EkaLore believes that BAD (Big Data/AI/Data Science) Projects face challenges common to many high visibility projects but with some unique twists due to the expense of the human and tech resources required. Inflation, war, the aftermath of the pandemic, and consumer demand shift cause what EkaLore terms “Challenging Times”, an environment where previously untouchable projects and concepts are being questioned at a far higher rate than current managers have ever experienced.
This post outlines key areas to improve the viability of BAD Projects in the face of increased and changing management requirements for the value returned.
Challenging Times are also changeable times. What was a great use of time when internal ROI requirements were 5%? It might be a waste of time if those requirements increase to 15%. To guard against this, enterprises need to re-affirm and analyze BAD Projects with assessment and evaluation even as projects continue. The scale and scope of BAD Projects provide opportunities to refine and re-focus work efforts even as projects progress. BAD Projects need this continuing analysis to ensure resulting benefits evolve and meet needs even as enterprises react with agility to challenging times.
Completing BAD Projects successfully often requires management attention. Challenging Times to increase the competition for that attention. Understanding how to communicate to upper management and (probably even more importantly) what to communicate is a required skill to keep the resources and time commitments to bring the project home.
Without the management attention mentioned above, many large BAD Projects have reached completion only to find their results to be irrelevant or discarded in the face of new realities. Examples abound in government (defense, social services, others), commercial enterprises (logistics, energy, marketing), and non-commercial (technology deployment, records). Once running, the sheer organizational inertia and personalities (politics) of a BAD Project fuse to reduce the interest in reviewing alignment or escalating commitment. Or in other words - BAD Projects become terrible projects one day at a time.
Other elements key to Bad Project success are priority access to resources, resource commitments in the face of escalating (growing) needs, and the time needed to start getting benefits. Scheduling and activity focused project planning lead to the well-known “90% done” effect, where substantial amounts of elapsed time, resources, and attention are disproportionately spent on “the last 5-10%” of a BAD Project. A simple test of a BAD Project risk is to evaluate the level of benefit if a project were canceled at 10%, 25%, 60%, or 90% of ‘planned activity.’ When ‘benefits’ or ‘rewards’ are mismatched with effort levels, the risk (and unrecovered resource spend) climbs. Unmatched efforts and benefits are frequent in BAD Projects. When the unmatched level changes from initial assumptions, a clear warning for management attention is present. A successful BAD Project wins the support of decision-makers by building their confidence that status and activity reports relate to risks and rewards.
BAD Projects take substantial time to accomplish results and gain benefits at scale and scope across enterprises. That time is measured in quarters, not weeks. Successful efforts require management attention and priorities combined with the time needed for a successful effort. Reaping substantial benefits requires sustained attention to the alignment of efforts, results, and assumptions.
BAD Projects require getting critical resources on a timely basis. Failing to provide essential resources on a timely basis is also a warning sign that management priority access (even if set at a ‘low priority’) is likely to raise the risks of bad outcomes. Frequently the competition for resources diverts resource priorities from all except the most exciting efforts. Implementing a new machine learning algorithm is more exciting than validating data sources. Even though a new ML model may have lower benefits than having more valid data.
One way to mitigate the risks to BAD Projects from failures of priority access to resources can be to clearly communicate up front critical resource needs/timing and revisiting as part of normal status reporting. The proverb “… for want of a nail…” began in military settings and applies to BAD Projects. A successful BAD Project wins access to critical resources by showing the costs (and required ripple effects) deserve priority by relating timely resource to benefits.
BAD Projects are especially susceptible to time changes. The largest factors in driving costs and risks of failures for large projects are schedule compression (too big a scope or scale in the time allotted) and schedule overrun (taking too long to consistently perform tasks required). BAD Project plans need to build in pre-analyzed “adjustments” and “refinement milestones” where changes in tasks, activities, and spend are adjusted to time scales and resources actually accomplishing results. A successful BAD Project already plans for adjustments to assure completion and rewards.
Successfully completing a BAD Project requires the same approaches as other enterprise processes: agility (such as preplanned adjustments when schedule slip), resilience (seeing alternatives; resource paths), and sustainability (maintaining good management attention).
You can read more about how to handle BAD (Big Data/AI/Data Science) projects at www.ekalore.com/bad-project-blog
Commentaires