Market Valuation

Three playbooks: trading comps, precedent transactions, IPO valuation and pricing. All three run the same pipeline — screen a set, build multiples, take percentiles, adjust for how the subject differs, apply to the subject's metrics — so that pipeline is given once below and each playbook covers only what is distinctive. The method in SKILL.md applies throughout and is not repeated here.

---

The multiples pipeline

Sourcing rule, before anything else. Every peer, transaction, multiple, share price and premium in this family is either supplied by the user, retrieved from a source you can name, or absent. Do not populate a comps or transaction table from recall. Specific multiples and deal terms held in model memory are unreliable and stale in a way that is invisible in the output — a table of plausible-looking fabricated multiples is the single worst failure mode in this family, because it survives review. Where the set is incomplete, deliver the framework with the rows you can fill, label the gaps, and say what the user needs to pull.

Screen, then defend the screen. State inclusion criteria — business model rather than sector code, size band, geography, growth profile, margin profile — and then the exclusions: names that pass the obvious screen but are not comparable, each with a reason. The exclusion list is where the judgement is, so it goes in the output rather than being done silently.

Build multiples consistently. Enterprise value = market cap (on diluted shares, treasury-method) + total debt + minorities + preferred − cash. Numerator and denominator must match: EV against revenue, EBITDA and EBIT; equity value against net income. Use the same period basis across the whole set — LTM against LTM, forward against forward — and say which. Negative or near-zero denominators make a multiple meaningless; suppress the cell rather than printing a large number.

Percentile statistics. For each multiple, report 25th percentile, median, 75th percentile, mean and sample size, with the subject alongside:

Metric	25th	Median	75th	Mean	n	Subject

Median is the anchor; the mean is shown only so a skew is visible. Name any outlier distorting the set and show the median with and without it.

Premium or discount to the set. The subject is not the median. For each factor state a direction and a magnitude, with the comparison that supports it: growth against median growth, margin against median margin, scale, and business quality — revenue recurrence, customer concentration, switching costs. Aggregate to a single stated premium or discount and apply it. Do not skip to a conclusion multiple without showing this step.

Implied range. Apply the 25th and 75th percentile multiples to the subject's metrics for enterprise value, then bridge: EV − net debt − minorities − preferred

• non-operating assets = equity value, ÷ diluted shares = per share. Present low, mid and high for each multiple used — the football field.

Reliability disclosure, always. Report the true comparable count, not the row count. State whether the set is thin, whether any sub-segment has n < 5, and how old the most relevant data point is. A median of four loosely-similar names is a number with a false air of precision, and the disclosure is what stops it being read as one.

---

Trading Comps

What the public market pays today for businesses like this one.

When: "comps", "comparable companies", "trading multiples", "what's it worth against peers", "EV/EBITDA", "is it cheap".

Phases:

1. Peer selection methodology. The screen and the exclusions, per the pipeline. Defend every company in and every company out.
2. Comps table. One row per peer: company, ticker, EV, LTM revenue, LTM EBITDA, EV/revenue, EV/EBITDA, EV/EBIT, NTM P/E, revenue growth, EBITDA margin.
3. Statistical summary. The percentile table, run on LTM and NTM EV/EBITDA, LTM EV/revenue, NTM P/E, revenue growth and EBITDA margin.
4. Premium / discount analysis. Per the pipeline, factor by factor.
5. Implied valuation range. Football field output across the multiples used.
6. Quality disclosure. Per the pipeline.

Required tables: comps table; percentile summary; premium/discount build; implied range by multiple.

Challenge test: which single peer, removed, most moves the median — and whether the subject's premium or discount survives the peer set being one name thinner.

Inputs to gather: subject name and description; subject LTM revenue, EBITDA, EBIT and net income; NTM estimates if available; diluted shares and net debt; industry sub-sector; known direct competitors; any known premium or discount factor such as family control, dual-class shares or regulatory overhang.

---

Precedent Transactions

What acquirers have historically paid for control. Historical clearing prices are evidence of what buyers did pay, not of what this asset is worth today — carry that distinction into the output rather than treating the median as a target.

