name: retrospective-essay-jan-2027 type: brand-asset / long-form-essay-outline status: drafted by Casper (agency lead) 2026-05-12 — STRUCTURAL FRAMEWORK + PROMPTS ONLY. The most personal essay in the §4.3 set; 80% of content is Michelle's unique reflection. This outline provides structural prompts (not content) for Michelle to react to, plus voice anchor + anti-patterns + Penn's assembly process. target_publication: michiu.nl/posts/wat-ik-fout-deed-2026 (or similar — Michelle confirms slug at publication time) target_publication_date: first Tuesday of January 2027 (originally first Tuesday of January 2026 per §4.3 cadence; that slot was missed — calendar shifts forward one year; the 2026 retrospective covers 2025 conceptually but the annual cadence starts properly Jan 2027 covering 2026) target_length: 1500–1800 NL words — shorter than other §4.3 essays. Reflection essays compress better than arguments. target_voice: Michelle first-person primary; Penn-Godin structures (+10 from §6.2 deltas) but with story-shape pulled tighter than the position or sourcing essays; Sutherland accent for vulnerability paragraphs (oblique register for honest disclosure) references:

• docs/strategy/2026-baseline.md §4.3 row 1 (the essay brief)
• docs/strategy/2026-baseline.md Tactic 3.4 (the wrong-first-version pattern)
• docs/strategy/2026-baseline.md §2 trend 3 (trying-is-cool / strategic rule-breaking — this essay is the public expression of this trend)
• docs/strategy/2026-baseline.md §3.3 #9 (Don Julio Art-of-Hospitality precedes rankings — vulnerability-as-credibility move)
• docs/strategy/2026-baseline.md §1.4 (cast-is-the-brand — chef-voice essay anchors the cast-led brand)
• docs/strategy/2026-baseline.md §4.1.1 (GEO baseline — this essay feeds memory_documents.chunk_set='chef_voice')
• docs/yue-do.md (canonical reference; this essay references Yue-do but doesn't argue for it like the position essay does)
• docs/penn-press-pitches-q1-2026.md §4 (Versprille — her UvA philosophy background + cook background means she rewards reflection-with-substance; this essay is a Versprille-natural fit)

Retrospective essay — Jan 2027 v1 outline ("What I got wrong" — annual)

"Eén stuk per jaar, in januari, over wat ik in het voorgaande jaar fout heb gedaan. Niet als zelfkastijding — als oefening. Een keuken die niet vertelt wat ze veranderd heeft, vertelt eigenlijk dat ze niet heeft veranderd."

A 1500–1800 word essay for michiu.nl/posts/wat-ik-fout-deed-2026. Michelle-voiced; Penn structures + assembles. The annual wrong-first-version essay per Tactic 3.4 — the public expression of §2 trend 3 (trying-is-cool again). The most personal essay in the §4.3 set; the framework below provides prompts Michelle reacts to during the fill-in session, not content for her to approve.

1. Brief

Why the Jan 2026 slot was missed and what that means. Per §4.3 row 1 brief, the wrong-first-version essay publishes the first Tuesday of every January. The Jan 2026 slot was missed during the strategy-doc compile cycle (Jan-Mar 2026 was internal-Michelle priorities not external-publication priorities). This isn't a strategic failure — it's a calendar slippage. The annual cadence starts properly Jan 2027 covering 2026 (the strategy-pass-4 year). After Jan 2027, every January in subsequent years is a wrong-first-version essay covering the year that just closed.

Why this essay is the most Michelle-dependent of the §4.3 set. Sourcing essay's specifics come from suppliers Penn can document. Position essay's specifics come from lineage Penn can capture in an interview. Signature-dish essay's specifics come from one named dish Penn can describe. Sake essay's specifics come from a curriculum the sommelier authors. This essay's specifics come from Michelle's own reflection on what she got wrong — there is no other source. Penn's role here is genuinely smaller than in the other essays: Penn provides the structural prompts, captures Michelle's reactions, assembles into a publishable shape, and gates against the §4 anti-patterns.

Why the prompts-not-content approach. Other §4.3 outlines have placeholder content (the cha siu dish, the leveranciersnota card, the Yue-do philosophy-name). This outline has placeholder questions — because the content can only come from Michelle's year-in-review thinking. Penn's working session with Michelle is 3-4 hours during which Penn reads each prompt aloud, Michelle responds, Penn captures verbatim, Penn structures.

