Methodology

How we decide what belongs in the record, how we weigh evidence, and where the limits of coverage lie.

Syglum is a source-verified record of material competitive activity across 30 global law firms. Every published record is grounded in documented source material. Each event includes the source URL, supporting evidence and a confidence label, so users can trace the factual basis directly. Automated systems may assist with processing and presentation, but they do not replace the underlying evidence standard.

This page explains what we track, how we decide what to include, and how we evaluate and present evidence.

What Syglum tracks

We record material events across the 30 covered firms in five categories: partner and team moves; major transactions and client mandates; litigation and regulatory matters; legal tech and AI; and firm strategy and offices. Each category is defined by the kind of competitive signal it carries, not by the source that reported it. A partner move is tracked because it changes the talent and client relationships available to a firm; a transaction is tracked because it signals market position and client access; a litigation matter is tracked when it places the firm in a material public role; a strategy or office move is tracked when it materially changes the firm’s footprint or competitive posture.

What gets included

Our default threshold for transactional records is one billion dollars in disclosed or reliably estimated value. At or above that level, a deal, mandate, or similar matter is ordinarily eligible for inclusion. We make explicit exceptions regardless of value for partner and team moves, litigation and firm-as-party matters, flagship mandates that shape a practice area, and first-of-kind legal technology or AI deployments.

Items that fall below the threshold and do not meet an exception are not published. An inclusion decision can be revisited if new information changes the assessment.

How sources are evaluated

We rank sources by proximity to the underlying fact. Primary records come first: court filings and dockets, regulatory filings, transaction documents, and other official records. Firm and client announcements come second, treated as direct but interested statements. Reputable legal and business reporting comes third, including Reuters, Bloomberg Law, Law360, and Law.com, where the reporting is attributed and the publication maintains a corrections policy. Official social media posts are used only when no better source exists and the post comes from a verified account of a party to the event.

We do not rely on anonymous message boards, unverified social accounts, or recycled aggregators as primary support.

Confidence labels

High confidence means the material facts of the record are supported by a primary or direct source, or by multiple credible corroborating sources that independently confirm the key details. Medium confidence means credible reporting supports the event, but a primary source is unavailable or one material detail is less directly supported.

There is no Low tier. Records that do not meet our publication bar are held back until better evidence is available or the event is no longer competitively relevant. A Medium label is not a weakness; it is an honest statement about the state of the public record.

Evidence and conflicts

Every published event includes a source URL, the exact evidence quote, the named source, and corroboration where available. When credible sources disagree on a material detail, we preserve the disagreement in a source note rather than smoothing it away. If new evidence contradicts an earlier record, we update the record and note the correction. The goal is traceability: a reader should be able to see why a record says what it says and decide whether that basis is sufficient for their purpose.

Briefs

Briefs are generated only from records already published in the database. Query filters determine which records are eligible before generation. Each substantive statement in a Brief must resolve to one or more eligible event records, and citations are rendered from those event IDs rather than created by the language model. When too few qualifying records exist, the system returns an insufficient-activity message instead of broadening the query.

Update cadence and corrections

The database is swept twice daily for new source material. Records are updated as new evidence emerges, not frozen at first publication. Each record carries a version number, an update timestamp and the date it was last verified against its sources, and a material correction is reflected in the record's evidence and source notes.

Coverage limitations

Syglum is public-source intelligence. It is selective, not exhaustive. Some events surface later than others because source availability varies by jurisdiction, practice area, and the willingness of the parties to announce a matter. Absence from Syglum does not prove absence of activity. Firm-issued materials may reflect the firm’s own characterization of an event, and we label them accordingly. Users should treat the product as a carefully curated research aid, not as a complete market census.

Additional detail for research teams

Syglum is not a substitute for proprietary firm intelligence or paid legal research services. It is designed to give partners and competitive-intelligence professionals a shared, evidence-based starting point for understanding the moves of the 30 covered firms. We do not publish the underlying monitoring list, crawl schedule, scoring formulas, or QA internals. What we do publish is our epistemology: the rules above, applied consistently, with the source trail visible on every record.

If you are evaluating Syglum for your firm and would like a walkthrough of the coverage model, please reach out.

Methodology updated: September 2026. Questions about this methodology or a specific record can be sent to methodology@syglum.com.