Public record31 exhibits are out.Oklahoma · West v. State Farm

BadNeighbor.law · A Denham Law public-records project

Case file West v. State FarmRecord 31 exhibits · 285 pages Public record

Like a
goodbad
neighbor?

A good neighbor doesn’t treat your roof like a profit lever.

We went page by page through all 31 numbered exhibits for the edits, oddities, internal doubts, tracked reversals, monitored scripts, and agent warnings that did not make the headline. Policyholder advocates should see the deep cuts.

Take me past the headlines Open the court filing

Their chartExhibit 10

Actual State Farm chart tracking claims closed without payment and later paid as a total roof.

WEST00001274 · FILED EXHIBIT 10

7,474

“overturn cases” shown, Jan–Oct 2021

The exhibit says a status change does not always mean the first call was wrong. Click for the full context.READ IT YOURSELF ↗

The stuff between the headlines

Their words get weirder in the small print.

The blockbuster figures matter. So do the group chats: salivating emoticons, “lol,” a tongue-out face, a renamed target, a training-paper-trail debate, and alarms from inside State Farm’s own agent network.

Exhibit 8 · State Farm IM archiveMarch 3, 2021

The dashboard had them making faces.

No parody needed. After discussing a Wind/Hail dashboard and a focus on Texas, Illinois, and Georgia, the filed transcript records this exchange:

Kathy
“Will that work for you?”

Nicole
“i’m drooling :o3:o3:o3”

Kathy
“lol!”

Nicole
“Passion project...I am so excited and optimistic (and I believe steeped in reality) that we WILL realize improvements in Wind/Hail trends in ’21”

Kathy
“Silver platter for you.”

Nicole
“:P”

Kathy
“lol...”

Context: Excited chat tone is not proof that an individual claim was handled incorrectly. It is, however, the tone used in the filed record while discussing the Wind/Hail dashboard and expected trend improvements.Exhibit 8 · PDF page 46 · HOSIERGAR0072891PROD ↗

Filed Exhibit 8 State Farm instant-message transcript containing I am drooling, three colon-o-three faces, lol, Silver platter for you, and a tongue-out face.

The actual group chat ↗

Exhibit 7 · One transcript, two topics

$78.8 million.
Then, 22 minutes later: “jerk move.”

After the percentage-point calculation and some scheduling messages, the transcript veers into meeting-room drama. These lines concern a separate office exchange—not the claim math above them.

Tom: “did Eric just tell me not to speak in the Dec meeting?”

Nicole: “He may have - was it a joke?”

Tom: “don’t think so”

Nicole: “jerk move”

Tom: “doesn’t like some of our direction but it’s not his call”

Nicole: “so not cool”

The roughly 22-minute interval is calculated from the timestamps on the page. The office comments do not establish anything about claim outcomes.Exhibit 7 · PDF page 45 · WEST00080627 ↗

Filed Exhibit 7 State Farm instant-message transcript showing the 78.8 million dollar calculation followed later by jerk move and so not cool.

Read the page in order ↗

Four bonus lines from the record

Corporate chat bingo was undefeated.

Every quote below links to the filed page. The punchlines are real; the caveats stay attached.

Ex 7 · p.44 · WEST00080626“cool beans”“Or am I just snowing myself”

The same internal reality check asks whether the results that matter actually prove improvement.

Questions in a chat are not an admission that the results were false.Ex 9 · p.47 · HOSIERGAR0070281PROD“Good grief”“Youre good.”

The plot twist: “I likely was not supposed to be invited to this call.”

The same chat discusses review thresholds and tougher coverage conversations; the invitation mistake proves no misconduct.Ex 26 · p.143 · HOSIERGAR00002589PRODGolf balls:
a two-way peril.Accidental list. Intentional list.

The formal guide puts “golf balls” in examples of both accidental and intentionally inflicted mechanical roof damage.

Classification depends on the facts. The guide does not prove fabricated damage in any particular claim.Ex 17 · p.75 · WEST00004150“bleeds State Farm logos”#ClaimTimeIsGameTime

An agent used both phrases while warning leadership that the roof-claim system was “not only broken.”

