MARKHAM
Research
Brief · MKM-R-2026-011

The cost of the slow decision

Slow decisions are not a culture problem. They are an unpriced cost line.

Quantifying decision latency across 40 organisations, with a method for pricing your own.

Length12 pages
Samplen = 40
Period2024–2026
AuthorsMarkham Institute
ReferenceMKM-R-2026-011
Version1.0 · Current
Download the excerpt (PDF)Request the full 12-page edition
The summaryA nine-minute read

Slow decisions are not a culture problem. They are an unpriced cost line.

Across 40 organisations we timed the same thing: the interval between a decision becoming necessary and the decision being made and communicated. The median organisation loses nine working weeks a year to decisions waiting in queues — not being analysed, not being debated, simply waiting for a forum with the authority to decide. At mid-market scale that latency prices out between $2M and $14M a year in delayed value, stalled programmes and repeated work.

This brief publishes the measurement method in full. Latency is observable: it can be read out of calendars, approval trails and steering-committee minutes without a single interview. The brief walks through the four highest-cost decision classes, shows the latency distribution across the sample, and prices a worked example so a reader can run the same arithmetic on their own organisation in under a week.

Key findings
9 wks

Median working time per year that necessary decisions spend waiting in queues, across the 40-organisation sample.

34 d

Median latency for cross-functional decisions — against 4 days for decisions held by a single named owner.

$6.8M

Median annual carrying cost of decision latency in the sample, priced against delayed value and rework.

71%

Share of measured latency attributable to missing decision rights, not missing information.

Inside the report4 chapters · 12 pages
01
Latency is measurableThe definition, the instrument, and why calendars tell the truth that interviews do not.
3 pages · 4 min
02
Where the weeks goThe four decision classes that carry most of the cost, and the distribution across 40 organisations.
4 pages · 5 min
03
Pricing your own latencyA worked example, from approval trail to annual cost, reproducible in under a week.
3 pages · 4 min
04
What actually reduces itNamed owners and a fixed cadence outperform every delegation memo in the sample.
2 pages · 3 min
If you only act on four things

The findings, as Monday-morning decisions.

a

Time your last ten significant decisions from necessity to communication. That number is your baseline, and it is almost certainly larger than you think.

b

Separate latency caused by missing information from latency caused by missing authority. In the sample, authority is the larger problem by a factor of three.

c

Give every recurring decision class a single named owner. Shared ownership is measured here as unowned.

d

Put the remaining decisions on a fixed cadence, so they wait days for a scheduled forum rather than weeks for an improvised one.

Methodology & governance
Sample40 organisations, $50M–$500M revenue, drawn from the wider transformation cohort. Each contributed a minimum of 60 timed decisions.
MeasurementLatency is read from calendars, approval trails and minutes — the interval from decision-necessary to decision-communicated. No self-reporting.
PricingCarrying cost is priced against the value case of the delayed work item, discounted for attribution. The arithmetic is published in full in chapter 3.
InstrumentThe measurement instrument is published as the Decision Velocity Index (MKM-F-007) and may be applied without licence.
Citation

Markham Institute, The cost of the slow decision, MKM-R-2026-011, v1.0 (May 2026). Citation permitted with attribution.

Revision history
v1.0 · May 2026First publication. Sample of 40 organisations, 2024–2026.