◆ Contrarian Corpus
Empirical view

The anatomy of an activist deck

Every chart below is computed from 34,166 slides across 848 decks from 38 firms. No hand-curation; no anecdote. Each section ends with the one thing it tells us.

Decks catalogued 848 38 firms · 2005–2026
Slides classified 34,166 17.4% contain a chart
Callouts extracted 26,677 search callouts →
Notable quotes 10,384 search quotes →
What the corpus tells us

Four findings that actually change how you'd write one.

Only 17.4% of slides actually have a chart.

The dominant slide is text-driven: a headline, a paragraph or bullet list, a boxed callout, and a footnote. Charts are the punctuation, not the sentence.

89.8% of all charts are bars or lines.

Pie and donut charts combined are under 5.0%. The pro preference is for charts that let readers compare magnitudes directly — which pies famously don't.

45.5% of slides carry a footnote.

Sourcing isn't optional. Almost every claim is anchored to a public document (earnings call, filing, press release). This is what separates a short-seller memo from a blog post.

36.1% of slides have a callout box.

The callout is where the author tells you what to think: the one sentence they don't want you to miss. In the corpus there are ~25k callouts — the searchable phrase library.

What's actually on a slide

Frequency of each element across all classified slides. A slide is a composition — usually headline + paragraph + callout + footnote.

Headlines and paragraphs on nearly every slide; charts on only ~17%.

Takeaway: The canonical slide is text. Headlines appear on 68.5% of slides; footnotes on 45.5% (pros source almost everything); callout boxes on 36.1%. Data tables are on 17.1% — tables remain the pro's weapon of choice when the story is "read these numbers yourself".

▸ all elements (36) — click any to drill down

The chart vocabulary

When activists do put a chart on a slide, what kind? (7,099 total charts.)

Bars and lines make up ~90% of all charts; pies and donuts under 5%.

bar charts 3,995

the workhorse — peer benchmarks, time series, comparisons

line charts 2,379

price history, trend breaks, margin erosion over time

waterfall 269

sum-of-parts reveals — the classic activist 'value bridge'

pies / donuts 352

revenue / segment mix; rarely used, quickly replaced by tables

other 104

org charts, flow diagrams, bubble plots, radar — niche uses

Takeaway: 89.8% of all charts are bars or lines. Pies and donuts together account for 5.0%. The bias is toward chart types where the human eye can compare magnitudes directly — which is precisely what pie charts fail at.

▸ all 16 chart types — click to see the slides

Slide types

What role each slide plays in the argument. Top 20 types.

Villain critique, CEO-quote, peer-gap and demand-list dominate the taxonomy.

▸ all 38 slide types — hover for description

Narrative function

What each slide is doing rhetorically.

Expose contradiction is the single most common narrative function.

Takeaway: The single most common narrative move is expose contradiction — management said X, reality shows Y. Together, expose contradiction + diagnose problem + name villain account for a huge share of slide-time. Activists don't argue; they expose.

▸ all 20 narrative functions — hover for description

The wording of activism

Words grouped by what they do rhetorically. Counts across all 37,061 callouts and quotes combined. Brand names and filler omitted.

Takeaway: Activist writing rides on three pillars — an epistemic hedge ("we believe", "appears to"), a comparative frame ("below peers", "despite"), and a named antagonist (CEO, board, directors). The nouns do the pointing; the verbs do the framing.

What gets measured

The 15 most-cited primary metrics across the corpus.

EBITDA, revenue and margin lead; unit-level metrics are rarer.

Takeaway: Share price is #1, but the meaningful action is in quality-of-earnings metrics — EBITDA, margins, TSR, free cash flow. These are the levers activists argue can be moved. Revenue and sales are mentioned less often: top-line doesn't create alpha, conversion does.

Signature moves by firm

For the 8 most prolific firms, % of that firm's slides that are each type. Normalized per-100 so bigger libraries don't dominate.

Each fund's deck mix is a rhetorical fingerprint, normalized per 100 decks.

Slide density

How crowded is the average slide?

Balanced layouts dominate; overcrowded slides are rare.

Deck length

How many slides activists pack into a deck.

Most decks fall between 25 and 70 slides; a long tail extends past 100.

Most prolific firms

Deck count by firm.

Pershing, Muddy Waters, Elliott and Spruce Point lead by volume.

Deep patterns

What the correlations reveal

Four cross-cuts the raw distributions hide: thesis quality, slide-type clustering, deck length vs. narrative score, and short-sellers vs. activists.