This is one agent’s warning—and her report of a roofer’s inverted slogan—not a company-wide finding.

The dashboard crushEX 08

The dashboard came on a silver platter.

“i’m drooling :o3:o3:o3”

The same exchange calls the Wind/Hail dashboard a “Passion project” and predicts improved 2021 trends.

Internal State Farm chat calling the Wind and Hail dashboard a passion project, saying I am drooling, and calling it a silver platter.

Actual excerpt ↗

Exhibit 8 · PDF page 46 · HOSIERGAR0072891PROD ↗Keep it honest: Enthusiasm about a dashboard does not prove that any individual claim decision was wrong.

The euphemism editEX 15

The target did not disappear. The word did.

“Changed Target to Selection”

The recap also says 13 of 21 slides were deleted and an execution-level control section was removed. The next page says a graph lost its glide path and baseline.

State Farm deck-edit recap stating Changed Target to Selection and Took off Execution level control section.

Actual excerpt ↗

Exhibit 15 · PDF pages 66–67 · WEST00020078–79 ↗Keep it honest: This is a draft-deck recap. The exhibit does not prove that the edits falsified a result.

The paper-trail paradoxEX 25

Do not write “mandatory.” Do track completion.

“all should attend but not mark as mandatory”

The draft says that label could become a problem if someone failed to attend—then asks for a “mark complete” requirement. Texas and Oklahoma were proposed for early emphasis.

Draft State Farm training message discussing avoiding the word mandatory while tracking completion, with Texas and Oklahoma highlighted.

Actual excerpt ↗

Exhibit 25 · PDF page 132 · WEST00073885 ↗Keep it honest: This was a draft training communication, not proof of the final wording or who completed it.

The agent alarmEX 17

One State Farm agent put it at 8 out of 10.

“we end up paying what we should have in the beginning”

The agent said the delay left a mad customer and tarnished reputation. State Farm’s summary says her outreach followed a review of 10 prior roof claims.

State Farm agent email saying eight out of ten times the company ended up paying what it should have paid in the beginning.

Actual excerpt ↗

Exhibit 17 · PDF page 75 · WEST00004150 ↗Keep it honest: That is one agent’s account from a small, non-random set—not a company-wide error rate.

The slide about the emailsEX 19

The email-risk slide wanted… fewer words.

“limited information (visual cues—not a lot of content)”

The same outline lists email capture, litigation searches, and the risk of messages being taken out of context; details would go in speaker notes.

State Farm outline listing litigation email searches, out-of-context risk, and a slide with limited information and details in speaker notes.

Actual excerpt ↗

Exhibit 19 · PDF page 86 · WEST00025909 ↗Keep it honest: This may be ordinary presentation and litigation hygiene. It does not direct anyone to destroy or conceal records.

The number they questionedEX 13

Even Claims was surprised by 98%.

“98% approved does not seem to be consistent with prior year findings.”

The email asks whether the data came from overturn calls or from structured phrases tied to small hail and light wind.

Internal State Farm email questioning a 98 percent approval figure and asking whether it came from overturn calls or structured phrases.

Actual excerpt ↗

Exhibit 13 · PDF page 57 · WEST00021540 ↗Keep it honest: The referenced spreadsheet is absent. No denominator, state mix, or definition of “approved” appears here.

Four more buried details

You could not make up the footnotes.

Ex 18 · p.77“A secret society”

An agency sales leader’s description of how agents felt Claims shared information.

Ex 27 · pp.155–1562 hours → 1:04?

The shortest seven training segments totaled 64 minutes after seven segments could satisfy the requirement. The emails do not show who watched which mix.

Ex 30 · p.171≥50 and ≤50

The same slide labels “light wind” both 50 mph and up and 50 mph or less—apparently a typo inside a threshold-driven program.

Ex 26 · pp.140–141Inspectors can damage roofs, too.

The guide says handler-caused footfall goes into the adjustment—and no deductible applies when that is the only damage. This one cuts the other way.