Why this essay matters strategically (per §2 trend 3):

• Brandwatch reports +25% growth in positive mentions of "cringe" in 2025 + the end of "blanding" + cultural reward for brands that feel real. The wrong-first-version essay is the most direct expression of this trend a restaurant can publish.
• Versprille (per docs/penn-press-pitches-q1-2026.md §4) has a UvA Philosophy + Ethics background + actual professional kitchen time. She rewards reflection-with-substance more than any other NL Tier-1 critic. This essay is a Versprille-natural fit for her column.
• The Brand Memory chef_voice chunk-set feeds every helper that reads brand context — Julius warning-engine, Penn pitch drafting, Sun campaign briefs, Carlos GBP description writing. A chef-voice anchor with vulnerability + specificity is the strongest single source for these chunks.

What this essay must NOT do:

• No "lessons learned" framing. Corporate-jargon trap. The §3.3 #9 Don Julio rule: name the labor of the change, not the abstract takeaway.
• No false-modesty. Has to be specifically wrong + specifically what's now different. "I work too hard" doesn't count.
• No mock-bragging-by-confession. The humble-brag trope ("I realized I cared too much about quality") is the AI-slop version of this essay genre. Julius warns; Penn rejects at edit time.
• No retrospective on things not yet changed. Every wrong-first-version must have a paired now-version. If Michelle still doesn't know what to do about a problem, it belongs in the next essay (or in §F "what I still don't know"), not in §C/D.
• No solo-insight framing. The §3.3 #9 vulnerability-as-credibility move grounds in relationship. Michelle names the person who pushed back / the guest who said something / the staff member who refused — and what they said. Without the relational anchor, the reflection reads as performance.
• No banned phrases per 0038. Specifically: no journey, no growth, no transformation. The wrong-first-version pattern uses concrete verbs ("I changed", "we stopped", "we started") not abstract nouns.

What this essay must DO:

• Open with the practice of writing this essay every January (frame the genre before the content).
• Reference Tactic 3.4 implicitly (the essay IS Tactic 3.4) without academic citation.
• Walk two wrong-first-version → right-now-version trajectories (specific, named, dated).
• Identify one pattern those two trajectories share — the Penn-Godin pattern-recognition payoff.
• Name one thing Michelle still doesn't know (the §3.3 #9 named-refusal move applied to learning).
• Close with the commitment to the next year's essay — operating cadence as commitment.

2. Structural outline — paragraph by paragraph

The essay has 7 paragraphs (~210–250 NL words each). Penn provides the structural prompts and assembles; Michelle's voice carries everything; this is the lightest Penn-drafting load of the §4.3 set.

§ A — The practice (~200 words)

[Michelle's opening — Penn captures verbatim during fill-in session,
then lightly edits. Approximate shape:]

"Sinds [jaar — TBD; either 2024 if this is the 4th annual essay, or
2026 if this is the first formal one in the new cadence] schrijf ik
één stuk per jaar, in januari, over wat ik in het voorgaande jaar
fout heb gedaan. Dit is dat stuk voor [jaar — 2026 if Jan 2027
publication]."

[Eén alinea over waarom de oefening bestaat. Niet als
zelfkastijding. Als manier om te checken dat de keuken het afgelopen
jaar daadwerkelijk iets heeft geleerd — en niet alleen iets nieuws
heeft gedaan. Specifiek waarschuwen voor het verschil tussen "deze
keuken is veranderd" en "ik heb dingen anders gedaan." Het eerste is
zeldzaam; het tweede is alledaags.]

Een keuken die niet kan vertellen wat ze veranderd heeft, vertelt
eigenlijk dat ze niet heeft veranderd.

Why this opening:

• First-person Michelle from sentence 1 — distinct from other §4.3 essays which open with Michiu-the-venue
• Frames the genre before the content — the annual essay practice IS the thesis statement; reading this once tells the reader what kind of essay this is
• "Niet als zelfkastijding" — Penn-Sutherland accent for the oblique opening; refuses the masochism reading without using the word
• Closing line "Een keuken die niet kan vertellen wat ze veranderd heeft, vertelt eigenlijk dat ze niet heeft veranderd" — the §1.1 specificity rule applied to retrospective writing; the most quotable line in the essay; Penn engineers this even if Michelle's verbatim is rougher

§ B — Two wrong-first-versions, from this past year (~300 words split into two ~150-word subsections)

Structural beat: TWO trajectories. Each: what was wrong → what specifically changed → who/what triggered the change. The §3.3 #9 vulnerability-as-credibility move applied twice.

