The 90-second claim tends to read as a speed argument. It is not, primarily. It is an argument about what changes when speed stops being a constraint — and specifically, what different questions a professional services firm starts asking about its allocation process when the time cost of each decision drops from hours or days to under two minutes.
Firms that reduce allocation time to 90 seconds don't just do the same thing faster. They start doing different things. The questions shift from "how do we get to a recommendation" to "is this recommendation the right one, and how would we know."
What the current time cost buys
A 12-day committee allocation cycle is not 12 days of evaluation. It is mostly coordination latency: time waiting for people to respond to emails, time scheduling a meeting, time preparing documentation, time for the meeting itself, and time for the decision memo to circulate afterward. The actual evaluation time — the discussion of which consultant fits this engagement — might be 45 minutes to two hours of a senior resource manager's week, plus the committee meeting time.
The rest of the 12 days is friction cost. That friction cost is not purely inefficiency — some of it is appropriate caution around significant deployment decisions. But the friction cost is applied uniformly regardless of engagement complexity. A straightforward 6-week project extension goes through the same 12-day cycle as a novel 20-week transformation engagement with a first-time client. The first decision does not require 12 days. The second might.
When speed is no longer the bottleneck, quality becomes the question
In a firm running allocation at 90 seconds per engagement, the first thing that changes is the type of conversation that happens around the ranked output. When the scoring pass completes in 90 seconds and the decision-maker is looking at a ranked list with dimension breakdowns, the conversation stops being about discovery — who should we even consider for this engagement — and starts being about validation: does this recommendation make sense, and are there any factors the model didn't have access to that should change it?
That shift in conversation type is not trivial. In a discovery-driven allocation discussion, the committee spends most of its cognitive budget generating and evaluating candidate options. In a validation-driven discussion, the cognitive budget goes toward checking whether the model's reasoning is sound and contributing the contextual knowledge that isn't in the data. The output of a validation conversation is generally higher quality than the output of a discovery conversation operating under time pressure with incomplete information.
The override pattern question
When allocation speed is no longer a constraint, firms start paying attention to override patterns. In a 12-day committee process, overrides happen but they are not easily visible as a pattern: each override is embedded in a specific committee discussion, and the reasoning is often implicit or informal. In a scored system where every override is logged with a reason code, the patterns become visible after 60 to 90 days of operation.
A mid-size firm that went live with scored allocation in early 2024 found, after three months, that 31% of their overrides cited "client preference for a named consultant" as the reason — a category that wasn't available in the original reason code set and had to be added retroactively. Another 22% cited "development opportunity" — a deliberate choice to staff a consultant for growth reasons rather than pure fit. Both categories are legitimate. Neither was visible as a systematic allocation consideration before overrides were tracked. The firm added "development allocation" as a distinct workflow within the scoring system so that deliberate development decisions were logged and tracked separately from pure-fit decisions, giving the resource management team visibility into how development allocations were being used without conflating them with fit-driven overrides.
Calibration as an ongoing practice, not a launch event
One of the operational changes that follows 90-second allocation is that model calibration becomes a routine activity rather than a launch task. When the scoring model is fast enough to use on every engagement, the feedback loop from decisions to outcomes to model adjustment shortens from quarters to weeks. A firm can run a calibration review after 30 scored engagements and adjust weights based on the override patterns and outcome data from that cycle, then run the next 30 under the adjusted weights and observe whether the override rate changes.
This calibration cadence is not possible with a 12-day committee process. By the time 30 engagements have completed and the outcome data is in, months have passed and the allocation decisions themselves are obscured by the complexity of the committee process. Speed is what makes iterative calibration operationally feasible.
The coverage problem: what gets allocated carefully
A persistent issue in professional services allocation is that the formal process gets applied selectively. High-profile engagements, flagship clients, and complex projects go through committee. Smaller engagements, extensions, project addons, and internal work get handled informally — a partner calls a resource manager, a name gets mentioned, an allocation gets made. The informal allocation is undocumented: no override logging, no utilization check, no domain expertise verification.
At 90 seconds per pass, there is no economic case for informal allocation. Running every engagement brief through the scoring process costs nothing in time and adds full documentation and utilization visibility regardless of engagement size. The result is a complete allocation history — not just the decisions that were considered important enough to document — which is the foundation for the calibration feedback loop and for any meaningful analysis of allocation quality at the practice or firm level.