The dashboard room

The little charts tell the big story.

These demonstratives use only numbers printed in the exhibits. The reconstructed views make the comparisons legible; every panel also includes the actual filed page and the limitation that belongs beside it.

Exhibit 4 · The benchmark gap

85% versus 8–10%.
65% versus about 5%.

An internal chat says State Farm’s analysis found far more full-roof replacements for small hail and light wind than an asserted industry average. It says the dollars narrowed the “focus of opportunity.”

No denominator, geography, time period, or underlying benchmark study appears on this page.Exhibit 4 · PDF page 35 · HOSIERGAR0071137PROD ↗

0%50%100%

Small hail

85%

8–10%

Internal State Farm figureAsserted industry average

Light wind

65%

about 5%

Internal State Farm figureAsserted industry average

See the sentence in the file ↗

Exhibit 10 · Call quality tracking

Closed without payment.
Later paid as a total roof.

The exhibit’s line is labeled “# of Overturn Cases.” The ten displayed monthly values total 7,474 from January through October 2021.

7,474displayed cases

The page expressly says the change may not mean the first call was incorrect; one example is a withdrawn claim later resubmitted for inspection.Exhibit 10 · PDF page 51 · WEST00001274 ↗

465

Jan

376

Feb

464

Mar

520

Apr

1,008

May

888

Jun

914

Jul

1,299

Aug

950

Sep

590

Oct

Monthly values transcribed from the filed chart. No inference added.

Actual State Farm call quality tracking chart showing total roof payments and overturn cases by month.

Read the chart and its caveat ↗

Exhibit 30 · February versus June 2021

The scripts went up.
Roof approvals went down.
The buckets changed, too.

The slide puts rising phrase/documentation usage above lower manager approval percentages. But its February and June approval categories are not the same.

No claim counts, denominator, geography, selection rule, or statistical method is disclosed. This is not a clean like-for-like comparison and does not prove causation.Exhibit 30 · PDF page 173 · WEST-RFP44-Sub. Prod.000127 ↗

Documentation / phrase use

AccuWeather phrase use22.3%→64.2%

QFC + ISO documented18.7%→42.8%

Scope sheet documented18.5%→57.5%

Full-roof TM approval

Wind only80.9%

→

Light wind only42.7%

Wind & hail73.8%

→

Combined45.8%

Hail only54.8%

→

Small hail only49.3%

Actual Wind and Hail February 2021 versus June 2021 slide showing phrase use, documentation, and manager approval percentages.

Notice the category labels ↗

Exhibit 7 · Their hypothetical

One point.Five thousand claims.$78.8 million.

That is the arithmetic in the internal chat: about 5,000 claims for each percentage point, multiplied by $15,769 average severity. The same thread then wandered into “jerk move” and “so not cool.” Corporate chat is still chat.

Exhibit 7 · PDF page 45 · WEST00080627 ↗

1 percentage point$78.8M

5,000 claims × $15,769 average severity, using the internal email’s arithmetic.

Move their hypothetical

1 point10 points

Demonstrates the math stated in Exhibit 7. It is not an estimate of wrongful denials, damages, or what any court has found.

Actual State Farm internal chat calculating 78.8 million dollars per percentage point and later saying jerk move.

The math—and the office chat—on one page ↗

A chart is not a verdict.It is a roadmap to the questions State Farm still needs to answer.

The neighbor test

The slogan is warm.
The documents aren’t.

Insurance is a promise bought before disaster. When internal records put financial opportunity, management controls, payment ratios, and indemnity reductions into the same story, policyholder advocates should ask who the system was built to serve.

What policyholders buy

Inspect the damage.
Apply the policy.
Pay what is owed.

What the released record discusses

Opportunity.
Controls.
Ratios.
Less indemnity.

A neighbor helps fix the roof. A claims system should never put the financial target first.

The receipts

Don’t trust our summary.
Read their pages.

Every image below is a direct crop from a filed exhibit—not reconstructed text. Open the full page, inspect the surrounding language, and decide what deserves an answer.