[Michelle picks one DISH-level reflection AND one OPERATIONAL-level
reflection. Penn prompts via the §3 fill-in questions below.]

**Het eerste — een gerecht dat ik [jaar] op de kaart had en [jaar] niet
meer.**

[Michelle's verbatim. The shape Penn cues toward:
  - Welk gerecht was het, hoe lang stond het op de kaart
  - Waarom dacht ik destijds dat het werkte
  - Wat een specifieke gast / kok / sommelier zei dat het anders deed
    werken in mijn hoofd
  - Wat er nu in plaats van staat OF welk gerecht het is geworden na
    de aanpassing
  - Concrete details — geen abstracties. Een ingrediënt-vervanging. Een
    techniek-aanpassing. Een seizoen-beslissing. Een naam.]

**Het tweede — een gewoonte in de keuken-organisatie die ik [jaar] had
en [jaar] niet meer.**

[Michelle's verbatim. The shape Penn cues toward:
  - Welke gewoonte / protocol / werkwijze het was
  - Hoe lang we het zo deden
  - Wanneer ik (of iemand anders) zag dat het niet meer werkte
  - Wat de aanleiding was — een dag, een gesprek, een incident, een
    klacht, een eigen observatie
  - Wat we nu doen
  - Concrete details — pas-tijden, FOH-protocollen, leverancier-besluiten,
    seizoens-aanpassingen. Geen "we werden efficiënter."]

Why this paragraph:

• TWO trajectories (not three, not one) — three reads as listicle; one reads as anecdote; two reads as pattern in formation
• The two are different category (one dish-level, one operational-level) — this lets §E's pattern-recognition do its work; if both were dish-level, §E couldn't generalize
• Penn's prompts (the bullets) are NOT in the final essay — they're the fill-in scaffolding; Michelle's verbatim replaces them in the published version
• The "wat een specifieke gast / kok / sommelier zei" prompt forces the relational anchor — refuses the solo-insight framing per §4.5 anti-pattern below

§ C — One thing in the brand-or-marketing layer (~200 words)

Structural beat: the third wrong-first-version is at a meta level — something about how the restaurant talked about itself that needed correction. This is the most subtle and most Yue-do-relevant reflection because it gets at the brand-positioning evolution.

[Michelle's verbatim. Penn prompts via the §3 fill-in questions.
Examples of what she might surface:
  - "I called what we do 'fusion' for years and stopped because [a
    specific moment / press critique / guest comment]"
  - "Our website said 'Asian cuisine' until I read [our position
    essay / a piece of press / something a colleague published] and
    realized 'Cantonese-Japanese' was always more specific and more
    accurate"
  - "I underestimated how much the room's design was telling people
    what to expect, until [specific moment]"
  - "We had a 'signature dish' that wasn't actually the most defensive
    dish — [the actual signature dish per the Jul 2026 essay] is, and
    the change happened because [specific moment]"
]

Dit derde stuk is moeilijker dan de eerste twee. Bij een gerecht of
een protocol is "verkeerd" zichtbaar. Bij hoe wij over onszelf
schrijven, is het meestal pas achteraf duidelijk — door iemand
buiten te horen wat ze begrepen, en te merken dat dat niet was wat
ik bedoelde.

Why this paragraph:

• The brand-or-marketing reflection is structurally higher than the operational reflections — it sits outside the kitchen, judging how the kitchen has been described
• This paragraph specifically lets Michelle reflect on the Yue-do philosophy-name decision retroactively (was the journey of finding Yue-do the right name a thing this year's essay can name?), OR on the signature-dish identification process, OR on the position-essay framing emergence
• "Bij hoe wij over onszelf schrijven, is het meestal pas achteraf duidelijk" — Sutherland accent extended; the oblique register works here because brand-self-perception is genuinely oblique

§ D — The pattern across the three (~200 words)

Structural beat: Penn-Godin pattern-recognition payoff. What do the three reflections have in common? The reader has been given specific cases; this paragraph rewards them with the general principle.

