of its military equipment was remnants from WWI, including armored tanks much inferior to the German Panther and Tiger versions. The German Messerschmitt fighter plan was faster and more maneuverable than the British Hurricane. The island country seemed to have a single advantage: the Royal Navy then considered the best in the world and ruler of the seas. But, as the combatants discovered, sea power can be overcome by airpower, so that controlling the oceans is, at best, a temporary advantage.

England’s true advantage in WWII was its overlooked ability to capture and analyze enemy information. For example, during WWI, the British persuaded America to enter the war by exposing a secret message from German foreign minister Arthur Zimmerman encouraging his ambassador to forge a secret wartime alliance with Mexico. In the years following, the Germans developed the “Enigma,” a machine that coded messages with a series of three to five rotating wheels so that another device could only read the contents of a missal with the proper settings.

Unknown to the Germans, a secret group of decoders working in Bletchley Park broke the code early in the war, enabling British supply ships to evade U-boats critical to the island’s population and resupply of war materials. According to one General, “The knowledge not only of the enemy’s precise strength and disposition but also how, when and where he intends to carry out his operations brought a new dimension to the prosecution of the war.”

Breaking the Project Data Code

Of course, the PMO has it much easier than those who worked at Bletchley Park. Computers are far more powerful, and modern business thrives on collaboration over secrecy. The strategic advantage that project data offers is also precious. Companies that can move fast and deliver change projects rapidly tend to outlive their less nimble competitors. But in order to achieve that strategic advantage, organizations need to understand themselves at least as well as they know their competitors.

Know the enemy and know yourself, and you can fight a hundred battles with no danger of defeat. If ignorant of both your enemy and yourself, you are sure to be defeated in every battle.

Sun Tzu – The Art of War

The value of information in the development of business strategy is not a new concept. Harvard Business Professor Michael E. Porter, the author of Competitive Strategy, wrote in 1985 that the new information revolution would:

• Change industry structure and alters the rules of competition.
• Create a competitive advantage by giving companies new ways to outperform their rivals.
• Spawn whole new businesses, often from within a company’s existing operations.

Though Porter focused more on the advantages that information technology would bring to his value chain concept, he noted that companies could gain extra profits by selling customer information to retailers, market research companies, and food processors.

Why 35 years on from Porter’s predictions, are we still struggling to derive competitive advantage from Project Data?

Information versus Project Data

While many use the terms “data” and “information” interchangeably, they are decidedly different. Data is a collection of unorganized facts expressed in raw forms of numbers, statements, and characters. Examples include 941581965, R[.!o(@7pH=3, or X+Y=Z. For most purposes, the facts do not mean much with context. Information is knowledge gained from processing, interpreting, and organizing facts, i.e., data.

The nature of the relationship between data, information, and knowledge continues in flux among information scientists, especially information and knowledge. Many assert that knowledge is the result of processing data into information, then into knowledge. Others insist that knowledge is necessary to convert data to information. The theoretical arguments are interesting but do little to resolve the problem expressed by Peter Diamandis, Founder of the X-Prize: “Every second of every day, our senses bring in way [more] data than we can possibly process in our brains.”

Today, organizations collect more data than at any time in human history. There are great reservoirs of facts concerning every aspect of modern life, relationships, movements, and emotions in print and digital form, and the totals are growing exponentially. Some futurists suggest that human knowledge doubles today every 13 years and will drop to 12 hours when the Internet of Things (IoT) matures. The World Economic Forum estimates that the amount of data created daily will be 463 exabytes (1000 bytes to the sixth power) in five years. For comparison, scientists estimate that the total of words spoken by humans since speech appeared at least 50,000 years ago into 5 exabytes. The ability to select, process, and organize data into information and subsequently apply the knowledge gained to company operations is a substantial competitive advantage.

The challenges of working with Project Data

Project Data initiatives are not for the faint-hearted! Here are some of the common challenges that PMO teams are faced with they embark on project data mining projects.

Are PPM Tools up to the job?

