Management Presentation Review

SkillMonitoring & ops

Reviews a management presentation for what it omits -- moved metric definitions, cohorts that stopped being shown, bridges that do not tie -- when you need the questions to ask in the session itself.

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the Management Presentation Review skill

What this skill tells your AI

The instructions your AI receives, as published by andreworia/claude-finance-skills in packs/private-equity/skills/management-presentation-review/SKILL.md and read by ahel’s review.

When to use

Use this skill on receipt of the management presentation and before the session is held, and again when a second-round deck must be diffed against the first. The deck is a sales document from the adviser who wrote the CIM: the disclosed numbers have been vendor-checked, so the signal is in what was dropped.

What it does

Produces a review organised around omission: a page-by-page diff against the CIM, a register of metric definitions that moved, the cohorts that stopped being shown, a bridge tie-out, a realised-versus-annualised EBITDA restatement, and a three-tier question list for the session, the data room and what to hold back.

Method

Step 1 -- Diff the deck against the CIM, page by page

Log every addition, deletion, re-cut chart and changed axis. Two tellings of the same story differ only where someone decided they should, so treat each difference as intentional and ask what it is doing.

Step 2 -- Chase every metric definition

Recurring revenue, active customer, adjusted EBITDA, retention (gross or net, logo or revenue): request the written definition and the metric restated on last year's definition. A definition that moved between the CIM and the deck deserves more diligence time than any number that was disclosed.

Step 3 -- Find the cohort that stopped being shown

Look for a cohort chart truncated to three years where the CIM showed five, a segment folded into "other", a churn table now shown only in aggregate. Ask for the missing period in the original format, not a new one.

Step 4 -- Tie the bridges

The revenue and EBITDA bridges must sum to the stated totals and reconcile to audited or QoE figures, with "other" small and run-rate items labelled. A bridge that does not tie is the opening question of the session, not a footnote for a workstream.

Step 5 -- Separate realised from annualised

Cost programmes shown at run-rate, price rises effective mid-year, contracts signed but not live: restate EBITDA on realised figures only, then recompute the multiple the indicative offer implies on that base. The gap between the two multiples is the negotiating position.

Step 6 -- Test the pipeline slide against history

Compare weighted pipeline with actual bookings for the last four quarters. If the conversion the plan requires exceeds anything the business has achieved, the growth case is an assumption and should be underwritten as one.

Step 7 -- Write the questions in three tiers, then run the session

Tier one: asked live, where the answer is a reaction rather than a number. Tier two: data-room requests, where the answer is a number. Tier three: held back, because the question reveals a kill criterion or tips a hypothesis early. In the session, record who answers what -- when the chief executive takes a question addressed to the finance director, note it.

Inputs

  • The management presentation, and any prior version of it
  • The CIM and the teardown memo
  • Audited accounts or the QoE draft, where available
  • The vendor's stated KPI definitions
  • Issues log, outstanding DDQ items, session agenda

Output format

  • CIM-to-deck diff, page by page
  • Definition register: metric, definition used, prior definition, effect on the number
  • Omissions list, each with the disclosure requested and the format required
  • Bridge tie-out, shown in prose, never as a table
  • Realised versus annualised EBITDA, with the entry multiple recomputed on both
  • Three-tier question list, each question with an owner
  • Post-session note recording who answered what

Example

Fictional worked example -- Calder Diagnostics.

The deck bridges FY23 reported EBITDA of GBP 16.0M to FY24 adjusted EBITDA of GBP 21.4M: volume +3.2, price +1.8, cost programme +2.1, investment -1.7. The bridge ties. Two omissions do not: the CIM showed FY23 reported EBITDA of GBP 15.2M, and the GBP 0.8M difference is an unfootnoted reclassification of capitalised development cost; and the GBP 2.1M cost programme is four months of actual savings of GBP 0.7M annualised, so GBP 1.4M is run-rate rather than realised.

On realised EBITDA of GBP 20.0M, the indicative 9.5x offer of GBP 203.3M is 10.2x. Tier-one question, asked live: which four months, and what has the monthly run-rate been since they ended?

Signals

GitHub stars
20
Forks
4
Last commit
Sep 2026
Advanced
Item type
skill
Key
management-presentation-review
Source
github.com/andreworia/claude-finance-skills