[Michelle's voice — Penn assembles her observations from the
working-session into a coherent pattern statement.]

[The shape: "Drie wrong-first-versions, drie verschillende niveaus —
het gerecht, het protocol, de manier waarop ik erover schrijf. Wat
ze delen is [X]. Wat [X] mij dit jaar heeft geleerd is [Y]."]

[Common patterns the prompt suggests Michelle might surface:
  - "Ik laat iets te lang staan. De fout in alle drie gevallen was niet
    de oorspronkelijke beslissing — die was redelijk. De fout was
    dat ik te lang heb gewacht voordat ik herzag."
  - "Ik vertrouw nog steeds te veel op mijn eigen oordeel. In alle
    drie gevallen was iemand anders eerder dan ik in het zien dat
    iets niet meer werkte."
  - "Ik onderschat de kosten van consistentie ten opzichte van de
    kosten van verandering. In alle drie gevallen was de keuken-arbeid
    om iets aan te passen kleiner dan de cumulative gast-ervaring van
    het ongewijzigd laten."]

[Michelle picks the genuine pattern; Penn helps articulate it.]

Why this paragraph:

• Pattern-recognition is the Penn-Godin "story-arc payoff" — three specifics resolve into one general observation
• The prompts force Michelle into a structural observation rather than a moral one (the failure-mode of retrospective essays). "Ik laat iets te lang staan" is structural; "I need to be more open-minded" is moral. Structural wins per §1.1 specificity rule applied to retrospective register.
• This is the paragraph that earns the most LinkedIn-excerpt traffic — a Michelle-named pattern reads as expertise; Penn engineers the language to be quotable in 1-2 sentences

§ E — What I still don't know yet (~180 words)

Structural beat: the §3.3 #9 Don Julio named-refusal move applied to learning. Michelle names ONE thing she doesn't yet know how to handle. This paragraph is what separates a credible retrospective from a corporate one — the publicly-undecided puts vulnerability under specific contract.

[Michelle's voice — one specific open question or unresolved issue.
Penn prompts via the §3 fill-in.]

[The shape: "Eén ding waar ik nog niet uit ben in [jaar]: [specific
named issue]. Ik [describe what I've tried so far + why none of
those approaches has been the right one yet]. Ik vermoed dat de
oplossing [direction of suspicion], maar ik ben er nog niet."]

[Examples of what Michelle might name:
  - A specific dish she's trying to make work and hasn't yet
  - A specific operational tension (e.g., between staff continuity and
    creative restlessness)
  - A specific brand-positioning question (e.g., how to handle the
    Mid-Autumn tradition for international guests who don't recognize
    it)
  - A specific business-model uncertainty within the strategic
    boundaries (out-of-venue catering NOT, but maybe a chef's-table
    capacity question)
]

[Closing line — something like: "Dit is wat het volgende stuk in
januari [year+1] hopelijk beantwoordt." Penn engineers the closing
to bridge to next year's essay.]

Why this paragraph:

• The publicly-undecided is the §3.3 #9 hospitality-discipline move at the brand-narrative level. Don Julio's Art of Hospitality didn't claim certainty; it claimed thoughtful attention.
• Versprille will lift this paragraph if she pitches the essay — her column rewards reflection that admits its boundaries
• Closing sentence bridges to next-year's essay — establishes the annual cadence as ongoing practice, not single-publication event

§ F — The practice, continued (~150 words)

Structural beat: §A's frame closed — the year-in-review essay as a practice (not an event). Commits to next year's cycle.

[Michelle's voice — short, declarative, closing.]

[Approximate shape: "Volgend jaar januari is er weer een stuk. Wat
ik dan zal hebben geleerd weet ik nu nog niet. Wat ik wel weet is
dat als ik in januari [year+1] niets nieuws te schrijven heb, dat
betekent dat ik het hele jaar [year] niet goed heb geluisterd."]

[One closing sentence that mirrors the §A opening — Penn-Godin
circular structure. The essay opens with the practice of writing
this; it closes with the commitment to the next iteration.]

Maasstraat 102. Amsterdam-Zuid. Yue-do.