When organizations decide to invest in Project Portfolio Tools (PPM), they often have business cases that talk about conducting project data analytics and generating insight that informs strategic decision-making. Many of these implementations fail to live up to the hype. Such tools are great at capturing project information, but such data points have limited value on their own. Consider the difference between having a list of planned dates vs high-confidence predictions on which projects will be late, and what remedial actions could improve the outcome. Too often, we can click icons to generate the former, but not the latter.

But we can’t blame the tools entirely. Many PPM implementations invest in an ‘MVP’ and leave behind teams who do not have sufficient skills or investment to continue to evolve the tooling to achieve the desired outcomes. Also, to effectively turn information into project data, and project data into actionable insight, we need to see the whole picture… which is rarely the case. While many PPM tools offer multiple capabilities, the reality is that project data are usually split across different products and repositories. Project Finance may be tracked in SAP, while Project Plans reside in Monday, Developers manage their backlogs in Jira, and reports are often hacked together with Excel, Email, and SharePoint. Successful PPM implementations would require cross-system data-mining and process analytics, not to mention insight into happiness quotients and other team data.

Are Projects Unique?

Another challenge we face with gaining a competitive advantage from Project Data is that project managers love to insist that their projects are ‘unique‘. Whilst some PMs can argue passionately about how their project is a special case, skilled PMO and Portfolio Leaders quickly identify patterns in project delivery and draw distinctions between the ‘Runner‘ projects which can be heavily templated, and the genuine ‘strangers’ with high levels of Novelty, Complexity, Pace, and new-tech.

We need to be honest about the dichotomy we create when delivering projects within organizations. When we spin up projects, we create mini-startups that are designed to deliver on a clear goal, at pace, and usually in spite of the silos and accepted ways of working within the organizations. We value project managers who can cut through the noise and get things done despite obstructive SLAs and ‘the way we do things around here‘. Small wonder then, that when it comes to project data, we often find it to be lacking. The data generated by projects may be valuable to the Portfolio and the organization, but it is often not valuable to the project team. Taking time to produce accurate, clean data are rarely a priority – getting deliverables out of the door with the right mix of time, cost and quality is the only game in town.

Is there enough data?

When PMO people talk about Project Data analytics and its potential, it doesn’t take long for the conversation to move into Machine Learning and AI. But whilst, ML and AI have great potential to offer the world, PMO teams working in isolation soon discover that they simply do not have access to the volumes of quality data that are required.

In order for a Machine Learning algorithm to perform predictions, some training data are required. How much data depends on how accurate your results need to be. The amount of data will depend on the number of predictors you want to use, how accurate you want your model to be, and what margin of error you require. Things get more complicated still when you consider how much change your organization goes through, and how project frameworks vary. Usually, there is insufficient training data of sufficient quality for PMO teams to train algorithms based on their own data sets.

Capturing the Strategic Advantage of Project Data

In order to capture the strategic advantage of Project Data, Project Management Offices and Portfolio Managers need to consider the following:

• Define a data and analytics strategy. PMO teams need to be clear on what they want to achieve from any data and analytics work they undertake. Get agreement from stakeholders on what you are aiming to achieve, and what data you will need to capture to achieve your goals.
• Get Project Manager buy-in. Any project data initiative starts with having some meaningful data. Moving the PM mindset from merely focusing on the project deliverables to building high quality data for the business requires a shift in outlook. Systems such as Kotter’s eight step change model can be used to create a sense of urgency, and enlist volunteers who are excited about the initiative and are empowered to deliver outcomes.
• Develop the right technical skills for your PMO. Managing data is a complex process requiring a variety of diverse skills. PMO Analysts need:
  • Business Process skills to understand and plan for the creation of reliable data,
  • Design skills to prepare for systems where data will be used of stored,
  • Technical skills to be able to adminster the software systems where data is maintained,
  • Data Analysis skills to understand issues and problems discovered in project data,
  • Analytic skills to interpret data and apply it to new situations,
  • Vizualization and storytelling skills to help people understand the data,
  • Strategic thinking skills to see opportunities to use data to serve the business.
• Timescales. Building datasets from scratch takes time. If you are starting from nothing, then you will need several projects to complete a full lifecycle (possibly including benefits realization, depending on your goals). Sponsors need to understand that such initiatives are not ‘quick wins’, more ‘slow-burns’. Building a quality data set takes time and is a mid-term investment.

