Most B2B lists are not wrong. They are aged. A contact who was the right person eighteen months ago now holds a different title, works somewhere else, or both. The account is still real; the reason to call it has expired.

That gap, between an accurate record and a current reason to act, is where outbound programmes lose budget without ever being able to point at the failure.

The pain point: a clean record and an expired reason

Picture the standard artefact. A list of accounts and contacts, refreshed quarterly, enriched with industry codes and headcount bands. By most measures it is clean.

Then a sales director asks one question about one account: why this one, this week? The honest answer is that it appears on a list, scored well against a firmographic profile, and has not been contacted in ninety days.

None of that is a reason a buyer would recognise. It is a reason a system produced.

What the public record actually provides

Before designing a process around buying signals, it helps to be precise about what a primary source gives you. The SEC's EDGAR system works as an example because it is public, free and documented.

The SEC describes EDGAR as free public access to millions of documents filed by publicly traded companies and others, and publishes a Latest Filings view of submissions as they enter the system, plus daily listings by form type within the past week.

Two features matter for prospecting. Full-Text Search covers filings submitted electronically since 2001, including attachments such as exhibits, and supports keyword, ticker, company name, CIK and reporter searches with Boolean operators and date filters. The EDGAR APIs expose submission history by filer and XBRL data from financial statements.

Read that as a capability statement. Structured, dated, attributable events about named legal entities are available without a subscription.

Three limits that change how you build

Search is precise, not conversational

The Full-Text Search FAQ states plainly that natural language search is not currently supported. Filing search rewards defined terms and defined events. It does not infer intent.

That should shape both the pipeline and the claims made about it. A monitor configured for a specific form type and a specific phrase is a precision instrument. Calling it an intent engine overstates it.

Filing date is not reporting date

The EDGAR browse view lists form type, filing date and reporting date side by side, and they routinely diverge. A document published this week can describe a period that closed months earlier.

This is where "current" quietly becomes ambiguous. If recency is defined only by the publication timestamp, a pipeline will surface stale context wearing a fresh date, and a rep will discover the mismatch during the first call.

An entity is not a buyer

Filings name legal entities. They do not name the person who owns the problem, holds the budget, or controls the calendar. Everything between the filing and a named role is inference, and inference should be labelled as inference.

Where qualification actually breaks

  • Contact drift. The record is accurate and the human has moved. Enrichment fixes the field; it does not restore the relationship.
  • Inference inflation. A predicted need is retold across two handoffs as an observed fact, and nobody can say where the claim originated.
  • Missing provenance. No citation means no reconstruction. A rep inherits a queue and cannot answer "why now?" without asking the person who built it.
  • Undefined disposition. Opportunities are worked or ignored, but the reason is never recorded, so precision never improves.

Each is a documentation failure before it is a technology failure. That is encouraging, because documentation is cheaper to repair than data.

Seven decision criteria

  • Citation. Can each opportunity point to a document, a form type, a filing date and a passage?
  • Recency definition. Is "current" defined by publication date, reporting period, or both, and is that stated on the record?
  • Observed versus inferred. Are filed facts and predicted needs stored in separate fields, with the inference labelled?
  • Entity resolution. Are subsidiaries, former names and name variants reconciled, so one event does not create six duplicates?
  • Role path. Is there a stated, checkable route from entity to role, or is the contact simply the first record returned?
  • Disposition capture. Does every worked opportunity record a reason when it is disqualified?
  • Precision measurement. Is the programme scored on the share of worked opportunities that produced a substantive conversation?

The seventh is the one most often replaced with a volume target. Volume reports easily and says nothing about relevance.

Three options and their real costs

Buy a refreshed list. Fastest route to more contacts and current firmographics. The cost is structural: firmographics describe a company, not a moment. You still cannot answer "why now" with anything better than profile fit and contact gap.

Build a signal pipeline. Monitor defined filing types and defined events, resolve entities, route to owners. Maximum relevance and full provenance. The cost is permanent: someone maintains the term list, the entity mapping and the deduplication rules, and someone owns the disposition data.

Signal feed plus human qualification. Automate detection, keep judgement with a person. Most common in practice, and the option most likely to drift, because the handoff point is where accountability usually goes unassigned.

None of these is wrong. The error is choosing one by default, then discovering at the first pipeline review which one you actually built.

A 30-day pilot that produces evidence

  • Week one, define the event. Pick one filing type or one hiring pattern that plausibly precedes a need you can serve. Write down the exact terms and the date window.
  • Week two, record the citation. For every opportunity, store the document reference, form type, filing date and reporting period. If the record cannot carry those fields, fix that first.
  • Week three, work a small batch. Include two accounts you expect to disqualify. Time how long a rep takes to reconstruct why an account is in the queue without asking anyone.
  • Week four, score and decide. Measure precision, reconstruction time and the distribution of disqualification reasons. Extend or narrow the term list on that evidence.

What to measure, without inventing a number

Four indicators are worth baselining before the pilot and re-measuring afterwards, and none requires an estimate you cannot defend.

Precision: the share of worked opportunities that produced a substantive conversation. Latency: elapsed days between filing date and first touch. Reconstructability: the share of worked opportunities whose reason for contact could be recovered from the record alone. Disposition coverage: the share of closed opportunities carrying a recorded reason.

Baseline all four. A defensible improvement claim comes from a measured before and after inside one operation, not from a benchmark borrowed from a vendor deck.

Where DealFORGE fits, stated factually

JWilliams & Associates positions DealFORGE as finding current B2B lead opportunities using buying signals from official filings, hiring activity and predictive business needs. The intended use is evidence-based opportunities to research and contact, not guaranteed purchases or closed accounts.

That distinction is the substance of this article. A signal identifies an account worth investigating and supplies a defensible reason to investigate now. It does not establish that the organisation has a budget, an owner, or a problem you can solve.

For evaluation, DealFORGE is offered as a free trial that provides at least 50 leads participants can keep and work independently, with no obligation to purchase or to pay for continuing lead-agency work. Those are the terms. No trial duration, credit-card requirement, signup URL, refund terms or conversion outcome is added here.

No customer results, win rates or savings are claimed, and none should be inferred.

Evidence notes and limits

Evidence was reviewed through 19:00 UTC on 29 October 2026. The description of EDGAR, its Latest Filings view, the coverage and operators of Full-Text Search, the API capabilities, and the distinction between filing date and reporting date are drawn from SEC documentation and help pages published on sec.gov.

Those pages are tool documentation, not research on contact decay or pipeline conversion. No decay rate, response rate, win rate or cost-per-lead figure is asserted here, because none was verified in the material reviewed.

The DealFORGE description and trial terms are limited to owner-approved facts current as of 27 October 2026. The criteria, options, pilot plan and metrics are analysis and suggested method, not a tested protocol.