Why this paragraph:

• "Als ik in januari [year+1] niets nieuws te schrijven heb..." — the §3.3 #9 named-refusal move at the next-year level. Michelle is publicly betting that the keuken will have learned something; failure to learn becomes publishable.
• "Maasstraat 102. Amsterdam-Zuid. Yue-do." — same closing as the other four §4.3 essays per §5.0.1 rule 1; consistency anchor

§ G — Footnote / next-essay reference (~80 words)

Andere lange-vormen op michiu.nl die naast dit stuk passen: het
sourcing-stuk van [maart vorig jaar], het positie-stuk van [september
vorig jaar], het signature-dish-stuk van [juli vorig jaar], het
RAI-stuk van [november vorig jaar], en het sake-programma-stuk van
[mei vorig jaar]. Wie deze hele set leest, leest een jaar in zes
stukken. Dit was de afsluiting van [year-1]. De openings van [year]
volgt in maart, met een sourcing-update.

Why this footnote:

• Names the five sibling essays — establishes the §4.3 essay calendar as a set, not isolated pieces
• "Een jaar in zes stukken" — the most quotable closing-line for press; positions the retrospective as the capstone of an integrated annual content arc
• Sets up the next March sourcing-update — annual continuity

3. Specificity checklist — Penn's fill-in prompts for Michelle

This essay's "checklist" is unusual: it's questions Penn asks Michelle during the 3-4 hour working session, not placeholders Penn fills with documented facts.

3.1 Questions for §B trajectory 1 (the dish-level wrong-first-version)

• Which dish did you take off the menu in 2026 that you'd had on for 6+ months prior? OR: which dish on the menu now is a dramatically different version of one you served earlier in 2026?
• What did you originally think made that earlier version work?
• Was there one specific moment — a guest comment, a staff observation, a critic's review, a colleague's question, a personal cooking moment at home — that flipped the assessment?
• Name that person + the moment.
• What did you actually change? A technique? An ingredient? A pairing? A portion? A presentation?
• Why did the change need to be made that way and not some other way?

3.2 Questions for §B trajectory 2 (the operational wrong-first-version)

• Which protocol, habit, or operational pattern did you change in 2026 that had been running for 12+ months prior?
• This could be: FOH timing, pass timing, supplier cadence, dietary-collection process, group-booking handling, staff scheduling, training rhythm, anything you do in the keuken or out front that you do differently now than at the start of 2026.
• What was the moment that triggered the change?
• Who was involved — Anna? Sommelier? Pastry lead? A specific guest interaction?
• What does the new way look like specifically?

3.3 Questions for §C (the brand-or-marketing-layer wrong-first-version)

• Did you call Michiu something in 2026 that you wouldn't call it now? (Cuisine framing? Position framing? Style framing?)
• Did a piece of press, a customer comment, or an internal conversation make you realize you'd been describing the venue wrong?
• The Yue-do philosophy-name decision is itself a major candidate for this section — when did you realize "Cantonese-Japanese" needed a named-philosophy framing rather than a hyphenated description?
• Or: was there a moment when the website / IG bio / menu header / press one-liner said something you wouldn't write now? What changed?

3.4 Questions for §D (the pattern across the three)

• Reading back the three trajectories — is there a structural pattern? Penn proposes three candidate patterns (see §D placeholders); Michelle picks the genuine one or proposes a fourth.
• Specifically — does the pattern relate to timing (when you saw vs. when you acted), trust (who you listen to), cost-assessment (consistency vs. change), or something else?
• One sentence that captures the pattern. Penn engineers the language to be quotable.

3.5 Questions for §E (what you still don't know)

• One specific thing you're working on now where you haven't found the right answer.
• Within the strategic scope per §11.2 — NOT a Michelin participation question, NOT a second-location question, NOT a catering question. Within the keuken or FOH or brand-narrative scope.
• What have you tried so far that hasn't quite landed?
• Where do you suspect the answer lies?

3.6 Working-session format

• Total time: 3-4 hours, single Saturday morning OR two 90-min sessions across two Saturdays (Michelle may want think-time between Q's; check at outset)
• Setting: quiet, recorder on, Penn with notebook + checklist; Michelle with no phone (this is reflection work; interruption-cost is high)
• Order: §A opening → §B trajectory 1 → §B trajectory 2 → §C → §D pattern → §E unknown → §F closing
• After session: Penn assembles verbatim into structured draft (~6 hours); Michelle reads through + light voice-pass (~1 hour); Julius banned-phrase check; ship

4. Essay-specific anti-patterns

Beyond the universal docs/yue-do.md §5 + the §4.3-essay-set common anti-patterns:

4.1 No "lessons learned" framing

The corporate-jargon trap. Per §3.3 #9 Don Julio rule: name the labor of the change, not the abstract takeaway.