When: "precedents", "transaction comps", "what have deals cleared at", "control premium", "what would a buyer pay".

Phases:

1. Screening criteria. Industry definition and how narrow; date range with the reason for the cutoff; deal size band; public against private targets; and the exclusions — distressed sales, minority stakes, unusual structures — each with a reason. Split strategic from financial buyers at the screen, not afterwards; they price differently and a blended median hides it.
2. Transaction table. Date, target, acquirer, deal size, EV/EBITDA, EV/revenue, premium to the 30-day unaffected VWAP, buyer type, stated rationale.
3. Statistical summary, split by buyer type. Percentiles and sample size in three columns — strategic, financial, combined. Where either sub-sample is under five, say so in the table rather than in a note beneath it.
4. Premium analysis. Median and mean control premium against the unaffected price, then premium by deal size, by buyer type, and by point in the cycle.
5. Market context adjustment. Have multiples in this sector expanded or contracted since these deals, and what drove it — rates, credit availability, sentiment, regulation. Apply an explicit haircut or uplift to the historical median and justify the size of it. An unadjusted median from a different rate environment is the most common error in this analysis.
6. Implied range. Applied separately at strategic and financial buyer multiples, since the two produce different answers and the user needs both.
7. Reliability disclosure. Age of the most comparable transaction, dataset size, and any segment where n < 5.

Required tables: transaction table; percentile summary split by buyer type; premium analysis; implied range at strategic and financial multiples.

Challenge test: how the implied range moves if the date range is halved — and whether the conclusion depends on transactions struck in a materially different rate environment.

Inputs to gather: subject name and industry; subject LTM revenue and EBITDA; the plausible buyer universe, both strategic names and sponsors active in the sector; any known approach, process or rumour; current credit availability and M&A conditions; the reason for the transaction — sale, restructuring, or IPO alternative.

---

IPO Valuation and Pricing

Where the deal should price. Pricing is a negotiation between an issuer maximising proceeds and investors requiring upside to participate, so model both sides rather than producing a single fair value.

When: "IPO", "listing", "how should we price it", "float", "what's the range", "is the deal priced to go".

Phases:

1. Offering structure. Primary shares raising new capital against secondary shares taking proceeds out, and who is selling; total offered against total post-IPO shares to give the free float; the greenshoe, conventionally up to 15% of base deal size; specific use of proceeds — debt paydown, growth capex, working capital — rather than a general statement; and the insider lock-up. Then pre-money equity value, post-money (pre-money plus primary proceeds), and dilution to existing holders as primary shares over post-IPO shares.
2. Comparable IPO analysis. Sector listings over roughly the past 24 months: company, date, IPO price, first-day close, 30-day and 6-month return, EV/revenue and EV/EBITDA at IPO, pre-IPO growth. This table is subject to the sourcing rule above and is frequently the one that gets fabricated — leave it empty and ask rather than filling it from recall. Conclude with sector IPO sentiment, hot, neutral or cold, and the evidence for the call.
3. Valuation range from at least three methods. Trading comps with a stated IPO discount applied and the discount justified; comparable IPO multiples from the table above; a DCF, which carries weight on the roadshow and little in the book; and a sector-specific metric where one governs. Reconcile them into low, mid and high per share with the implied multiple at each point, and explain any method you are down-weighting.
4. Buy-side perspective. At the midpoint, what return does a long-only investor need to believe is available, and what growth and margin path delivers it. State whether that path is consistent with the company's own history. Then the first-day pop: the sector median first-day return implies the deal is being priced at a stated discount to fair value, which is the cost of a covered book.
5. Pricing recommendation. A range, an expected midpoint, one paragraph of rationale, and the top three risks to pricing.

Required tables: offering structure and dilution; comparable IPO table; valuation range by method with implied multiples.

Challenge test: the price at which the buy-side return case stops working, and how far that sits above the proposed midpoint. If that gap is thin, the deal is priced to break.

Inputs to gather: company name and description; LTM revenue and growth; LTM EBITDA or the path to profitability; pre-IPO shares outstanding; target primary raise; secondary shares with seller and size; comparable public companies; recent sector IPOs; use of proceeds.