HR as DJ

The core insight was that capacity is not just a number. It is a distribution problem. A firm might have 100 billable people, but if 80 of them are at the senior partner level and only 10 are at the senior consultant level, the portfolio is structurally constrained regardless of total headcount.

I designed a sunburst visualization that reveals how uncertainty permeates the capacity distribution, with opacity encoding confidence levels: practice on the outer ring, skill band in the middle, current utilization in the center. The allocation model estimates the likelihood that a practice area can deliver 5 senior consultants for a 6-month banking engagement, surfacing the confidence levels behind each match rather than pretending to know the answer.

The Chrome app extends this into a search interface. Search for a skill, a practice area, or a specific name, and immediately see availability ranges, bench depth with confidence intervals, and deployment history.

A resource manager accustomed to the precision of a staffing spreadsheet must now navigate probabilistic uncertainty, where 88 consultants represent a distribution of confidence levels, not a fixed count. The interface design challenge was specific and daunting: how do you present capacity uncertainty in a way that supports deterministic staffing decisions without pretending the uncertainty does not exist?

You do it not by hiding the distribution, but by making uncertainty visible as uncertainty. A manager who can see the full capacity landscape with its confidence intervals and understand the mechanism behind each match recommendation is far more equipped to make a defensible staffing decision than someone who sees a single “best fit” score and has no idea where it came from.

This model ultimately became the precursor to a service blueprint mapping the user’s transition from bench awareness to capacity planning and finally to staffing action, with explicit decision points where confidence levels determine whether to proceed or seek more information.

The platform’s information architecture was structured as a spatial reasoning interface, not a static dashboard. The core architectural problem was that each consultant has 3 distinct analysis axes, practice area, skill band, and availability percentage, and the firm’s capacity distribution changes daily, with varying degrees of certainty.

The defining design decision to manage this cognitive load was progressive disclosure. High-level capacity scores with confidence intervals for initial portfolio screening. Highly detailed individual profiles with availability ranges for specific staffing decisions. A unified treemap that encodes uncertainty visually, block opacity reflects confidence in availability, so users can toggle between structural clarity and probabilistic honesty without losing their analytical context.

A forward-looking capacity slider replaced the traditional series of static reports in favor of a continuous view showing how bench availability migrates from today out to the next quarter, with confidence bands widening as the timeline extends. The goal at every single level was to be decision-ready, not data-dumped, and to be honest about uncertainty, not decorative about it.

The entire approach crystallized in the treemap analysis view. This interface relies on a custom squarify algorithm showing precisely where capacity is concentrated, how deep the bench runs, and which practice areas may be able to deliver on demand, with the critical design decision that a senior consultant at 60% availability appears as a translucent block, not as a solid block.

The design decision that truly matters here is that this view shows a resource manager evaluating a banking engagement whether the firm may have 5 senior consultants available or merely 5 junior associates, and shows the confidence level behind that estimate. That singular capacity distinction changes the staffing decision entirely. It changes the proposal. It gives a manager something tangible they can put in front of a delivery board and defend, not an abstract utilization score, but a physical mechanism of capacity with its uncertainty made visible rather than hidden.