Front door for the teardown. Full reasoning and sourcing in consolidated-brief.md; per-area evidence in findings/.

The one-sentence version

Ladder sells a coach-led strength app for $29.99/mo, has ~400k members and roughly $100M ARR — and it gets there by spending 56 cents of every revenue dollar on marketing while spending 7 cents on the coaches and studio everyone assumes are the product.

What it feels like to use

You pick a coach. Every Sunday night, six workouts appear for the week. Each one is a named, branded session — "PEACH PALACE", "PYTHON PUMP" — with a coach talking in your ear, video demonstrations, and a timer. Your teammates, who are all doing the same workout that week, can send animated emoji with sound onto your screen while you are mid-set. On Sunday at 6pm, the week expires and a new one appears.

It feels like having a coach.

What it actually is

A broadcast. Everyone on your team gets the identical workout. The onboarding quiz — 27 screens on the web — decides which team you land on and which marketing segment you join, and then touches nothing else. There is no auto-regulation: the app tells you the sets, reps and effort, and you decide the weight.

The coach in your ear is a recording on a fixed timeline. Swap an exercise mid-workout and the video changes, but the coach carries on cueing the movement you just replaced.

None of this is a flaw. It is a very well-executed trick: it buys the feeling of a personal plan at the cost of a database lookup, and that gap is where all the margin lives.

How the economics work

A coach writes about 12 workouts per 8-week cycle. Alternating A and B weeks stretch those 12 into six weeks of "new" content, and the cycle repeats. Across 22 teams that is roughly 1,700 authored sessions a year serving nearly 6,900 delivered workout-days — a 4:1 leverage ratio.

The exercise demo videos are not a treadmill either. Each team gets one shoot day of ~400 clips, then that library serves for two to three years. One team's November 2023 batch was still in use in August 2026, and nobody noticed.

What genuinely recurs is the coach's voice: about 30 short audio clips per session, every session, in a specific named person's voice. That is the real content cost — and the real moat, because the value is that a real person people follow said it.

Where the money actually goes

Not into content. Into buying customers. Ladder's $90M from General Catalyst is not equity — it is a marketing-spend facility that covers about 80% of monthly sales and marketing and is repaid out of the revenue from the customers it buys. Actual equity raised is about $37.5M, and there has been no new equity in 21 months.

The 20-odd coaches are best understood as an acquisition channel wearing a content costume. You are not hiring a programmer; you are renting an audience.

Why the rating is 4.95

Because 80% of Ladder's members were not using a fitness app before Ladder, and half had never used one at all. It is not beating rivals in a head-to-head — it is converting people who were not in the category. The complaints in the low reviews are almost entirely about bugs, billing and customer support. Practically nobody criticises the training.

Where it is open

• It cannot hear you. No voice input anywhere, no logging from the Apple Watch. You log sets by opening a sheet during your rest break. A paying customer asks for exactly this in a review.
• The coach cannot respond. Pre-recorded audio on a fixed timeline. A genuinely conversational coach is a category Ladder cannot enter without breaking the 4:1 leverage its economics depend on.
• Barbell basics are structurally diluted. Only 10 of 22 teams even list a barbell in their equipment; more than half are dumbbell-or-lighter. Across 154 session slots in the whole catalogue, not one is named after a squat, bench, deadlift or press — they are named "Glutes", "Arms x Abs", "Hourglass Upper Body". Ladder's own data has a flag for whether a movement is an essential one. They know; they choose otherwise.
• Self-inflicted friction. Switching teams is capped at twice a month unless you pay annually. Coach messaging moved behind a paid upgrade in 2026. Cancelling requires emailing support three business days ahead.
• Six years in, still English-only and iPhone-only. Android and corporate wellness were both promised in the 2024 raise and neither has shipped.

The honest caveats

• The App Store screenshots are from ~2024 and do not show the current app. Ilia's own hands-on notes are the freshest evidence in this file.
• Team count is either 22 or 27 depending on which API read you trust. Both were taken.
• One of Ladder's four advertised awards is inflated: Women's Health gave them "Best Strength Training App", not "Best Overall".
• Research ran out of web-search budget partway through, and one researcher was cut off by an account limit after filing. Several questions are parked rather than answered, and the brief says which.

Synthesis of six Area Findings. Every claim below traces to a Finding; Finding tags are [A1]–[A6] and the underlying source tags live in each Finding file.

1. Bottom line

Ladder is a performance-marketing company that ships a broadcast fitness product, and it is very good at both. Almost everything that reads as personalisation is theatre, and the theatre is deliberate — it is what makes a cheap, highly-leveraged content operation feel like a coach who knows you.

Three numbers carry the whole story:

The company reports >300k paying members and "approaching $100M ARR" (Jan 2026), and 400k members, 80% women (Apr 2026) [A6].

