Qualifying from Public Filings

The workflow

Play 31

Qualifying from Public Filings

1Name the fact that qualifies

The qualifying question

FootprintOwned vs leasedWho maintains it

The fact that decides fit, which firmographic databases do not carry. Every field allows unknown.

2Route public and private

Target accounts, pulled into a Clay tableName, domain, country, CRM id

Public filerThe filing is the answer

10-K Item 220-FFiscal year

Pin the run to one named filing.

Source

SEC EDGAR

Read

Claude

Private companyNo filing

Published totalsStore countFormat sizes

Published totals first, then a build-up per format, flagged as an estimate.

Search

ExaFirecrawl

Read

Claude

3Record only what is disclosed

The source disclosesThe record gets

Square footage split by tenure→Owned and leased records

Split counts, one total→One record, counts as fields

Store count, no size→Build-up per format, estimate

Nothing usable→Unknown, with the reason

Never apportion a total by a site-count ratio

4Guard in code, not in the prompt

Gate1

No blanks

Every field returns a value, or unknown with a reason.

× Blank fieldRe-run

Gate2

No hidden estimates

Arithmetic in the method forces the estimate flag on.

× Unflagged mathFlagged as estimate

Gate3

Source attached

Each field carries its filing, fiscal year and URL.

× No sourceHeld back

5Audit, then load

Check against the filing

Open a sample of rows and match each value to the page it cites. Report coverage per field.

Write to the CRM

Value, source and estimate flag land together. Unknown becomes the rep's first question.

Write

HubSpot

Accounts qualified on cited facts

The stackHubSpotSEC EDGARClaudeExaFirecrawl

Oriel · GTM plays library

Open full size

The diagram as text

1. 1

   Name the fact that qualifies

   The qualifying question: Footprint, Owned vs leased and Who maintains it. The fact that decides fit, which firmographic databases do not carry. Every field allows unknown..

2. 2

   Route public and private

   Target accounts, pulled into a Clay table (name, domain, country, crm id): Tool: HubSpot.

   Public filer (the filing is the answer): 10-K Item 2, 20-F and Fiscal year. Pin the run to one named filing. Source: SEC EDGAR. Read: Claude. Tool: SEC EDGAR.

   Private company (no filing): Published totals, Store count and Format sizes. Published totals first, then a build-up per format, flagged as an estimate. Search: Exa and Firecrawl. Read: Claude.

3. 3

   Record only what is disclosed

   The source discloses → The record gets

   • Square footage split by tenure → Owned and leased records
   • Split counts, one total → One record, counts as fields
   • Store count, no size → Build-up per format, estimate
   • Nothing usable → Unknown, with the reason

   Never apportion a total by a site-count ratio

4. 4

   Guard in code, not in the prompt

   Gate 1: No blanks. Every field returns a value, or unknown with a reason. Rejected: Blank field, Re-run.

   Gate 2: No hidden estimates. Arithmetic in the method forces the estimate flag on. Rejected: Unflagged math, Flagged as estimate.

   Gate 3: Source attached. Each field carries its filing, fiscal year and URL. Rejected: No source, Held back.

5. 5

   Audit, then load

   Check against the filing: Open a sample of rows and match each value to the page it cites. Report coverage per field..

   Write to the CRM: Value, source and estimate flag land together. Unknown becomes the rep's first question. Write: HubSpot. Tool: HubSpot.

   Outcome: Accounts qualified on cited facts. Tools: HubSpot.

The stack: HubSpot, SEC EDGAR, Claude, Exa and Firecrawl.

Open the diagram full size (1180 px wide)

The play in six steps

1. 01

   Step 1: Name the fact that actually qualifies

   Data vendors sell industry and headcount. The fact that decides fit, such as footprint, tenure or who carries the maintenance, usually lives only in filings.

   • Write down the question the buyer would ask on a first call, then the field that would answer it.
   • Test the obvious proxy on real accounts before trusting it. Owning buildings is not the same as maintaining them: a net-lease landlord owns everything and fixes nothing.
   • Give every field a closed set of answers, and make unknown one of them.
   • Unknown is an acceptable answer. Blank is not.
2. 02

   Step 2: Route public filers and private companies

   Public and private companies need different research methods, so every account is routed before anything is read.

