Skip to main content
L3i
L3i Data

Financial statement analysis, traced to the filing.

L3i Data turns SEC EDGAR filings for roughly 8,300 public issuers into structured financials, ratios, and comparables — and lets you query all of it in plain language, with every answer cited back to the source filing.

~8,300
Issuers with structured financials
2009
Structured XBRL financials, tagged and normalized
65+
Structured, extracted & derived datasets

Coverage is stated precisely, not rounded up. Structured financials begin with XBRL tagging in 2009, with full-text keyword search across the filing archive. Filing types include 10-K, 10-Q, 8-K, proxies, Form 4, 13F, 13D/G, and S-1 — enriched with FRED macro series and Census business-formation data so a company reads against its market.

One foundation

Ingest once. Analyze every way.

The filings are normalized a single time, then reused across every view — so a ratio, a trend, a peer set, and an AI answer all resolve to the same audited source. No re-keying, no drift.

SEC EDGAR filing
normalized once
Normalized statementsRatios & trendsCommon-size analysisPeer comparablesFlux / varianceCited answers
What you can do

Read the statements, build the comparables, ask the question.

The financial-statement layer is normalized once and reused everywhere — so a ratio, a peer set, and an AI answer all resolve to the same audited source.

01

Structured financials & ratios

Normalized income statement, balance sheet, and cash flow for ~8,300 issuers, with derived ratios computed consistently across companies and periods — no re-keying from PDFs.

02

Comparable-company screening

Build peer sets from disclosed financials and generate interquartile ranges on the margins and returns that matter, with each comparable tied to the filing it came from.

03

Ask in plain language

Pose a question the way you’d ask a colleague. L3i returns a direct answer grounded in the filings, with citations to the exact passage — not a general-purpose guess.

04

Filing browser & full-text search

Move from a screen to the underlying document in one step: 10-K, 10-Q, 8-K, proxy, Form 4, 13F, 13D/G, and S-1, all full-text searchable.

Comparables

From one company to a whole industry.

Roll a peer set into a single size-weighted industry markup — the pricing-power lens behind the De Loecker / Roosevelt Institute studies — and track it across years, so one firm’s cost shock doesn’t distort the read.

Industry markupTechnology hardware · FY2023
FirmRevenueMarkup
Apple$383B60.3%
Dell$88B6.5%
HP Inc.$54B7.5%
Garmin$5B30.0%
Logitech$4B10.0%
Size-weighted industry markup45.4%

Why size-weight?

Weighting each firm by its share of industry sales reflects real pricing power — a few dominant firms move the aggregate. The same five-firm set reads very differently three ways:

Median10.0%
Equal-weighted22.9%
Size-weighted45.4%

Tracked over years

Size-weighted markup by fiscal year

Illustrative peer set on the COGS + SG&A basis; each markup traces to the firm’s 10-K. Apple’s 60.3% is computed from its FY2023 filing.

Arm’s-length range

A defensible range, with outliers flagged.

The interquartile range is the standard way to set an arm’s-length band — it keeps the central distribution and drops the extremes. L3i builds the range from your comparables and flags anything beyond the Tukey fence, so a single outlier can’t quietly stretch your position.

Arm’s-length rangeOperating profit markup · 9 comparables
050%100%150%200%250%Arm’s-length rangeMedian 67.4%Q1 22.7%Q3 89.4%Tukey fence 189.5%Meta 233.2% · flagged
Q122.7%
Median67.4%
Q389.4%
IQR66.7%
Flagged out1

Illustrative comparable set on the COGS + SG&A basis. The same range can be built on asset-adjusted bases — operating ROA or the Berry ratio — for cross-industry comparability.

Three ways in

Same data foundation, whichever way you work.

Web app

Browse & screen

A company and filing browser with screens and comparable sets, built for analysts who want to move fast without writing code.

REST API

Pull it into your stack

Query normalized financials and citations programmatically and drop them straight into your models and pipelines.

Ask AI

Query in plain language

A natural-language layer over the same corpus that answers with source-traced citations — so the reasoning is auditable, not opaque.

Built to be defended

The answer is only useful if you can stand behind it.

Every figure is traced to the filing it came from and carries a reliability verdict — so the reasoning is auditable, not a black box.

CLEANVerified against source. The value ties cleanly to a tagged filing passage and passes consistency checks.
WATCHUsable, with a note. Minor normalization or restatement nuance you should read before relying on it.
FLAGGEDNeeds a human. The disclosure is ambiguous or inconsistent — surfaced for a practitioner to resolve, not silently filled in.

Source-traced by design

Every number links back to the filing it came from.
Operating margin · FY202329.8%
Return on assets · FY202327.5%
Current ratio · FY20230.99
Traced to:AAPL 10-K FY2023 · Item 8

See your own coverage, on your own tickers.

Walk through L3i Data with your companies loaded, and see the citations resolve to the filings you already know.