4.2 No false-modesty / humble-brag

The most common failure mode of retrospective essays. "I worked too hard" / "I cared too much about quality" / "My pursuit of perfection..." — Julius warns at draft; Penn rejects at edit.

4.3 No retrospective on things not yet changed

Every wrong-first-version in §B + §C must have a paired now-version. If Michelle is still working out what to do about a problem, it belongs in §E ("what I still don't know") — NOT in the trajectory sections. This is structurally important because:

• §B + §C are about closure (you decided, you changed)
• §E is about openness (you're still working on it)
• Mixing the two collapses the essay's argumentative weight

4.4 No publishable insight without a relational anchor

Every wrong-first-version names a specific person who pushed back / a guest who said something / a staff member who refused / a critic who commented. Without the relational anchor, the reflection reads as solo-insight performance — which the Versprille / Broekaert audience punishes.

If Michelle can't name a person for a given trajectory, that trajectory doesn't go in the essay this year. It can go next year if a person eventually surfaces.

4.5 No "growth" / "journey" / "transformation" vocabulary

Per 0038_banned_phrases_2026_felt_not_told.sql — these are the AI-slop self-help register that this essay must specifically avoid. The wrong-first-version pattern uses concrete verbs ("I changed", "we stopped", "we started", "we kept doing X because [reason]") not abstract nouns.

4.6 No publish-without-Michelle-voice-pass

This essay is uniquely Michelle's. Penn assembles; Michelle owns. The voice-pass step is non-negotiable — Michelle reads through the assembled draft AND makes her own edits in her own register. If Penn's assembly reads more like Penn than like Michelle, the essay defaults to delay-publication-by-a-week rather than ship-with-Penn-voice.

5. Voice anchor — what Michelle's voice should sound like

This essay is the most personal voice-test in the §4.3 set. Reference anchors:

Michelle-first-person primary:

• Sentences in past tense for §B + §C trajectories
• Sentences in present tense for §A + §F frame and §E unknown
• The "ik" (I) appears more than in any other §4.3 essay — but doesn't appear more than necessary; sparseness signals seriousness

Penn-Godin (+10) — story-shaping accents:

• §B's two trajectories follow micro-story structure (set up → triggering moment → resolution)
• §A and §F circular framing
• §D pattern-recognition payoff

Sutherland accent (oblique register for vulnerability):

• §A "niet als zelfkastijding — als oefening" — refuses the masochism reading without naming it
• §C "meestal pas achteraf duidelijk" — admits the brand-self-perception is genuinely oblique
• §F "als ik niets nieuws te schrijven heb, dat betekent dat ik het hele jaar niet goed heb geluisterd" — Sutherland mock-self-criticism that's actually self-discipline

Voice anchors NOT to drift toward:

• ❌ Guidara warmth — wrong register; reflection isn't hosting
• ❌ GaryVee kinetic — completely wrong; reflection is paced and considered
• ❌ Kahneman decision-design at scale — works for argument-essays (position, sourcing, RAI), not for reflection. This essay has minimal Kahneman accent compared to the other four.

6. Cross-link to Brand Memory + the chef_voice chunk-set

Per §4.3 row 1 brief: this essay's chunks land in memory_documents with chunk_set='chef_voice'. Once published, the chunks are retrievable by:

• Sun when drafting Campaign briefs that need Michelle-voice
• Penn when drafting press pitches that quote Michelle directly (especially Versprille — her UvA philosophy background means she'll lift reflection-with-substance quotes)
• Carlos when writing GBP descriptions / Schema.org Person (Michelle) markup that needs Michelle-voiced source material
• Julius when checking outbound copy attributed to Michelle — does it match her actual voice patterns?
• Siren when generating review-replies that need to read as Michelle's voice rather than the helper-default

The chef_voice chunks complement the cuisine_position chunks (from the position essay) and the signature_dishes chunks (from the signature-dish essay). The three together form Michelle's published voice corpus.