   • Pull target accounts from HubSpot into a Clay table: name, domain, country and CRM id.
   • Route each one: public filer, private operator, private landlord, or not relevant to the question.
   • Public filers go to the filing. Private companies go to published totals and a build-up.
   • Tag out-of-scope segments instead of deleting them, so reopening one later costs nothing.
3. 03

   Step 3: Read the filing for public companies

   For a listed company the annual filing already holds the answer, so the research starts at SEC EDGAR, not a search engine.

   • Name SEC EDGAR as the first source and the latest 10-K as the document. Item 2, Properties, holds site counts, owned versus leased and total square footage.
   • For foreign filers, use the 20-F or the annual report and say which one was used.
   • Claude reads the section and returns structured fields, each with the filing, fiscal year and sec.gov URL behind it.
   • Pin each run to one named filing. Without it, the same company returns different numbers on different runs.
   • Split records only where the filing splits them.
4. 04

   Step 4: Research private companies without guessing

   Private companies file nothing, so the method changes and every estimate has to say it is one.

   • Use Exa and Firecrawl to find published totals: press releases, investor pages, landlord tenant lists.
   • With no published total, build one up: store count times typical size, per format, with the working shown.
   • Never apportion a total by a site-count ratio. Owned and leased sites are rarely the same size.
   • A labeled estimate beats a blank. An unlabeled estimate is worse than both.
5. 05

   Step 5: Enforce the rules in code

   A rule in the prompt loses to the model often enough to matter. A check that runs after the model does not.

   • Reject any row with a blank field. Unknown must come with a reason.
   • If the method text contains arithmetic, force the estimate flag on, whatever the model returned.
   • Mark any ratio-based figure as apportioned and keep it out of the CRM.
   • Hold back any field with no source attached.
   • Cache every raw response so a parse failure can be re-read without paying for the research twice.
6. 06

   Step 6: Audit, then load to the CRM

   Coverage and accuracy are measured before a single field reaches a rep, because a wrong cited number is worse than no number.

   • Open a sample of rows and check each value against the filing it cites.
   • Report coverage per field, on the in-scope accounts and on the full list separately.
   • Write the fields to HubSpot with their source and estimate flag beside them.
   • Every unknown becomes a qualifying question the rep asks on the first call.

Tools used

The five tools this play names. Each links to the other plays that use it.

HubSpotSEC EDGARClaudeExaFirecrawl

Related plays

Play 30

Public Records Signals

1Pull public records on a schedule

Government APIsFree, read-only

Disaster declarationsWeather eventsSEC filings

Daily for declarations, weekly for filing phrase searches.

Pull

Plain GET on a schedule

Portals with no APIReal browser

Court docketsPermits

Throttled, sessions recycled. Portals behind a CAPTCHA are out.

Drive

BrowserbaseFirecrawl

One row per recordDate, source URL, verbatim excerpt

2Match each record to an account

Entity resolutionThe hard part

A record names a legal entity or a place, not your CRM account.

Name

Strip Inc, Corp, LLC, The

Place

County or address to known sites

Against

HubSpotSalesforce

MatchConfidence

Name and location agree→High: routes on its own

Name only→Medium: review queue

Fuzzy or alias→Low: review queue

3Qualify in two gates

Gate1

Not suppressed

Customers, open deals and competitors come out first.

× Customer or open deal

Gate2

Actually relevant

Claude reads the excerpt and gives a one-line reason.

× Not relevant

4Tier what the record allows

Act aloneThe event is the reason

Owner alertThis week

Record link, excerpt and matched sites, straight to the owner.

Alert

Slack

Log

HubSpot

One more factPerson or site first

Find the person

The right contact, then a checked email.

Find

AI Ark

Email

ProspeoFindymail

Verify

BounceBan

Rank onlyContext, no message

Score only

Adds weight to the account score. Never opens a message.