The opportunityEX 03

When “full roof replacements” become a “bucket of opportunity.”

An internal instant-message transcript discusses tracking small-hail and light-wind claims and matching them to roof-replacement payment codes.

Exhibit 3 · PDF page 34 · WEST00011454

State Farm internal instant-message transcript with the phrase full roof replacements is our biggest bucket of opportunity highlighted.

ACTUAL EXHIBIT

The numberEX 06

“About a $1.4B decrease in indemnity.” Their figure. Their email.

The email also compares total-to-partial roof-replacement ratios after the wind-and-hail initiative began and expanded.

Exhibit 6 · PDF page 43 · WEST00020276

State Farm internal email reporting about a 1.4 billion dollar decrease in indemnity from 2020 to 2021.

ACTUAL EXHIBIT

The nameEX 15

Quality materials consolidated into a “Fix Profit Task Force Deck.”

The released email says several slide decks were combined and describes planned changes to controls, targets, selections, and findings.

Exhibit 15 · PDF page 66 · WEST00020078

State Farm internal email naming a Fix Profit Task Force Deck.

ACTUAL EXHIBIT

The handoffEX 29

The guidelines did not simply disappear. They moved into standard procedures.

A wind-and-hail training email says FME guidelines were sunset and moved to SCPs and operation guides.

Exhibit 29 · PDF page 164 · WEST00016926

State Farm training email stating FME Guidelines sunset and moved to SCPs and OG.

ACTUAL EXHIBIT

Context is not a loophole.It is why every screenshot links to the complete filed page.

The playbook

This does not read like one spicy email.

Read together, the exhibits describe design, controls, rollout, measurement, and continuity. That pattern is why the records matter beyond any single headline.

Name the opportunity

Internal records connect roof-replacement frequency to indemnity spending and potential financial impact.

Exhibits 3–5

Build the controls

Plans describe manager review, scripts, training, dashboards, targeted reinspections, and tracked completion.

Exhibits 9, 23, 25–30

Scale the program

A rollout deck records a Texas pilot, expansion across Texas, and implementation in all states by December 2020.

Exhibit 30

Measure the result

Later records track roof-payment ratios, approval percentages, reason codes, and large reported indemnity decreases.

Exhibits 6–7, 10, 13

The record alleges a sequence:A program. A rollout. A reported result.

For policyholder advocates

Turn outrage into informed pressure.

These records are most powerful when advocates use them precisely—connecting the corporate program to the actual path of an individual claim.

Teach from the page

Use the actual screenshots, exhibit number, Bates number, and full-page link. Let the record—not a slogan—lead the conversation.

Ask about the claim path

Who inspected? Who reviewed the roof recommendation? What dashboard, reason code, script, training, or written procedure touched the decision?

Demand the missing data

Ask for Oklahoma cohorts, claim counts, definitions, methodology, manager-review records, successor procedures, and training histories.

Share without overclaiming

Say what the documents show, link the full context, and be clear that State Farm disputes wrongdoing and the merits remain undecided.

Ask on every affected roof file

Who made the call—and what program, metric, script, training, manager review, or procedure shaped it?

No more neighborly fog

State Farm should answer these. In public.

01
Why was the full-roof-replacement rate discussed as a financial opportunity?
02
What caused the reported $1.4 billion decrease in indemnity—and how much came from fewer total roof replacements?
03
Did manager review improve claim accuracy, reduce valid payments, or both?
04
Which procedures replaced the FME guidelines, and how did they affect Oklahoma policyholders?
05
How many claims were selected, reviewed, reclassified, reinspected, or paid differently because of the program?
06
Were policyholders told that these controls or financial measurements could affect the path of their roof claims?

State Farm’s position

State Farm disputes wrongdoing and says it evaluates claims fairly and pays what it owes under the policy.

Read its response ↗

What is decided

The court has not decided the merits. Public release is not a finding that any particular claim was underpaid.

Why publish anyway?

Because unresolved questions this serious deserve evidence, scrutiny, and answers—not another slogan.