7. Pre-publication checklist

• Working session completed with Michelle (3-4 hours; recorder; checklist-driven)
• Penn's assembled draft (~6 hours work post-session)
• Two §B trajectories each have: specific dish/protocol/etc. + triggering moment + named person/source + concrete change
• §C brand-or-marketing-layer reflection has: specific framing change + triggering moment
• §D pattern identified by Michelle (not Penn); Penn engineers the language
• §E unknown is specific, scope-appropriate (per §11.2 strategic boundaries), and bridges to next year's essay
• Julius banned-phrase check passes — specifically §4.5 vocabulary (no growth/journey/transformation)
• No false-modesty / humble-brag patterns (per §4.2)
• Every wrong-first-version has a paired now-version (per §4.3)
• Every reflection has a relational anchor — a person named (per §4.4)
• Michelle voice-pass — Michelle reads through and edits in her register (per §4.6)
• Final approval from Michelle per §8.4 escalation
• Carlos prepared to publish at michiu.nl/posts/wat-ik-fout-deed-2026
• Schema.org Article markup drafted (Carlos):
  • author = Michelle Zhang (Person schema)
  • datePublished = 2027-01-05 (first Tuesday of January 2027)
  • keywords includes §4.3 row 1 LLM target queries (chef voice, philosophy, what is the story behind X restaurant)
  • articleSection = "Chef Voice"
• m_campaign tagging on internal links = longform_wrong-first-version_2026
• Versprille pitch optional follow-up (the essay's voice-anchor is Versprille-natural; Penn evaluates pitch in Q1 2027 once published)

8. Decision log

• 2026-05-12 — Drafted outline by Casper (agency lead) per §4.3 row 1 brief. Frame is Jan 2027 retrospective covering 2026 (the Jan 2026 slot was missed; annual cadence picks up properly Jan 2027 going forward).
• Pending — Michelle's 3-4 hour working session with Penn. Should be scheduled for late November or December 2026 so the essay is ready by first Tuesday of January 2027.
• Pending — Penn assembles draft; Michelle voice-passes; Julius check; ship 5 January 2027.
• Pending — Penn evaluates Versprille pitch as Q1 2027 follow-up once essay is published.
• Pending — Annual repetition: the same §3.6 working-session format applies to Jan 2028 (covering 2027), Jan 2029 (covering 2028), etc. Document the format as a recurring Sun-coordinated calendar item (cultural_moments row: annual-wrong-first-version-michelle, moment_date='2027-01-05', recurring pattern).

9. Related references

• Strategy doc: docs/strategy/2026-baseline.md §4.3 row 1 (essay brief); Tactic 3.4 (the wrong-first-version pattern); §2 trend 3 (trying-is-cool / strategic rule-breaking); §3.3 #9 (Don Julio Art-of-Hospitality vulnerability-as-credibility move); §1.4 (cast-is-the-brand — chef-voice anchor); §11.2 (strategic boundaries — what §E "what I still don't know" cannot include)
• Canonical Yue-do reference: docs/yue-do.md §3 (closing geographic+philosophy anchor per §5.0.1 rule 1)
• Companion essays: docs/sourcing-essay-mar-2026.md + docs/position-essay-sep-2026.md + docs/signature-dish-essay-jul-2026.md + docs/rai-corporate-essay-nov-2026.md + docs/sake-essay-may-2026.md (the five sibling essays in the §4.3 set; this essay is the capstone)
• Press pitch: docs/penn-press-pitches-q1-2026.md §4 Versprille (her cook-philosopher background = natural fit for the reflection-with-substance register)
• Voice baselines: 0039_voice_distributions_2026.sql — Penn-Godin +10 + minimal Kahneman (reflection essays compress Kahneman; story-shape carries the load)
• Banned phrases: 0038_banned_phrases_2026_felt_not_told.sql — Julius runs against draft; §4.5 vocabulary trap (growth/journey/transformation) is the highest-frequency failure mode of this essay genre

10. §4.3 essay set — completion status

With this outline, the §4.3 essay set is 6 of 6 outlined:

All six are structurally complete; each has Michelle/sommelier/Anna fill-in requirements documented; each cross-references the others into a coherent annual content arc. The §4.3 bi-monthly cadence is fully drafted at outline-level.