$41M
Outside counsel spend FY27 YTD
0.2% under plan · plan $42M
$9.4M
In-house legal cost · 38 FTE
19% of total legal cost
2,544
Open matters
363 older than 12 months
$479
Blended billed rate per hour
+4.3% YoY · rate +$20 · mix +$17
$2.9M
Guideline recoveries YTD, reconciled
est. $1.4M still leaking
$149K
Billed by unapproved timekeepers this quarter
4 timekeepers · held
The write-down is the wrong trophy
Invoice adjustment rate ranks Pembroke Shaw first; combined defense plus indemnity on the same 186-claim cohort ranks it last.Adjustment rate · ignore for panel choice
Pembroke Shaw
4%
Calder Rowe
5%
Harbour Vale
6%
Thornbury Quill
7%
Combined defense + indemnity vs cohort · the sort
Thornbury Quill
−8%
Harbour Vale
median
Calder Rowe
+4%
Pembroke Shaw
+30%
Cheapest on the bill, most expensive on the claim. Pembroke Shaw tries claims the cohort settles at L160; partner on production; rate explains about a third of the gap. A rate negotiation does not fix this firm. The move: Model moving 40% of Pembroke Shaw's QLD motor volume to Harbour Vale and Thornbury Quill — $1.4M/yr expected saving at equal indemnity ratio; Pembroke Shaw retains the insured relationships it is incumbent on. Open panel review.
YTD spend by matter type
Bar = spend · tick = plan · red = more than 8% over planClaims litigation · motor BI
$19M / $20M
Claims litigation · property & liability
$11M / $11M
Coverage disputes
$4.1M / $3.6M
Recovery & subrogation
$2.3M / $2.5M
Commercial & corporate
$2.9M / $2.4M
Employment
$1.1M / $1.0M
Regulatory · APRA / ASIC
$800K / $900K
Privacy & cyber
$400K / $300K
How this is computed
DefinitionSum of paid invoice lines by the matter's type in the taxonomy (one type per matter, set at intake, reviewed by the classifier); plan from the approved budget by type.Sourcesinvoice_line · matter · budget (ClaimCenter, AP feed)Coverage · n2,544 open + closed matters with FY27 spend · 100% typedFreshnessnightly; AP reconciliation daily
Open matters by type, stage and age
IntakePleadingsDiscoveryTrial prepSettlementClaims litigation · motor BI
1,240
Claims litigation · property & liability
610
Coverage disputes
96
Recovery & subrogation
410
Commercial & corporate
140
Employment
28
Regulatory · APRA / ASIC
11
Privacy & cyber
9
| Type | Open | < 90 d | 90–180 | 180–365 | > 365 d | Median age |
|---|---|---|---|---|---|---|
| Claims litigation · motor BI | 1,240 | 397 | 372 | 322 | 149 | 190 d |
| Claims litigation · property & liability | 610 | 122 | 159 | 195 | 134 | 214 d |
| Coverage disputes | 96 | 19 | 25 | 31 | 21 | 260 d |
| Recovery & subrogation | 410 | 131 | 123 | 107 | 49 | 150 d |
| Commercial & corporate | 140 | 77 | 39 | 20 | 4 | 95 d |
| Employment | 28 | 9 | 8 | 7 | 3 | 170 d |
| Regulatory · APRA / ASIC | 11 | 2 | 3 | 4 | 2 | 240 d |
| Privacy & cyber | 9 | 3 | 3 | 2 | 1 | 120 d |
How this is computed
DefinitionStage from the matter's phase (UTBMS phase of the latest billed work, or the matter-system status where it is maintained); age = today − opened date.Sourcesmatter, invoice_line phase codes (ClaimCenter)Coverage · n2,544 open matters · stage known for 96%Freshnessnightly
Blended rates, variance and staffing mix
Billed rates by level, this year vs last, against the panel market median| Level | Last year | This year | Market median | Hours share | Benchmark share |
|---|---|---|---|---|---|
| Partner | $640 | $668 | $655 | 31% | 24% |
| Counsel | $560 | $585 | $570 | 9% | 10% |
| Sr associate | $430 | $452 | $445 | 26% | 28% |
| Associate | $330 | $341 | $338 | 24% | 28% |
| Paralegal | $190 | $196 | $192 | 10% | 10% |
Variance decomposition. Blended rate $459 → $479: +$20 from rate increases at constant mix and +$17 from a heavier senior mix than the benchmark. The mix effect is the negotiable one.
