NeuraCap

Methodology

How a figure gets verified

A number is only useful if you can say where it came from. This page sets out how NeuraCap builds a figure, what it does when the evidence is thin, and what it refuses to publish. The same method sits behind the free industry research and behind every company report.

The rule

Two independent sources, or it is not a figure.

Every figure that reaches a report is checked against a second, independent source before it is used. Across the data set, 97.5% of figures are confirmed that way. We publish the rate rather than rounding it to 100%, because the gap is the part you need to know about.
97.5%of figures confirmed against two independent sources
  • What counts as independent

    Two readings of the same press release are one source. A figure taken from a filing and matched against a transcript, a consensus estimate or a market data feed is two. Independence is about the path a number travelled, not how many places it appears.

  • What the 2.5% means

    Roughly one figure in forty cannot be matched. Usually that is a disclosure gap rather than a contradiction: the company does not break the line out, or only one party to a transaction has said anything on the record.

  • Where you can check

    Every report ships with a sources page naming the classes of evidence behind the analysis. The seven classes, and what each one contributes, are set out on coverage and data.

Plausibility gates

The multiples you will not see.

A comparable set is only as good as the outliers someone had the discipline to remove.

Every multiple passes through a plausibility gate before it reaches a chart. A multiple that fails is excluded. It is not footnoted, greyed out or plotted at the edge of the axis with an asterisk. It is not plotted at all.

The reason is arithmetic rather than taste. A multiple is a ratio, and a ratio stops carrying information as its denominator approaches zero: a company sitting at break-even produces a number in the hundreds that says nothing about how the market values the business. Leave three of those in a peer set and the range widens by tens of percent, the median drifts, and the midpoint — the figure people actually quote — becomes an artefact of the outliers.

This matters most in the football-field chart, where several methods sit side by side and the eye reads the overlap between them as the answer. A range built on implausible inputs is worse than no range, because it looks like evidence.

When a multiple stops carrying information

  • The denominator is negative or close to zero, so small changes in earnings swing the multiple wildly.
  • A one-off item — a disposal gain, an impairment, a settlement — is doing the work that operating performance is supposed to do.
  • Deal terms were never fully disclosed, so the implied multiple is a reconstruction rather than a fact.
  • The enterprise value and the earnings beneath it describe different periods, so the ratio compares two different companies.

Calibrated language

Why a report says NeuraCap reads, not because.

Financial writing has an old habit of turning a correlation into a cause, because causes read better. NeuraCap does not. Where a report interprets, it says so in the first person, so the interpretation can be separated from the evidence and argued with.

The convention applies wherever the platform writes prose: report narratives, sector commentary, and the answers the Financial Assistant gives when you ask it a question about a company. Nothing is stated more confidently than the evidence supports.

It also holds the line that matters most to a reader who has to defend the analysis downstream. Sourced data, calculated metrics and interpretation are three different things, and a report that blurs them is impossible to check. Coverage and data sets out how the three are kept apart.

Written this way

“Multiples across the peer set compressed through the period. NeuraCap reads the compression as the market pricing in a slower renewal cycle.”

The observation and the reading are separate sentences. You can accept the first and reject the second.

Not written this way

“Multiples compressed because renewals slowed.”

One clause, two claims, no attribution. The data can show both facts and still not establish the link between them.

Valuation basis

The basis is chosen before the answer, and stated.

Not every sector should be valued on the same line of the income statement. NeuraCap states a primary basis for the sector, and carries EV/Revenue alongside it as a cross-check.

Primary basis, stated up front

A basis chosen after the answer is known is not a basis, it is a justification. The report names the primary basis for the sector before it shows a number, so you can disagree with the framing rather than only with the output.

EV/Revenue as the cross-check

Revenue is the line least affected by accounting policy, capital structure and one-off items. That makes it a weak primary answer for an established business and a strong check on one, which is exactly how it is used.

