River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Organizational Design Assessment Template

Five documents and three sheets that score every seat's span against its archetype, then classify each recurring decision Clear, Diffused, Escalated or No Decider.

Free download  ·  No account needed

Decision Bottleneck Map · Bellhaven Specialty Insurance · 260 people, four offices · logged one quarter

Sorted on days to close, not on the org chart

DecisionNominal deciderReal agreersDaysClassification
Cross-office referral splitNo seat named0 named19No Decider. Closes at whichever executive is free
Vendor change over $10kOperations Director1 (the CFO)11Escalated. The CFO closes it, not the named seat
Underwriting exception >$50kSenior Underwriter2 named seats6Diffused. Both seats sign nearly every instance
New hire requisitionBranch Manager0-2, by office6Inconsistent. Three offices, three sign-off counts
Client fee discount >10%Account Manager1, rarely asked2Clear. The nominal seat is the real seat
PTO or scheduling conflictTeam Lead01Clear. Highest volume, least contested

Ten leadership seats, five reporting layers, every span defensible. Four of the six recurring decisions above still have an authority problem no org chart shows.

Search this and most results hand you the same instrument: a span calculator testing direct reports against a target ratio, or a maturity score across five levels like the maturity model pack already runs. Both grade the org chart. Neither asks whether the seat the chart names as a decider is the seat that actually decides. A company can clear every span and never have redrawn a box in years. It can still run an underwriting exception through two people's sign-off, or leave a referral dispute with no seat named for it.

This pack starts from the same span-and-layer check everyone runs, scored per seat against OMB's own guidance on spans and delegated authority, then adds what that check leaves out. Every recurring decision gets logged onto a Decision Bottleneck Map: instances, how many escalated above the nominal decider, and how many named seats had to agree at once. Three numbers, one of four classifications every time, Clear, Diffused, Escalated or No Decider, never assigned by feel. A Role Overlap Register then catches the human half: two seats whose occupants each describe one decision as theirs.

Bellhaven Specialty Insurance, 260 people across four offices grown by acquisition, scores a defensible span at ten leadership seats and five reporting layers. Its chart is fine. Four of its six logged decision types are not. One splits across two seats that both sign every instance, one has quietly moved to an unnamed seat, and one has no decider at all. A fourth closes differently, depending on the office's legacy company. Score the client on a rubric with the diagnostic assessment pack, or check what its policies cover with the risk and controls assessment.

Ten seats, six decisions, and four authority problems the chart cannot show

The Span and Layer Analysis, the Decision Bottleneck Map, and the Role Overlap Register.

Span and Layer Analysis

Every seat scored against the kind of work it supervises, not one flat ratio, built from the org chart and the headcount.

SeatLayerReportsArchetypeBand
CEO14Manager of function headsYes
COO24Manager of managersYes
CFO22Specialist function headYes
Regional VP Underwriting22Specialist function headYes
HR Director23Specialist function headYes
Operations Director35Supervisor of mixed leadsYes
Underwriting Director36Supervisor of professional staffYes
Branch Manager, Meridian38Coordinator (largest site)Yes
Branch Manager, Ashford35Coordinator of managersYes
Branch Manager, Colby & Trent Falls33Coordinator (smaller site)Yes

All ten leadership seats clear a defensible band. A fourth layer of team and department leads sits below these ten at a median of five reports, range three to nine, also within band.

Decision Bottleneck Map

Six decision types, one quarter of logged instances, classified from three numbers on each row alone.

RefDecisionLoggedEscalatedReal agreersDaysClass
D-01Underwriting exception >$50k3485.3%26Diffused
D-02Client fee discount >10%5824.1%12Clear
D-03New hire requisition2972.4%0-26Inconsistent
D-04Vendor change >$10k1788.2%111Escalated
D-05Cross-office referral split19100.0%0 named19No Decider
D-06PTO or scheduling conflict1452.8%01Clear

D-04 is the trap a flat escalation-rate read misses. Only one real agreer beyond the nominal seat, which usually reads fast, and still an eleven-day median because every instance visits the wrong desk first.