Associate-level work done by partners. 2,140 partner hours this year on tasks the classifier codes as associate-level (document review, first drafts, research): premium of $700K over the associate rate. Staffing rule LMG-3.x flags it; 61% is already appealed or waived.
How this is computed
DefinitionBlended rate = hours-weighted billed rate per level after reductions; market median from the consented benchmark pool (same level, geography, matter type); hours share vs the benchmark staffing mix for the same matter mix. Task level from the narrative classifier (method and confidence stored per line).Sourcesinvoice_line, timekeeper, benchmark_pool, classifier outputCoverage · nall paid lines YTD · level known for 99% of hoursFreshnessnightly; benchmark pool quarterly
Rate creep · three years
Top-3 firms by spend vs the rest of the panel vs CPI, same-timekeeper chained index| Dimension | Top-3 firms | Rest of panel | CPI |
|---|---|---|---|
| Partner | +11.4% | +6.2% | +7.1% |
| Sr associate | +13.1% | +7.0% | +7.1% |
| Associate | +9.8% | +6.4% | +7.1% |
| Sydney | +12.6% | +6.8% | +7.1% |
| Melbourne | +10.9% | +6.1% | +7.1% |
| Brisbane | +9.7% | +5.9% | +7.1% |
| Auckland | +8.2% | +5.5% | +7.1% |
Chained by timekeeper: the index follows the same people year to year, so promotions and office moves do not masquerade as rate rises. Senior-associate creep at the top firms is the fastest-compounding line.
How this is computed
DefinitionHours-weighted billed-rate index 2024 → 2026, chained within timekeeper; approved rate-card versions used to separate negotiated increases from off-card billing.Sourcesinvoice_line, timekeeper, rate_card_versionCoverage · n88% of hours chainable (timekeepers billing in consecutive years)Freshnessnightly
Actual cost against alternative fee arrangements
Shadow bill (hours × standard rates) vs the agreed fee · tells you whether the AFA is priced right| Arrangement | Fee billed | Shadow bill | Variance | Read |
|---|---|---|---|---|
| Early-resolution motor claims · fixed fee $2,850 Fixed fee per claim · 1,240 units | $3.5M | $3.9M | +$380K (+11%) | Shadow bill above fee |
| Property defense · capped fee $14K Capped fee · 180 units | $2.5M | $2.7M | +$190K (+8%) | Within collar |
| Subrogation · 12% contingency Contingency · 410 units | $610K | $580K | −$30K (-5%) | Fee holds |
| Coverage advice · monthly retainer Retainer · 12 units | $360K | $490K | +$130K (+36%) | Shadow bill above fee |
A shadow bill well above the fee means the firm is absorbing cost (renegotiation risk, or scope creep by you); well below means the fee is rich. Both are negotiation inputs, not savings.
How this is computed
DefinitionFee billed from AFA invoices; shadow bill from the firm's hours at approved standard rates (firms submit hours under the AFA terms); units from the matter record.Sourcesafa_agreement, invoice_line (fee + shadow hours), matterCoverage · n4 arrangements · shadow hours submitted on 94% of AFA invoicesFreshnessmonthly on invoice
Unapproved timekeepers
Billing without an approved rate or staffing approval · lines held, not paid| Timekeeper | Hours | Billed | First seen | Action |
|---|---|---|---|---|
| J. Marr Pembroke Shaw · Partner | 61 h | $39,650 | 12 Jul | Held |
| K. Osei Pembroke Shaw · Sr associate | 118 h | $53,100 | 3 Aug | Held |
| T. Nguyen Keaton Mallory · Associate | 74 h | $24,420 | 19 Aug | Held |
| Paralegal pool (4) Ashgrove Nunn · Paralegal | 212 h | $31,800 | 1 Sep | Held |
Approve adds the timekeeper to the approved list with a rate from the card; reject returns the lines to the firm with reason code LMG-1.4. Either way the decision is recorded against the firm's relationship profile.
How this is computed
DefinitionTimekeeper on a billed line not present in the approved-timekeeper list for the matter (or pool approval), or billing above the approved rate; new timekeepers are held at first appearance.Sourcesinvoice_line, timekeeper, approved_timekeeper, rate_card_versionCoverage · n4 timekeepers · $149K held this quarterFreshnesson invoice receipt
Firm efficiency index
100 = comparable cohort · below is better · same matter mix, same outcome band| Firm | Hours per claim vs cohort | Partner share | Write-down rate | Cycle vs cohort | Index |
|---|---|---|---|---|---|
| Thornbury Quill | 0.88× | 22% | 4% | 0.92× | 89 |
| Harbour Vale | 0.94× | 27% | 5% | 0.96× | 97 |
| Calder Rowe | 1.03× | 30% | 7% | 1.02× | 104 |
| Keaton Mallory | 1.08× | 34% | 8% | 1.06× | 111 |
| Pembroke Shaw | 1.31× | 41% | 11% | 1.18× | 130 |
Index = 0.45 × hours ratio + 0.25 × partner-share ratio + 0.15 × (1 + write-down rate) + 0.15 × cycle ratio. Inefficiency is excess hours and seniority at equal outcome, not a low write-down (see the trophy chart on the home screen).
How this is computed
DefinitionEach firm's matters matched to a comparable cohort (type, venue, stakes band, stage) with a minimum of 20; ratios are firm median ÷ cohort median; outcome band held constant.Sourcesmatter, invoice_line, cohort, outcome, timekeeperCoverage · n5 firms with n ≥ 20 matched matters · others abstainedFreshnessnightly; cohorts refreshed weekly
Billing guideline leakage
Recovered, reconciledRestored on appealIn processLeaked · unflagged (audit sample)Rate and timekeeper rules
$1.2M / $310K
Block billing and vague entries
$620K / $440K
Staffing (partner on routine work, >2 at hearing)
$480K / $380K
Expenses and disbursements
$290K / $110K
Pre-approval (research, travel, experts)
$310K / $190K
Flagged YTD
$3.9M
Recovered, reconciled
$2.9M
Leaked · unflagged estimate
$1.4M
Leakage is what the rules did not catch: estimated from a monthly specialist audit of a random 2% sample of auto-approved lines, extrapolated with a confidence interval. Block billing and staffing rules leak most; the fix is enforcement level (advise → specialist → auto) and rule precision, both set in Guidelines.
How this is computed
DefinitionFlagged = findings applied; recovered = reductions reconciled to payment net of reversals (ledger); restored = appeal reversals; leaked = audit-sample estimate of non-compliant amounts on unflagged lines, per rule family.Sourcesfinding, ledger_entry, appeal, audit_sampleCoverage · nall invoices YTD · audit sample 2% monthly (n ≈ 1,400 lines)Freshnessledger daily; audit monthly
Insourcing opportunities · recurring, high-volume work
Outside cost vs loaded in-house cost at the volume observed| Work type | Units / yr | Outside cost per unit | Outside spend | FTE needed | Net saving |
|---|---|---|---|---|---|
| Routine NDAs and vendor contracts Template + in-house paralegal | 480 | $1,850 | $888K | 0.9 | $699K |
| Standard subrogation demands In-house recovery team + automation | 1,100 | $620 | $682K | 0.7 | $535K |
| Subpoena and production responses In-house litigation support | 320 | $2,400 | $768K | 0.8 | $600K |
| Early-resolution motor claims (no suit) Partially: triage in-house, panel for suit filed | 1,240 | $2,850 | $3.5M | 3.5 | $2.8M |
$4.6M net annual saving if all four are insourced at a loaded cost of $210K per FTE, before recruiting lag. Candidates qualify on three tests: volume above 100 units a year, coefficient of variation of unit cost below 0.3 (the work is repeatable), and outside unit cost above 1.8× in-house. Everything else stays with the panel.
How this is computed
DefinitionUnits and unit cost from matters tagged routine by work type; FTE needed = units × median in-house hours per unit ÷ 1,600 productive hours; net saving = outside spend − FTE × loaded cost.Sourcesmatter, work_type, invoice_line, inhouse_cost_model (FTE register)Coverage · n4 work types pass all three testsFreshnessquarterly
Repeat players
Counterparties, opposing counsel and intermediaries that recur across matters| Party | Matters (36 mo) | Spend | Pattern |
|---|---|---|---|
| Plaintiff firm · Shine-type volume firm Opposing counsel | 212 | $6.1M | Settled 0.8× reserve · 190 d |
| Smash-repair chain (3 entities) Claimant / third party | 48 | $1.4M | Recovered 62% |
| National broker group Insured intermediary | 96 | $2.9M | Coverage dispute rate 2.1× book |
| Rehab provider network Treating provider | 140 | $900K | Medical cost 1.3× cohort |
Repeat players are where upstream fixes pay: a template clause, a settlement protocol, a panel-firm playbook or a supplier conversation changes the next twenty matters, not the next invoice.
How this is computed
DefinitionParties resolved to a master entity (name matching + registry ids) across matter_party; counted where the same entity appears on 3 or more matters in 36 months.Sourcesmatter_party, party_master, invoice_line, outcomeCoverage · nentity resolution confidence ≥ 0.9 on 92% of parties; the rest listed unresolvedFreshnessweekly
Moves on the table
Open units ordered by the move, not by invoice size · each has an owner, a due date and a follow-up observation| Move | Owner | Due | Evidence | Observed later |
|---|---|---|---|---|
| Authorize settlement on 38 QLD soft-tissue claims Reserve $60–120K; defense spend above 30% of reserve | Head of claims litigation | 9 Oct | Cohort n=186 · 0.80× reserve p50 | Outcome vs reserve at close |
| Approve routing rule: NSW property under $50K to Harbour Vale and Thornbury Quill Capacity and conflicts to confirm | Panel owner | 15 Oct | Cohort n=212/301 · p20–p80 shown | Cost per closed claim, 2 quarters |
| Fund or accept the $800K FY28 gap After $320K of modeled actions | General counsel · CFO | 31 Oct | Variance decomposition | Quarterly reforecast |
A saved scenario is not a decision; nothing here has a button until someone accepts the move. The month-18 metric is the share of moves followed and what those units then cost against their cohort.
How this is computed
DefinitionDecision records (F7.6): typed move with owner, due date, authority, evidence snapshot and a later observation (followed? cost vs cohort?).Sourcesdecision_record, matter, cohortCoverage · n3 open movesFreshnesslive
Where intelligence declined
Shown, not hidden: a refused comparison is a featureCoverage dispute CD-2211 · commercial property fire
Declined: First-of-kind in this book: cohort n = 6 after the widest rung (coverage only); minimum 20
Falls back to: Guidelines, rate card and a phase budget apply; re-checked monthly
Class action CA-0071 · add-on insurance
Declined: No comparable class actions in client history; pool not consented for this line
Falls back to: Phase budget from the firm's plan with a scope question at each gate
Smaller-panel ranking · NSW property under $50K
Declined: Dispersion check failed: cohort n = 31 but cost spread too wide to rank (CV 0.9)
Falls back to: Panel rotation by guideline; ranking withheld
Below the floor the screen says so and the file falls back to guidelines, the rate card and a phase budget. Buyers trust a system that declines more than one that always ranks.
How this is computed
DefinitionAbstain state (F6.6): cohort ladder exhausted (exact → drop stage → drop venue → coverage only) with n still below the pack minimum, dispersion check failed, or as-of date stale; every abstention is logged and counted.Sourcescohort, decision_recordCoverage · n3 abstentions open this monthFreshnessnightly
Discovery cost drivers
$9.8M of discovery / investigation spend YTD by component · driver metric vs comparable cohortDocument review (firm)
$3.9M
hours per 1,000 docs: 14.2 · cohort 11.0 · +29%
E-discovery vendor (processing + hosting)
$2.1M
GB processed per claim: 38 · cohort 22 · +73%
Medical records and IME
$1.6M
records per claim: 9.1 · cohort 7.4 · +23%
Depositions / examinations
$1.4M
per claim: 2.3 · cohort 2.0 · +15%
Expert reports
$800K
per claim: 0.9 · cohort 0.8 · +12%
The driver metrics are what to negotiate: review hours per thousand documents (technology-assisted review, staffing), data volume processed (custodian scoping, early case assessment), and expert count. Rates explain less than a fifth of the variance here.
How this is computed
DefinitionDiscovery spend = UTBMS L300 lines plus e-discovery vendor invoices and expert disbursements mapped to the matter; driver metrics from vendor invoices (GB, documents) and timekeeper narratives (hours per document batch).Sourcesinvoice_line (L300, E-codes), vendor_invoice, matter, cohortCoverage · nall claims with discovery spend YTD · vendor data on 81%Freshnessnightly
How this view is governed
Why the numbers can be trusted- One metric catalogue. Every tile reads a named metric with a published definition, grain and allowed filters. The dashboard and Ask Spend use the same catalogue, so a question and a tile never disagree.
- Coverage before value. Each metric reports how much of the population it covers (typed matters, recorded exposure, chainable timekeepers). Below a threshold the tile shows a floor and says so; it never fills gaps with estimates.
- Cohorts with a minimum n. Any comparison to "comparable" matters states the cohort definition and size; below n = 20 the comparison abstains.
- Ledger truth for money. Recoveries and savings come from the reconciled ledger, not from flagged amounts.
- Role scope. The GC sees all matters; claims leaders, legal ops and finance see their scope. Firms never see this view.