Thesis types ranked by narrative quality

Avg narrative_interest score (1–5) across decks tagged with each thesis. Sample size shown.

ThesisDecksNarrativeVisual
multiple rerating 219 4.12 2.95
strategic pivot 89 4.09 2.78
capital return 121 4.07 2.98
fraud exposure 357 4.06 2.35
breakup spinoff 134 4.04 3.11
cost cutting 58 3.98 2.88
operational turnaround 283 3.92 2.94
management change 270 3.89 2.84
undervaluation 191 3.89 2.86
governance board 413 3.85 2.76
sale of company 47 3.81 2.34
other 48 3.71 2.29

Takeaway: Activist defense decks (management side) score highest on narrative — they're rarer but more sophisticated. Fraud exposure has top narrative but the lowest visual craft — short-seller reports are utilitarian text-heavy documents that don't need to be pretty. Breakup / spinoff leads visual craft — the sum-of-parts reveal is inherently a designed-chart play.

Slide co-occurrence matrix

Of the decks containing slide type X, % that also contain Y. Pick any subset — the matrix rebuilds instantly. Covers all 16 types with ≥50 decks.

types shown
· ·

Rows = slide type (the delivery). Columns = narrative function (the objective). Each cell = % of that slide-type's slides doing that function. Darker = more concentrated.

Expose contradiction Name villain Diagnose problem Appendix Expose gap Transition Establish context Front matter Frame situation Quantify opportunity Cite precedent State demand Compare peers Decompose segments Introduce nominees Propose solution Show valuation bridge Pre-empt rebuttal Filler
Villain critique (10003) 44.0% 38.0% 15.0% 2.0%
Other / uncategorized (3319) 15.0% 36.0% 5.0% 22.0% 2.0% 5.0% 4.0% 1.0% 1.0% 2.0% 2.0% 2.0%
CEO-quote contradiction (3249) 96.0% 1.0%
Peer-gap chart (2270) 1.0% 69.0% 30.0%
Executive summary (588) 32.0% 2.0% 14.0% 47.0% 2.0% 1.0%
Demand list (1016) 87.0% 9.0%
KPI overview (1383) 3.0% 30.0% 3.0% 32.0% 2.0% 25.0% 2.0% 2.0%
Precedent transactions table (707) 99.0%
Valuation reveal (493) 4.0% 33.0% 61.0%
Sum-of-parts buildup (936) 2.0% 2.0% 20.0% 67.0% 2.0% 7.0%
Thesis headline (379) 27.0% 6.0% 47.0% 14.0% 2.0%
Timeline (460) 14.0% 5.0% 1.0% 27.0% 26.0% 6.0% 12.0% 6.0%
Process diagram (394) 14.0% 21.0% 2.0% 16.0% 10.0% 2.0% 1.0% 31.0% 1.0%
Before / after framing (321) 48.0% 6.0% 15.0% 2.0% 6.0% 9.0% 3.0% 7.0% 2.0%
Nominee biography (596) 98.0%
Comparison table (184) 23.0% 5.0% 23.0% 6.0% 3.0% 1.0% 24.0% 8.0% 8.0%

Takeaway: The matrix exposes that "slide type" and "narrative function" aren't identical — a ceo_quote slide almost always exposes contradiction (its whole purpose), but a peer_gap slide splits between expose_gap, compare_peers, and diagnose_problem. The same visual delivery can carry different rhetorical payloads.

Takeaway: Peer-gap + villain-critique co-occur in 84% of cases: the gap isn't just a chart, it's always attached to blame. Before/after framing virtually requires a peer-gap (92%) — you can't show "after" without the benchmark. The activist toolkit comes in clusters, not isolated moves.

Deck length vs. narrative quality

Contrary to "keep it short" advice — longer decks score higher on our narrative assessment.

PagesAvg narrativeDecks
1–30 3.50 479
31–60 4.20 161
61–100 4.19 124
101–200 4.60 62
201–500 4.80 15

Takeaway: In adversarial research, more evidence = more persuasive. The longest decks (200+ pages) average 4.80/5 — the top-tier narrative bracket. Short decks (<30 pages) average just 3.50/5: too little room to build the argument. The "less is more" mantra doesn't hold here.

Short-sellers vs. long activists

% of slides doing expose_contradiction — the signature move of both camps, but at very different intensities.

short-sellers (5×_) 37.3% Muddy Waters, Spruce Point, Hindenburg, Kerrisdale, Citron…
long activists (0×_ / 1×_ / 2×_) 14.7% Pershing, Elliott, Starboard, Trian, Icahn, Third Point…

Takeaway: Short-sellers expose contradictions at 2.54× the rate of long activists. Their entire output is about exposing — no propose_solution, no decompose_segments. Activists expose too, but they also have to build an alternative: turnarounds, breakups, capital-return plans. Different job, different grammar.

