Register now for KMWorld in Washington, D.C., November 15-18. Save an extra $200 with Super Early Bird savings!

Deep project management

Article Featured Image

Deep project management (anticipatory)

This consists in large part of integrating various “deep” and other related technologies. Two such technologies were on full display at the KMWorld 2019 conference this past November: text analytics and knowledge graphs. Other key aspects are complexity management and embedded semantics. Let’s take a look at each of these and how they apply.

Text analytics and knowledge graphs. Remember those old entity-relationship diagrams people used to manually piece together when constructing a relational database? Text analytics does that and more automatically by using a process called entity extraction. To learn more, Tom Reamy’s book Deep Text, published by Information Today, Inc., is an excellent reference.

Many text analytics platforms take the extracted entities, categorize them, identify their interrelationships, and visually present it all in a concept map called a knowledge graph. These graphs show relationships among the many moving parts of a project—tasks, schedules, resources, etc.—and identify interdependencies that fan out across the entire organization and beyond. The best of these tools also give you access to the underlying ontology behind the map.

Application to deep project management: Think of the mountains of unstructured and semi-structured documents, including policy guidelines, standards, emails, presentation slides, SMS chats, and written reports. All contain hidden nuggets, clues to what’s actually going on deep within a project. These faint, early warning signals usually get filtered out in traditional and even agile project management. Amplifying these signals and connecting the dots across them creates an early warning system, greatly reducing risk.

Complexity management. In previous articles, we’ve mentioned cyclomatic complexity, which analyzes the many interwoven paths through a work process, especially where decisions are involved. The greater the number of paths, the greater the chances for error, inconsistency, and unnecessary redundancy, all of which contribute to decreased efficiency and effectiveness and increased cost. This applies to the end product being developed as well as the project organization, including the people and workflow. Managing any one alone can be difficult, but managing all of these aspects together can be truly daunting.

Application to deep project management: The simple act of identifying and removing redundant or unnecessary paths results in fewer mistakes and reduced risk.

Embedded semantics. Embedded semantics is a means for documenting critical knowledge associated with each step in a process, along with the rationale for any actions taken (see “The curious case of a broken crumb trail” in the March 2013 issue of KMWorld). Semantic authoring tools are currently under development that allow embedding knowledge trails directly within project documentation. This removes much of the “guesswork” left behind by even the best text analytics platforms.

Application to deep project management: As errors in a project occur, or when parameters such as cost and schedule begin to drift out of range, the typical response is to conduct a forensic type of analysis in an attempt to find out “what went wrong.” Embedded semantics helps transition a project management system from a data-rich, context-poor environment to one that is both data- and context-rich, resulting in reduced ambiguity and fewer errors due to misunderstanding and misapplication of information.

An illustrative example

Imagine that during usability testing of a new software system, a set of additional requirements emerges. In traditional project management, this would kick off a lengthy process, starting with convening a change review board, followed by designing, estimating, approving, and implementing the changes. In agile project management, the first few changes happen smoothly. But as new requirements keep piling on, the project backlog grows along with the list of “bugs.” Either way, the project gets delayed. Eventually, things seem to be getting out of control and an “expert” is called in to fix the problem. Hint: if you need to call in an expert, that’s a clue that you were attempting to do something without having the requisite knowledge in the first place. That’s where deep project management comes in.

It’s easy to see the many disconnects (knowledge gaps) in this example. One type is the latent errors and missing information that occur as requirements are being developed, because the end users are not fully engaged in the requirements definition phase. Another is the preponderance of skills mismatches and semantic disconnects across the many different perspectives involved, including business and finance, enterprise architecture, quality assurance, safety, security, and legal, etc.

In deep project management, many of these disconnects simply wouldn’t happen. All key stakeholders would be involved from beginning to end and would be actively engaged in key decisions in every phase of the project. There would also be a clear understanding of the specific knowledge and skills required at each step, and assurance that people with that knowledge are present throughout. Needless to say, the opportunities for applying knowledge management to this latest generation of project management are many.

What it all means

Given the increased negative media exposure that comes from project failure, organizations need more tightly integrated, intelligent project management systems, in addition to people who have the requisite skills. This need will grow as systems continue to become more complex and timelines more tightly compressed.

Quantitative benefits include cost savings through error reduction, improved resource allocation, and better human capital management. Identifying and automating repetitive project management tasks alone can result in cost savings of 20% or more. Perhaps best of all, project management systems aided by machine intelligence could help identify potential “blind spots” and avoid project stakeholders having to say those dreaded words: “How did we miss that?”

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues