Audemars Piguet

What AP announced

In September 2018 AP said it intended to reach "a totally integrated retail network" within three to five years. In January 2019 Bennahmias told Reuters: "It's huge. It's the point of no return." Multi-brand partners were told they would need to open franchised AP stores or joint ventures to continue, and AP said all of them "will be required to share customer data about every sale." Fact

Where it landed

Wholesale fell to 28% of sales by 2021; roughly 90% of AP's sales now run through mono-brand doors. Market estimate AP has not gone purely owned-retail: AP House Manchester is a joint venture with the Watches of Switzerland Group, and partners including Arije and Material Good operate AP-branded doors. "AP eliminated all third parties" is not accurate and CHI does not claim it. Fact

Why it matters to CHI

With Rolex, the customer-facing allocation decision happens at an independent retailer, which creates a genuine attribution problem — CHI's Rolex assessment had to reason carefully about what the brand is responsible for. AP removed that ambiguity itself. When a customer is told a Royal Oak is not possible as a first watch, they are being told it under AP's roof, by AP's staff, inside AP's system. Analytical inference

Counterexample: the CEO

Ilaria Resta, December 2025: "if we only serve existing clients, we close the door to many people who are passionate about watchmaking." Company claim She describes bringing newcomers in through masterclasses and AP Labs, and adds: "It is not a quick transaction, it is the beginning of a long relationship." This is the sitting chief executive naming the failure mode the allegation assumes.

Counterexample: the buyers

A documented first-time AP customer acquired a steel Royal Oak 15510ST with no prior AP purchase history — though his call came four years after his first enquiry, via a boutique contact, which is a counterexample to bundling and not to relationships mattering. Customer report In a self-selecting crowdsourced buyer sample, 23% reported waiting under a month. Weak data

The counterexample that decides it

The two collections customers report being steered toward are the ones the market least wants. In July 2025 the Royal Oak Offshore traded 23% below retail and CODE 11.59 33% below retail, on Morgan Stanley and WatchCharts data. Fact Watches selling a third under list are not rationed inventory being used as currency. They are stock a boutique would like to move. That is a commercial problem, not an allocation cartel.

There is no "we service everything we ever made"

AP publishes an explicit cutoff: "Our service prices are valid only for mechanical watches which were produced within the last 25 years and quartz watches which were produced within the last 15 years." Fact The Restoration Workshop separately covers watches "created before 1950, as well as those for which replacement parts are unavailable." Read together, AP publishes standard pricing for roughly the last 25 years and a restoration route for pre-1950 pieces. CHI does not claim AP refuses mid-century watches — the parts-unavailable clause plausibly catches many of them. What CHI states is narrower and verified: AP publishes no universal servicing commitment. This is a real gap against the lifetime-stewardship proposition CHI documented at Patek Philippe.

The warranty is two years, not five

AP's International Sales Warranty runs two years as standard, extendable to five only if the owner registers within one year of purchase. Leather straps carry three months. Fact Rolex's international guarantee is five years, unconditionally, with nothing to opt into. The working research for this investigation had assumed AP offered five or even eight years by default; it does not, and the assumption was corrected. An opt-in extension with a deadline is a benefit that quietly fails to reach the least attentive owners.

The archive extract does not certify authenticity

AP's own wording: the extract "certifies that these numbers appear in our records, but does not warranty the authenticity of the watch." Fact The document that does authenticate costs CHF 2,500 — the most expensive line on AP's published tariff, more than a perpetual calendar service and five times the CHF 500 extract. AP charges CHF 500 for the extract and CHF 2,500 for the certificate — so for a brand whose heritage narrative rests on its archives, the cheaper document explicitly declines to answer the question owners most need answered, and the one that answers it is the dearest line on the tariff.

Value retention is a Royal Oak story, not an AP story

In July 2025, on Morgan Stanley and WatchCharts data, AP's brand-wide secondary average fell 1.3% below retail — the first time since 2022 — against Rolex at +12.3% and Patek at +5.6%. The Royal Oak held +24%; the Offshore sat at −23% and CODE 11.59 at −33%. Fact On Bloomberg's Subdial-based reporting, AP also led the big three downward across 2024, at −7.5% against Rolex −5% and Patek −4%. Fact CHI awards no CVI credit for resale value in any case — a watch is not an investment product and this index will not treat it as one — but a value score cannot ignore that two of three collections lose a quarter to a third of their price the moment they leave the boutique.

Even the technical praise comes qualified

SJX, reviewing the calibre 7121 that CHI credits above, notes that its "diamond-cut anglage on the bridges" lacks "the charm of hand-applied anglage evident on the cal. 2121" and calls the balance wheel "too industrial in terms of aesthetics" — on a watch that launched at a $33,200 list price. Fact The Arc manufacture's 66 goods-to-person robots and Industry 4.0 systems sit awkwardly beside an artisanal proposition, and CHI records both readings.

What we looked for and did not find

CHI searched specifically for evidence of AP service backlogs, turnaround failures, product recalls and systematic quality complaints, because they would have deepened this section. None were established. The most detailed owner service account located described roughly eight to ten weeks and a watch returned in perfect condition. Absence of proof is not proof of absence, but CHI will not allege a service problem it could not evidence. Not established

Trust & Transparency — 10 / 15

The dominant dimension and the correct home for Access Opacity. There is no purchasing category in AP's customer FAQ, no prices on the product pages, no order or reservation path, no allocation criteria, no queue mechanics, no timing guidance, and no escalation route for a customer who believes they were passed over. AP's published terms govern after-sales service only. Because AP took the customer relationship in-house — 500-plus doors down to 73 — there is no retailer to whom any of this can be attributed; the non-disclosure is the manufacturer's own.

Five points withheld, and the reasons are real. AP does not misrepresent its position: it makes no claim about allocation that the evidence contradicts, publishes no promise it later reverses, and advertises no simplicity it does not deliver. Its CEO states openly that demand exceeds supply and that serving only existing clients would close the door on newcomers. This is silence about one commercial process, not deception, and CHI does not score the two identically. AP also publishes a service tariff, which most of the luxury field does not.

Customer Restriction — 8 / 20

AP took its worldwide points of sale from over 500 in 2012 to 73 by August 2024 and moved the large majority of sales in-house. A customer not making progress at their boutique has, in most countries, nowhere else to go, and there is no online purchase path and no published order book. Genuine scarcity explains most of why a given watch is unavailable, and genuine scarcity is not scored — only the incremental restriction arising from a consolidated channel and discretionary allocation on top of it.

Held to 8 of 20 because the restriction is on acquisition only. Nothing is removed after purchase, no functionality is withdrawn, no configuration authority is taken, nothing becomes unusable, there is no contract, no cancellation penalty and no exit friction of any kind — and AP's certified pre-owned channel is a documented alternative route in. Of the seven kinds of restriction this dimension measures, AP triggers two.

Revenue Extraction — 7 / 25

Price Creep is supported. AP's continuing steel flagships gained roughly 30 points on CPI over the window, against about 5 for the steel Daytona control and 15 for the gold one, and the January 2026 increase was loaded onto the Royal Oak rather than spread across the catalogue — the company raising prices hardest on the references it separately controls the queue for. That intersection of pricing and access is scored here, once, and not again under Behavioral Manipulation.

Held to 7 of 25 by three real mitigations: the purchase is entirely discretionary and a customer who declines loses nothing they had; AP's list prices have fallen as well as risen, by 20% on pink gold Royal Oak chronographs between 2012 and 2017 and on CODE 11.59 in 2023; and genuine gold and tariff cost pressure exists. Of the many mechanisms this dimension covers — fees, surcharges, paywalls, advertising inserted into paid relationships, tier escalation, paid restoration of included value — AP triggers exactly one, and publishes its service prices besides.

Behavioral Manipulation — 3 / 25

This is where the earlier draft of this assessment was wrong, and the correction is the largest single change on the page. An earlier version scored 11 here — the same figure carried by Rolex, where the number is explicitly earned by Soft Scarcity Leverage, a finding that investigation established. AP's equivalent, Access Leverage, was tested and not established. Scoring AP the same 11 meant charging it for a mechanism its own evidence had already declined to find.