Hot areas for PMO Data innovation

So you want to begin to capture the strategic advantage of Project Data – where do you start? Here are some areas where we see PMOs beginning to use project data to achieve business benefits.

Mastering Project Risk Data

A sizable chunk of project management is about managing risk. Projects have varying strategies for capturing risks and implementing risk mitigation strategies. But often this data is forgotten about when the project ends. Augmenting risk logs with other data sources, allows teams to build data sets that will both model project risk more accurately, and even reduce risk on future projects. Here are a few examples of where Data Projects with Project Risk data could be beneficial:

• Identifying common risk profiles for different types of projects. Such profiles can be used to ‘seed’ risk logs for new projects, giving project managers a head start on avoidance and reduction strategies to increase the chances of project success.
• Modeling Project Risk Budgets based on historical data. By understanding the cost of risk on historic projects, Programme managers can have a better understanding of the appropriate risk budget to assign to new projects
• Auto-populating risks by task. Imagine a project manager entering a task in their plan. They click a button and the PPM system suggests risks that should be considered and added to the log. The list comes not from a generic list of risks, but from historic data from real-world projects that were delivered in the same organization or operating environment. Having such data at their fingertips would be invaluable to project planners and allow them to factor in contingency plans, start activities earlier, or chose strategies that minimise deliver risk.

Mastering Dependency Data

Mapping dependencies is time-consuming. But data projects can help. By building datasets and making data available to project and portfolio planners, we can move away from referring to the art of dependency management and start thinking of it more as a science.

• Modeling can help resource managers and portfolio managers model different scenarios to establish which delivery prioritisation approach results in the fewest cross-project dependencies
• Historic data can identify dependencies before project managers are even aware of them. A project manager adding a task to create a new server, can receive suggestions based on historic project plans about the dependencies that this requires on (for example) network teams, security teams and procurement.
• One of the major challenges with dependencies is vizualisation. By building dynamic data sets and creating visual representations of dependency maps across portfolios, they become visible. And when dependencies are visible, it is easy for people around the project, such as sponsors, stakeholders and operational teams, to better support project timelines and provide better project support. Too often project dependencies come as a surprise to operational teams. Modeling and visualisation can help.

Mastering Estimation

One of the most reliable ways of estimating projects is to use historical data. Using data from timesheets, and ‘actuals’, we can get a much better sense of how long activities take in real life, when the people who work in our organization undertake them using our organization’s unique processes and procedures.

• Datasets of actual costs and durations for activities can be used to estimate work with increased confidence
• Analysis modeling such as Monte Carlo analysis can be used to provide more reliable indications of when projects will complete.

Which factors count towards project success?

Sophisticated models can be applied to historic data to understand which factors truly drive success in your organization. Does increasing resources make a difference? What about the competency level of the project team? Does having the right sponsor increase the chances of project success? If you have a quality data set with sufficient data points, then these questions can be answered with regression analysis – helping you to infer causal links between interventions and project success. It doesn’t need to be limited to interventions either. These data sets can be used to answer questions such as:

• Which projects are more successful – small or large?
• What is the best time of year to start a project in our organization?

Modeling of this nature allows PMO teams to identify leading indicators of project success/failure that are uniquely in tune with the environment in which the project is being delivered.

Final thoughts

The transition from project support office to project data master is not easy, nor is it inexpensive. It takes investment in skills, systems, and of course, the commitment of the projects’ community. But the potential rewards are great. PMO and portfolio management leaders who chose to invest in this area can use data and analytics as the basis for strategic decisions that lead to competitive dominance. Portfolio Offices are all about delivering the right things and delivering things right. Becoming a data-driven PMO enables you to do both.