Two different genres, not one

Beyond the single "expose" axis, the two camps produce structurally different decks. Side-by-side on the four dimensions we grade:

activists n=495
Avg. pages37.3
Avg. narrative3.7 / 5
Avg. visual craft2.56 / 5
Tone (top)adversarial · 44.6%
short-sellers n=353
Avg. pages47.8
Avg. narrative4.02 / 5
Avg. visual craft2.2 / 5
Tone (top)adversarial · 90.7%

Takeaway: Short-sellers write 28% longer decks (47.8 vs 37.3 pages), score slightly higher on narrative, and are 91% adversarial in tone — versus 45% for activists. Activists lean on visual craft (lower narrative, higher craft); shorts lean on prose and sourcing. The shape of the deck signals which genre you're reading before you've seen a single number.

Activism, decade by decade

From a handful of campaigns to an industry

Public activist + short-seller decks in the corpus, by presentation year. The corpus draws from fund archives, so recent years are better covered — but even accounting for that, the shape is unmistakable.

Flat 5–10 decks per year through 2010, climbing to ~125 in 2024–2025.

Takeaway: The 2023-2025 window alone carries 376 decks — more than all 2005-2009 combined (16). Activism was a niche craft for 15 years and then became a normalized capital-markets genre. The grammar of the argument — expose, diagnose, demand — is now industry standard.

Different sectors, different arguments

What gets argued, by sector

Each row = one industry. Columns = the thesis types most commonly applied to it. Darker = larger share of that sector's pitches. Each cell links to browse the decks.

Heatmap of sector by thesis type. Rows are industries; columns are the thesis categories most commonly applied to each. Each cell shows the percentage of that sector's decks leaning on that thesis, and darker cells indicate higher share.
governance board fraud exposure operational turnaround management change multiple rerating undervaluation breakup spinoff capital return
consumer (443) 15% 16% 14% 9% 11% 9% 8% 5%
tech (359) 19% 16% 14% 15% 11% 5% 7%
energy (291) 24% 4% 16% 17% 4% 6% 16% 5%
industrials (261) 22% 16% 13% 15% 11% 8% 3% 3%
healthcare (188) 12% 27% 10% 10% 15% 8%
communications (164) 18% 18% 12% 10% 7% 8% 4% 8%
materials (162) 20% 15% 13% 12% 8% 8% 7% 6%
real estate (131) 19% 18% 4% 12% 12% 16% 3%
diversified (120) 11% 10% 16% 10% 8% 19% 7% 10%

Takeaway: Sector determines the rhetorical template. Energy and industrials lean hard on governance/management change — incumbent CEOs defending against a board shakeup. Tech and healthcare tilt toward fraud exposure and multiple rerating — the upside story is questioning the narrative. Consumer is the most contested arena, with every thesis type represented at meaningful share.

Valuation grammar

How a thesis puts a price on its claim

% of decks that explicitly show each valuation framework on a slide. Decks can use multiple — the number to watch is "not shown".

Multiple comparison leads; 28% of decks show no valuation framework at all.

Takeaway: 64% reach for multiple comparison — the workhorse of activist pricing. But 28% of decks don't show a valuation framework at all. That's not negligence: for a fraud exposé or a governance pitch, the argument doesn't price-target the upside, it questions whether the current price is honest. A "not-shown" valuation isn't missing — it's a signal about what kind of thesis you're reading.

Who's who

Brief glossary of the main voices

Pershing Square — Bill Ackman's fund; famously architected the Valeant short thesis and the Herbalife campaign. Most methodology-heavy decks in the corpus.
Elliott Management — Paul Singer's multi-strategy fund; most prolific activist by campaign count. Signature: exhaustive operational critique.
Muddy Waters — Carson Block's short-selling firm. Known for on-the-ground fraud investigations, especially in Asia.
Spruce Point Capital — NY short-seller; its research notes are branded "Spruce Point believes…" — the signature phrase you'll see recur in callouts.
Hindenburg Research — Nate Anderson's short-selling firm. Authors of the Adani and Nikola reports.
Starboard / Trian / Third Point — classic US activist funds; Starboard (Jeff Smith) is heavy on operational turnaround decks; Trian (Nelson Peltz) specialises in governance; Third Point (Dan Loeb) in sharper rhetoric.