2. What Ladder actually is

The product is a weekly broadcast, not a plan

• A Team = one named coach + one named program. 22 live [A4].
• A coach authors a 6-week series as alternating A/B week templates (~12 authored sessions), then a 2-week "Open Season" deload [A2]. Six workouts drop Sunday night.
• The plan expires every Sunday 6pm ET [A5]. Returning is structural, not encouraged.
• There is no auto-regulation anywhere. The app prescribes movement, sets, reps, rest and Effort%; the user picks the load, guided only by a Previous column [A2].
• Everyone on your team gets the identical workout. The cohort is the point — it is what makes Cheers possible (below).

The personalisation is segmentation

The onboarding quiz runs 95 live channel×team variants, 27 screens on web, 15 in-app [A3]. Its answers drive team assignment and Braze segmentation only. They do not touch the training [A3]. Combined with the absence of auto-regulation [A2], this closes the loop: Ladder asks a great many questions and then hands you a broadcast.

This is not a criticism. It is the trick. It buys the feeling of a bespoke plan at the cost of a lookup.

The coaching is pre-recorded and timeline-bound

The single most important mechanical finding in the Inquiry:

Swap an exercise mid-workout and the video changes, but the coach keeps cueing the original movement [A2].

The audio is a content asset, not a system — roughly 30 coach-voiced clips per session, each attributed to a named coach in an audioSpeakerUser field [A4]. It cannot respond to you. It cannot know what you did. Structurally, it never will without re-architecture.

The retention machine is duller and better than the gamification

Ranked by evidenced load [A5]:

1. The Sunday 6pm expiry — a hard weekly clock.
2. A weekly streak with a 3-workouts-per-week floor — forgiving enough that it rarely breaks, so it never stops working. Users state the causal chain themselves.
3. A named coach, bonded with parasocially. Reviews name Kelly and Lauren, not Ladder.
4. Cheers — teammates send animated emoji with sound into your screen while you are mid-set. Apple's Editors' Choice writeup leads with it. It only works because a cohort is doing the same workout in the same week.

Volts are inert — 0 mentions across 1,250 reviews, no documented earn/spend/decay, and there is no leaderboard at all [A5]. The visible gamification is cosmetic. The Rewards tab is a referral + merch wallet (30-day Guest Pass for a friend, $25 credit for you, spendable on apparel) — acquisition, not retention [A5].

3. How it makes money

The trial is a purpose-built product, not a sample

listTeams returns a dedicated Welcome Workout per team — flagged isPractice: true and dated 3000-01-04, a sentinel so it can never enter the real schedule. ~32 min, full body [A3]. Every team also ships a welcomeVideoURL, and a team cannot be recommended by the quiz without one [A3][A4].

And the sharpest move in the funnel: trialists are given ELITE-tier Coach Chat and a free Form Check, then lose both on converting to PRO [A3]. You are shown the expensive tier and downgraded into the cheap one.

Because the trial flag is theirs, they can print trial length at will — ~70 promo codes grant anywhere from 14 days to 12 months [A3].

The funding is not what the press says

"$105M raised" is misleading [A6]:

• Equity is only ~$37.5M — $6.5M seed, $12M Series A, and the SEC Form D shows $19.04M actually sold against a $15M announced round.
• The $90M from General Catalyst is a marketing-spend facility covering ~80% of monthly S&M, repaid out of acquired-customer revenue.
• One SEC filing ever. No new equity in 21 months.

Founded ~2018 in Austin by Greg Stewart and Tom Digan. Began as a personal-trainer marketplace at $60/mo, failed on quality control, relaunched July 2020 as coach-led Teams at $29/mo [A6] — which is why the bundle ID is still com.ladder.bootcamp.

4. The moat, correctly located

The intuitive read — "their moat is an expensive human content treadmill" — is wrong, and correcting it is the most valuable output of this Inquiry.

Synthesis: Ladder's supply chain is far more copyable than its brochure implies. What is not copyable is the coach's name attached to the voice and the coach's audience attached to the funnel. A challenger should assume it wins on programming and loses on distribution.

5. Where Ladder is open

Ranked by how cleanly a challenger could attack, and cross-referenced to Ilia's own notes.