Role Overlap Register

Two seats, each independently describing one decision as theirs, in their own words from an interview.

RefDecision in disputeSeat ASeat BStatus
RO-01Underwriting exception sign-offUnderwriting DirectorRegional VP UnderwritingOpen, linked to D-01
RO-02New-hire requisition approvalBranch ManagersHR DirectorUneven, linked to D-03
RO-03Claims-adjacent vendor selectionOperations DirectorUnderwriting DirectorLatent, no Map row yet
RO-04Off-cycle retention counter-offerHR DirectorBranch Manager, MeridianOpen, low frequency

RO-01 explains D-01's six-day median. Neither seat's job description names the other as a required check, so both sign nearly every instance rather than one.

What's in the pack

01

Span and Layer Analysis

Ten seats scored against the kind of work each one supervises, not one flat ratio, with the archetype and reporting layer named beside every span.

02

Decision Bottleneck Map

Every recurring decision logged with instances, escalation rate and real agreers, classified Clear, Diffused, Escalated or No Decider from three numbers alone.

03

Role Overlap Register

Two seats whose occupants each describe one decision as theirs, in their own words from an interview, with neither account judged the correct one.

04

How the Classification Works

The three numbers behind every classification and the working thresholds that set them, so a Diffused row and a No Decider row are never confused.

05

Current and Proposed Structure

The org chart as it stands today, and what actually changes once every fix is priced as a decision right rather than a reporting line.

06

Design Findings

The headline read off the Map and the Register together, naming which recurring decisions have an authority problem and which class each one belongs to.

07

Options Analysis

Two named options per broken decision, each effort-scored and sized against a comparable row already on the Map rather than against a guess.

08

Transition Note

The rollout sequenced by effort against what it removes, plus the exact sentence that announces a changed authority line without naming who held it before.

How to use it

  1. 1

    Open in River, or take it blank

    Open the pack in River and send the org chart, headcount and whatever already shows a decision closing, or download the Word documents and CSV sheets and work through them yourself.

  2. 2

    Log decisions before you diagnose

    Build the Decision Bottleneck Map first: how many instances of each recurring decision happened, how many escalated, and how many named seats had to agree.

  3. 3

    Check spans against the archetype

    Score every seat's span against the kind of work it actually supervises rather than one flat ratio, and read the reporting layers off the same chart.

  4. 4

    Write findings that name the decision

    Turn the Map and the Register into findings that name the seat and the decision, then size and sequence the fix the way the Transition Note requires.

Frequently asked questions

Is this template free?

Yes. The zip is Word documents and CSV sheets, no account and no card. Edit with AI is the optional half: the agent builds the Map and the Register from your own files and classifies every decision. The rest sit in the template library.

What format are the downloaded files?

Five Word documents and three CSV sheets in one zip, no conversion needed. The sheets carry the columns that do the work: instances, escalation rate, real agreers, and the classification each row computes to. Open them in Excel, Numbers or Google Sheets.

What does Edit with AI actually do?

It reads the org chart, the headcount and whatever evidence of a real decision closing exists, an approval trail, a ticket history or interview notes, then logs each recurring decision onto the Map and computes its classification.

How is this different from a span of control calculator?

A span calculator stops at one ratio, direct reports against a target number. This pack scores that too, then logs what a chart cannot show: whether the seat named as decider is the seat that actually decides.

Does this replace the org chart with a new one?

Rarely. In the worked example the proposed chart is identical to the current one, the same ten seats and reporting lines. Every fix changes who has to agree on a named decision, recorded in an authority schedule rather than a moved box.

What if two people both think a decision is theirs?

That goes on the Role Overlap Register as a named dispute, in each person's own words from an interview, with neither account treated as more correct. Bain's own framework is what the eventual fix gets checked against.

Can this diagnose a structure problem without it becoming personal?

That is the rule behind every document in the pack. A finding names the decision and the seat, a sign-off requirement, for example, never the person who currently holds it, which keeps the conversation about the structure.

Find the decision the chart can't show

Take the Word documents and CSV sheets blank, or send River the org chart and whatever shows a decision closing and get back a classified Map.

Edit with AI