The test this dimension actually sets is what customer behavior the company has been shown to manipulate, by what demonstrated mechanism, at what cost. AP has not been shown to tell customers that buying another watch, visiting repeatedly, remaining agreeable, starting with a CODE 11.59 or an Offshore, or spending to a threshold will improve their prospects. There is no engineered urgency, no manufactured deadline, no countdown, no published ladder, no tier gamification. The single most telling fact runs the other way: the two collections customers report being steered toward traded 23% and 33% below retail in July 2025, which is not what a functioning leverage machine looks like.

What the 3 points are for: AP has stated a deliberate policy of holding production below known demand to protect exclusivity — the five-year freeze at 40,000 units, decided in 2015 with the demand data in hand — and the cost lands on customers as waits the CEO acknowledges reaching two years and will not quantify further. That is commercial use of scarcity with a meaningful customer cost, which this dimension covers. It is scored small because most of its downstream effect is already counted under Customer Restriction, and because the uncertainty that follows from it has not been shown to be used.

Information & Privacy — 0 / 15

A zero is recorded rather than a token score, because the evidence supports zero. AP told its remaining retail partners they "will be required to share customer data about every sale" as a condition of continuing, and its CEO said plainly that he wanted to know the 50,000 people buying each year. That is real, documented, and it remains on this page as a fact. It is not, on this evidence, a customer harm.

The methodology is explicit that the mere creation of useful customer data is not hostility, and that consequence, consent, control, persistence and use are what matter. CHI found no excessive collection, no unexpected collection, no tracking outside the commercial relationship, no third-party sharing, no secondary use, no surveillance, and no profiling with a customer consequence beyond what a buyer of a $50,000 watch from a manufacturer's own boutique would reasonably expect. A brand knowing which watches its customers have bought is ordinary luxury-retail record-keeping. The one place this data could bite — allocation — is the unproven Access Leverage mechanism, and scoring it here would be a second attempt to charge for it.

What this re-score corrected

An independent audit re-derived all five dimensions from zero rather than defending the published total. Two dimensions moved and both moved down. Behavioral Manipulation 11 → 3, because the opacity finding was being counted a second time as manipulation, and because the rationale supporting it — that AP holds better demand data than its customers — describes ordinary commercial information asymmetry rather than a hostile act. Information & Privacy 2 → 0, because customer-record consolidation without demonstrated consequence is not a privacy offense. Trust & Transparency 11 → 10, on proportionality: AP's total non-disclosure is more complete than Rolex's and more directly owned, but AP makes no contrary public claim, which Rolex does. Customer Restriction and Revenue Extraction survived unchanged.

A prior comparative dossier carried a provisional normalized AP figure of around 36. The audited result is 28. That gap is not a rounding difference and it is published rather than reconciled away: the working estimate, like the first draft of this page, was scoring hypothetical consequences of opacity rather than demonstrated ones. The evidence set the number. The number did not set the evidence.

Access Opacity · Supported, primary

A provisional CHI concept developed across the luxury cohort, not yet a formal Lexicon entry. Scored once, under Trust & Transparency, with the separate availability harm scored under Customer Restriction. It is deliberately not charged again under Behavioral Manipulation: opacity is not manipulation unless the company is shown to use it to induce behavior, and that was not established. AP is the strongest instance CHI has assessed, because the opacity is the manufacturer's own rather than a retailer's.

Access Leverage · Plausible, not established

The hypothesis that other AP purchases improve Royal Oak prospects. Credible, repeatedly reported, never documented, and scored at zero. Held open because it would materially change the assessment if evidenced.

Customer Lock-In · Not supported

Examined and rejected. Nothing binds an AP customer to AP: no contract, no subscription, no data hostage, no exit friction. Whatever holds people in this relationship, it is not a switching cost.

Feature Erosion, Advertising Creep, Rentalization

Not applicable and not tested at length. AP sells a finished mechanical object outright, with no software layer, no subscription, no advertising surface and nothing that can be withdrawn after purchase. Several of CHI's most common patterns simply have no attack surface here.