When the two disagree

A gap between the earnings-based range and the revenue-based range is a finding, not a rounding problem. It usually points at margin, mix, or a period that is not representative — and the report says which, rather than averaging the two.

How the methods are assembled into a single range is set out in valuation analysis, and the transaction record behind the precedent method is described in M&A intelligence.

Exclusions

What gets left out, and where we say so.

Excluding evidence is only defensible if the exclusion is visible. Each of these is recorded rather than silently dropped.
  • Multiples that fail a plausibility gate

    Excluded from the peer set and from the chart. A ratio built on a denominator close to zero is arithmetically valid and analytically empty; publishing it would widen a range without adding evidence.

  • Figures with no second source

    The confirmation rate is 97.5%, which means roughly one figure in forty cannot be matched. Those figures are treated as unconfirmed rather than quietly promoted into the analysis.

  • Disclosure gaps

    Where a company does not break out a line, or where the parties to a transaction never disclosed terms, the figure does not exist. NeuraCap records the gap instead of estimating one into existence.

  • Periods that do not line up

    A multiple only means something when the numerator and the denominator describe the same period. Where they cannot be aligned, the comparison is left out rather than presented as like for like.

Every report ends with a sources page naming the classes of evidence the analysis drew on, so a reader who wants to go back to the filing can. You can see the convention in practice in any of the free published industry reports.

Limits

What this is not.

Being explicit about the limits is part of the method, not a disclaimer bolted on at the end.
  • Not a recommendation

    Nothing on this site or in a report is a recommendation to buy, sell or hold a security.

  • Not an offer

    Nothing here is an offer, or a solicitation of an offer, to buy or sell any financial instrument.

  • Not an appraisal

    A NeuraCap valuation range is analysis. It is not a formal appraisal of a business or an asset.

  • Not a fairness opinion

    No report constitutes a fairness opinion, and none should be presented to a board as one.

No fiduciary relationship is created by reading a report or using the platform. The analysis is evidence for your decision; the decision, and the responsibility for it, stay with you. The full wording is in the disclaimer and the terms of use.

Questions

Methodology, answered plainly.

How does NeuraCap verify a financial figure?
Each figure is checked against two independent sources before it is used in analysis. Across the data set, 97.5% of figures are confirmed that way. Where a second source cannot be found, the figure is not presented as a settled fact.
Why are some multiples missing from a comparables chart?
Because they failed a plausibility gate. A multiple that no longer carries information — most often because its denominator is near zero, negative or distorted by a one-off item — is excluded rather than plotted at the edge of the axis, where it would widen the range and move the midpoint.
What does it mean when a report says NeuraCap reads something?
It is calibrated language. A reading is attributed and can be argued with; a causal claim asserts something the data cannot establish. Reports state what the evidence shows, then say how NeuraCap reads it, and keep the two apart.
Which valuation basis does NeuraCap use?
A primary basis is stated for the sector before any number is shown, and EV/Revenue is carried alongside as a cross-check. Where the earnings-based range and the revenue-based range disagree, the disagreement is reported rather than averaged away.
Is a NeuraCap report investment advice?
No. Nothing NeuraCap publishes is a recommendation to buy or sell a security, an offer or a solicitation of an offer, an appraisal or a fairness opinion, and no fiduciary relationship is created by reading it.
Can I see the sources behind a report?
Yes. Every report carries a sources page setting out the classes of evidence behind the analysis, so a figure can be traced back to the kind of document it came from.

Keep reading

Where the method shows up.

Sources

Coverage and data

The seven source classes, what each contributes, how often it refreshes, and how sourced data is kept separate from calculated metrics and interpretation.

Platform

Valuation analysis

How the methods are run and assembled into a football-field range, with the plausibility gates applied before anything is plotted.

Research

Sector coverage

Sector research read on a repeating cycle, with the valuation spread and the transaction record set out the same way every time.

Get your custom-built strategic insights report today.

Public or private, any sector. Tell us the company and what you need to understand.