Approval Threshold Matrix¶
Template — instantiates Decision Load Management
A threshold table linking decision stakes to required review depth or approval level.
An Approval Threshold Matrix is a standing lookup table that answers one question before any deliberation begins: for a decision this consequential, who must sign off and how much review is required? Its rows are stakes bands — usually keyed to a monetary amount, a reversibility flag, or a risk tier — and its columns state the review depth and the named authority that band demands. The defining move is that the matrix is fixed in advance and consulted per decision: it is not a routing workflow that reacts to a queue, and it is not a narrative document of who-may-decide-what. It is a static grid whose only job is to convert a decision's magnitude into a required approval level, so that trivial choices clear at the front line while expensive or irreversible ones climb to the authority that should own them.
Example¶
A manufacturing company drowns in signature requests: a plant supervisor waits three days for a VP to approve a $400 replacement gasket while a $2M line retooling gets waved through in the same email thread as everything else. The finance team builds an Approval Threshold Matrix. Purchases under $1,000 are auto-approved against a standing budget and need only the requester's manager to be notified. From $1,000 to $25,000, the department head signs. From $25,000 to $250,000, the divisional VP plus a one-page business case. Above $250,000, or anything flagged irreversible (long-term contracts, capital assets), the capital committee reviews it at its monthly meeting with a full appraisal.
The matrix is printed on one page and wired into the purchasing tool as a set of routing rules. The immediate effect: the gasket clears in an hour, and the VP's attention is now spent only on the handful of decisions large enough to reach her row. Note that the numbers here are illustrative — the point is the shape, a small number of bands each mapped to a proportionate authority and review depth.
How it works¶
The matrix is built, not improvised. First, agree the stakes axis — the one or two dimensions (amount, reversibility, risk class) that best predict how much scrutiny a decision deserves. Second, cut that axis into a small number of bands; too many bands recreate the load the matrix is meant to remove. Third, for each band name the required authority and the review depth (notification only, single sign-off, documented case, committee). Fourth, define what happens below the lowest threshold: a standing rule that lets routine items clear without a human gate at all. The finished artifact is a reference, consulted the way one reads a tax bracket — you locate the decision's band and read off its obligation. This is the archetype's "four-eyes" logic[n1] rendered as a proportionate schedule rather than a blanket rule.
Tuning parameters¶
- Band count and granularity — how many stakes tiers the matrix carries. More bands fit authority more precisely to consequence but add lookup friction and boundary disputes; fewer bands are faster but blunter.
- Threshold values — where the cut-lines sit. Set them low and too much rises to senior review (defeating the purpose); set them high and consequential decisions slip through under-scrutinized.
- Stakes axis choice — whether bands key on dollars, reversibility, risk class, or a composite. A dollars-only axis is easy but blind to cheap-yet-irreversible decisions.
- Review-depth ladder — how steeply obligation rises across bands, from a light notification to a full documented appraisal.
- Refresh cadence — how often thresholds are re-indexed to inflation, budget growth, or org change, so a once-sensible $25,000 line does not quietly become trivial.
When it helps, and when it misleads¶
The matrix's strength is proportionality: it reserves scarce senior attention for the decisions that actually warrant it and lets everything below the line move at speed, which is exactly the archetype's core allocation goal. It is also legible — anyone can see why a given decision needs a given signature.
Its classic failure is threshold gaming: when a hard cut-line at $25,000 exists, requests cluster at $24,900, decisions get split into sub-threshold slices to dodge the higher gate, and the matrix's discipline erodes. A dollars-only axis compounds this by treating a reversible $30,000 spend as weightier than an irreversible $5,000 policy change. And a matrix left un-refreshed drifts: yesterday's meaningful threshold becomes today's rubber stamp. The guarding discipline is to audit for clustering just under cut-lines, add a reversibility flag alongside the amount axis, and treat the thresholds as living values that are re-indexed on a schedule rather than set once and forgotten.
How it implements the components¶
decision_stakes_classification— the matrix is a stakes classifier frozen into a table: each row is a consequence band with an explicit boundary.choice_owner_map— every band names the authority that owns decisions of that magnitude, mapping stakes to accountable signer.default_or_delegation_rule— the sub-threshold rule delegates routine, below-the-line decisions to a standing policy so they clear without a human gate.
It does not detect degraded judgment or trigger a pause — that fatigue_signal work belongs to Fatigue Check Protocol — and it does not group like decisions into timed windows (batching_rule); that is Decision Batching Workflow.
Related¶
- Instantiates: Decision Load Management — the matrix supplies the stakes-to-authority mapping the archetype needs to reserve scarce judgment.
- Sibling mechanisms: Delegation Rulebook · Decision Triage Queue · Default Option Design · Low-Stakes Decision Automation
Editorial Notes¶
Form Classification¶
Form family: Rule, Policy & Commitment
Rationale: The matrix sets standing stakes bands and binds each to a required review depth and approval level, so its operative form is a conditional governance rule.
Nearest alternative: Representation, Specification & Plan — A threshold table carries the rule, but the mechanism constrains which authority and scrutiny a decision must receive.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Organizational & Management Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Management science maps decision stakes, reversibility, and risk to delegated authority and review depth.
Related originating lineages:
- Accounting & Auditing — Financial controls supply monetary bands and four-eyes rules.
- Operations Research — Decision-load allocation explains reserving senior attention for consequential cases.
Review resolution: Management decision rights and delegation make organizational management primary. Accounting materiality and authorization bands, together with operations-research allocation of scarce review attention, materially shape the threshold grid; legal authority is a constraint on some deployments rather than a separate origin.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The matrix answers how much review and whose signature by magnitude; a Delegation Rulebook answers which role may decide what by domain. They pair naturally — the matrix sets the height of the gate, the rulebook says who holds the key — but conflating them produces a document that is neither a clean lookup nor a clear grant of authority.
[n1] The four-eyes principle — the control convention that a consequential action requires approval by at least two people. A threshold matrix generalizes it into a graded schedule: the number of eyes rises with the stakes rather than applying uniformly. ↩