On

HubSpot

5Reach out with the record as evidence

State the public fact plainly

Something the buyer can check for themselves, with the source link at hand.

Email

Smartlead

Call

Nooks

A dated reason to call this week

The stackSEC EDGARBrowserbaseFirecrawlClayHubSpotSalesforceClaudeSlackAI ArkProspeoFindymailBounceBanSmartleadNooks

Oriel · GTM plays library

Play 30Outbound · ABM

Public Records Signals

5 steps · Email, Call, CRM

+8 more

Play 23

HubSpot Account Tiering & Scoring

1Learn from the deals you already closed

Closed won and closed lost, last 12 to 24 months

Closed wonWhat good looks like

ACVSales cycleExpansion

Start with the highest revenue customers. Look for patterns in industry, size and geography.

Closed lostWhat to rule out

Poor-fit industrySize mismatch

The disqualification reasons that keep coming back.

2Define segments and scoring logic

SegmentsOne per account

Company sizeIndustryBusiness model

Scoring inputWhat it checks

Firmographics→Size, industry, location

Technographics→Tech stack

Signals→Recent activity at the account

3Build the Clay model and tier

Core fields

CompanyDomainLinkedIn URL

Enrichment fields

EmployeesRevenueIndustryTech stack

Logic fields

SegmentTierLast enriched

Clay formulas assign the tier

Tier 1Strong fit

Tier 2Good fit

Tier 3Low fit

4Sync tiers into HubSpot

New HubSpot properties

Account tierSegmentKey firmographics

Company and contact data mapped field by field, so the sync never creates duplicates.

Data

Clay

Emails

Findymail

Verify

BounceBan

Write

HubSpot

5Keep tiers live with workflows

Tier workflowsProperty trigger

Company workflows with re-enrollment on, so the tier changes when the data does.

FallbackMissing tier

Accounts without a tier are flagged, not left to slip through.

FreshnessQuarterly

Re-enrich any account last enriched over 90 days ago, and re-check its contacts for job changes.

Job

LinkedIn

Monthly data review, quarterly strategy update

Every account tiered in HubSpot

The stackHubSpotClayFindymailBounceBanLinkedIn

Oriel · GTM plays library

Play 23ABM · Outbound

HubSpot Account Tiering & Scoring

7 steps · CRM

Play 21

Pain-Qualified Segments from Calls

1Gather calls and customers

Sales callsBuyer side

DiscoveryDemoClosing

What was true for the buyer when they chose to talk.

From

Fathom

Closed won and lostBacktest set

Close dateReasonAccount

The customers every segment is checked against.

From

HubSpot

2Mine the buying situations

One schema per call

SituationWhy nowAlready triedVerbatim painOutside-in fingerprint

Every situation is counted across calls. The buyers' exact phrases go into a language bank.

Read

Claude

3Situation plus tension

Segment anatomyTwo facts pulling apart

Doing

The thing they are doing

Hurts

The thing that makes it hurt

Now

A deadline, target or season

Number

The figure that costs the quarter

4Membership test, field by field

Condition isUse it as

Public: the field states it→Filter and copy

Proxy: the field suggests it→Score and ask, never assert

Not public→A question for the call

5Measure before building

Gate1

Count it

Free counts, with the query and the date recorded.

× Too small or too broad

Gate2

Eyeball five rows

Open the evidence. Two of five wrong means the test is wrong.

× Fix the test

Gate3

Backtest for lift

Customers in the 90 days before they bought, against a TAM sample.

× Low lift

6Build the list and the opener

Refreshable list

Discover on one filter, score on the rest. Unknown never passes. Emails re-verified on every refresh.

Find

Explorium

Evidence

Exa

Build

Clay

Verify

BounceBan

Opener in their words

Names the situation, says where you saw it, asks if the read is right.

Email

Smartlead

LinkedIn

HeyReach

Segments with measured lift, in the buyers' own words

The stackFathomHubSpotClaudeExploriumExaClayBounceBanSmartleadHeyReach

Oriel · GTM plays library

Play 21Outbound

Pain-Qualified Segments from Calls

6 steps · Email, LinkedIn

+3 more

All GTM plays