1. No voice input. None. Logging is a swipe-up sheet done in rest breaks, and no logging from the Apple Watch at all [A2]. A paying user in the reviews asks for exactly this. This is Ilia's note 1, independently evidenced.
2. The coach cannot respond. Pre-recorded, timeline-bound audio [A2]. Any genuinely conversational coach is a category Ladder cannot enter without rebuilding its economics — its whole 4:1 leverage depends on the audio being a fixed asset.
3. Basics are structurally diluted [A4], three measured mechanisms: - Only 10 of 22 teams list a barbell; 12 of 22 (55%) are dumbbell/kettlebell/ band/mat-only — those teams cannot program a squat, bench or deadlift. - Across all 22 workoutSplits — 154 session slots — not one is named for a squat, bench, deadlift or press. They are Glutes, Arms x Abs, Hourglass Upper Body. - Ladder's schema carries an essentialFitnessMovement flag — they own the distinction and program against it deliberately. The basics exist in the library; they are diluted in the programming. This is Ilia's note 4, confirmed with the supply-chain cause attached.
4. Self-inflicted friction: team switching capped at 2/month unless annual (cited in 2★ reviews as a reason not to buy) [A5]; coach 1:1 messaging moved behind a paid upgrade mid-2026 — one reviewer received a coach welcome message that "only costs another $500 to reply to" [A5]; cancellation by email, 3 business days ahead [A3].
5. Unshipped promises: Android and corporate wellness were both committed in the Nov 2024 raise; no evidence either shipped [A6]. English-only, iPhone-only after six years.
6. Complaints are never about training. Under the 4.95★, the low reviews are overwhelmingly bugs, billing and support [A5]. The product is loved; the company's operations are not.

6. Contradictions and reconciliation

Reconciliation note — team count. [A3] reports 27 teams and 27 Welcome Workouts from listTeams; [A4] reports 22 live teams with 22/22 welcomeVideoURL and practiceWorkoutID. Both read the same API at different times. Most likely reading: 27 total team objects including retired/unlisted, 22 currently live and purchasable. Not resolved — use "~22 live, 27 total objects" and do not state either as fact.

Reconciliation note — offline. A 2024 screenshot shows a downloaded state [A1]; a 2026 reviewer states offline access does not exist [A1]. Possible removal between versions. Parked.

Reconciliation note — the screenshots are stale. The App Store set is a ~2024 artefact — legacy 1242×2208 canvas, 2024 award laurels, in-app dates of Aug 2023 and Apr 2024, and no nutrition despite it shipping 2025-10-27 [A1][A6]. Every screen claim in [A1] is labelled seen (2024) vs changelog-only (current). Ilia's own hands-on notes are the most current primary source in this Inquiry.

Correction to an earlier orchestrator claim. Nutrition did not ship in v4.11.0 (Aug 2026). It launched 2025-10-27; the August release only adds a food library, and it is bundled free — retention and defence, not a new revenue line [A6].

Award claims, checked [A6]: iPhone App of the Year finalist verified (lost to Tiimo). CNET 2026 verified. Apple Editors' Choice unverifiable. Women's Health "Best Overall" contradicted — WH awarded "Best Strength Training App"; no "Best Overall" exists in that article. One of four badges is inflated.

7. Implications for the case study

Audience: a generalist tech company, open-ended brief, Ilia sets scope [STATE]. The panel does not know Ladder. The scoping decision is itself being graded.

The thesis that travels outside fitness:

Ladder looks like a content company and is priced like one. It is actually a paid-acquisition company with a 4:1 content leverage ratio and a coach who cannot hear you. The leverage is the product — and it is also the ceiling.

That argument is legible to people who have never lifted a weight, which is exactly the grading criterion. It rests on three verifiable numbers (56% S&M vs 7% content; 1,716 authored → 6,864 delivered; 80% new-to-category), not on fitness taste.

The wedge the evidence points at (proposed — Ilia deferred this question, so treat as a recommendation he can override): a coach that can hear you. Voice-first logging and conversational coaching. It attacks the one thing Ladder's economics forbid, it is the opening two independent Areas found, it matches Ilia's own first note, and it demos in sixty seconds — which matters more than feature completeness in a hiring room.

Prototype scope, recommended: the two screens that decide the app — in-workout logging and the rest timer — with voice as the input method. Explicitly do not build onboarding, settings, a social feed, or a content library. Say so out loud; the cut list is the seniority signal.

8. Open questions

• Live team count: 22 or 27. Unresolved between [A3] and [A4].
• Does an A-week repeat re-serve the same audio, or is it re-cut per air date? [A4] Changes the recurring media figure 4× (≈51k vs ≈200k clips/year). Unverifiable publicly.
• Volts semantics — streak, level, or vestigial. 0/500 reviews mention them [A1][A5].
• Offline download — present in 2024, denied by a 2026 reviewer [A1].
• Dollar cost per shoot day — not public [A4].
• Effort% definition — provably not %1RM (70%/12 reps → 100%/6 reps in one session); RPE re-skinned, undocumented publicly, likely a 2026 addition [A2].

9. Confidence map

Inquiry-wide caveat: the session's web-search budget (200) was exhausted mid-run, and Scout A4 was terminated by an account session limit after writing its Finding. No Finding is missing, but several carry explicitly parked items that more searching would have closed.

Finding A1 establishes that this set is a ~2024 artefact — legacy 1242×2208 canvas, 2024 award laurels, in-app dates of August 2023 and April 2024, and no sign of the nutrition feature that shipped in October 2025. Read them as Ladder-as-of-2024, not as the shipping app.