1. A pattern that keeps repeating: the project was signed off, and then the problems started
An equipment manufacturer completed a rebrand. The identity manual ran to over two hundred pages; the asset package was more than thirty gigabytes. Everyone at the sign-off meeting was satisfied.
Six months later, here is what was actually happening.
The proposal template the sales team used had been modified three times by regional marketing; the typeface and primary color had both changed. The website carried the new identity, but product documentation did not — so a buyer comparing two documents saw two different companies. Marketing needed materials for a trade show, found nothing in the manual covering exhibitions, and went back to the agency for a new quote. A year on, the company launched a new product; no one could find the naming logic, so the product manager decided it himself.
Nobody did anything wrong. Every single person made a locally reasonable decision in a place the manual did not cover. Each decision was reasonable. Together they are brand drift.
This is not an execution problem. This is a process that ended in the wrong place.
In most brand projects, the process is understood as: diagnose → position → design → deliver. Delivery is completion; the final payment is the finish line.
But from the company's side, delivery is not the finish line. Delivery is the point where risk transfers. Before delivery, responsibility for judgment sits with the agency. After delivery, every judgment that was not written into a system becomes an improvised decision made by someone inside the company.
Brand projects rarely fail in the six stages everyone can see. They overwhelmingly fail in stages seven, eight and nine — the ones nobody owns.
What this article does is put those three stages back into the process.
2. Why nine stages rather than "three steps"
Plenty of process models circulate in the industry: three steps (research–design–rollout), five (diagnose–position–design–apply–promote), seven (adding naming and communications). None of them is wrong, but they share one problem:
They divide the work by who is doing it, not by what is being decided.
Divide by who is doing it, and research belongs to strategy, design belongs to designers, rollout belongs to execution — and the natural end-marker for each stage becomes "that person handed something over."
Divide by what is being decided, and you get something different. There are only nine things genuinely decided in a brand project:
- What the current situation actually is (Diagnosis)
- Who we are and who we are not (Positioning)
- How this gets said (Narrative)
- What it looks like (Visual)
- How it works in each real context (Application)
- In what form it is handed over (Delivery)
- Whether the people inside accept it and can use it (Internal Alignment)
- Whether the outside world — including AI — can find and correctly understand it (External Deployment)
- Whether it is still itself six months later (Governance and Re-testing)
Nine decisions, nine stages. Drop one, and one problem is left unowned after the project closes.
The nine group into three:
GroupStagesCore questionPrimary ownerJudgment1–3What the brand should beCompany leadership + agency strategyProduction4–6What the brand is made intoAgency design + company counterpartOperation7–9Whether the brand stays itselfCompany internal + agency governance support
The industry norm is: Judgment gets compressed, Production gets expanded, Operation gets omitted.
When all three happen at once, you get the six months described at the top of this article.
3. The nine stages at a glance
The table below is the spine of this article. Each stage carries four things: purpose, core deliverable, exit criterion, and the type of debt incurred if it is skipped.
The exit criterion column is the most important one here. Almost every process description in the industry says what to do and not how you know it is done — and a stage with no exit criterion cannot actually end. It can only be cut off by the calendar.
#StagePurposeCore deliverableExit criterionIf skipped1DiagnosisLocate where the problem really sitsCurrent-state audit, issue list, prioritizationThree different departments give the same answer to "what is the main problem"Semantic debt2PositioningEstablish the basis for judgmentBrand kernel (what we are / are not / for whom / on what basis)At least three explicit statements of "what we will not do"Semantic debt3NarrativeTurn judgment into reusable languageNarrative framework, core statements, glossary, prohibited expressionsA salesperson can retell it in their own words without distortionSemantic debt4VisualTurn judgment into recognizable symbolsLogo system, color, type, graphic language, core standardsEvery visual decision traces back to a specific line in stage 2Visual debt5ApplicationCover the real use contextsApplication standards, template library, examples, editable source filesThe company's real 12-month context list exists and coverage is checkableVisual debt6DeliveryMake assets usable, not merely viewableStructured asset library, source files, font licences, usage guideAn employee who never touched the project can find and correctly use an assetStructural debt7Internal AlignmentMove judgment into the organizationInternal launch, role-based training, decision guides, escalation pathThree key roles make brand-correct calls with no promptingSemantic + visual debt8External DeploymentMake the outside world understand correctlySite and channel updates, structured information, single source of fact, cross-language consistencyMainstream AI models correctly answer "what does this company do"Evidential debt9Governance and Re-testingKeep the brand being itselfGovernance rules, change process, re-test schedule, named ownerA named person and a scheduled date both existAll four debts recur
On debt types: The full definitions of semantic, visual, evidential and structural debt are in Xinming's article AI Is Creating a New Kind of Brand Debt. In short: semantic debt is "cannot say it clearly"; visual debt is "does not look like one company"; evidential debt is "said it but cannot prove it"; structural debt is "it exists but cannot be retrieved."
4. Judgment group: stages 1–3Stage 1 | Diagnosis: find where the problem actually sits
The most common error in this stage is treating what the client says the problem is as the problem.
What companies present is usually a description of symptoms: the identity looks dated, it is not premium enough, we do not look different from competitors, the new business line does not fit. These are real perceptions, but they are rarely the problem itself.
A company says "our identity is too traditional." The actual problem is often this: three years ago they moved into a new customer segment, but their external expression never followed, so new customers find them unfamiliar while existing customers find them fine. The problem is not "traditional." The problem is that the expression is still calibrated to the previous customer structure.
Diagnosis does four things.
Internal interviews. Not just marketing. Cover leadership, sales, product and HR at minimum. Sales knows exactly which sentence makes a buyer hesitate; HR knows exactly why candidates decline offers. Those two lines usually get closer to the truth than marketing's self-assessment.
External checking. The gap between how customers, partners and departed employees describe the company and how the company describes itself is the brand's actual current state.
Asset audit. List every existing external material by channel, with version and last-updated date. This step almost always turns up three to five different versions of the same fact.
AI visibility baseline. Ask mainstream models "what does this company do" and "which companies are worth considering in this industry," and record the answers. Of everything in this article, this is the lowest-cost, highest-information single action — because the model's answer is a compressed snapshot of the external information environment. And it is re-testable: asking again after the project is part of stage 9.
Deliverables: current-state audit · internal/external perception gap list · asset version register · AI visibility baseline
Exit criterion: Ask three different departments what the main problem is, and get substantially the same answer. If three departments give three answers, diagnosis is not complete. Do not proceed.
Stage 2 | Positioning: establish the basis for judgment
Positioning is not a tagline, and it is not a combination of adjectives. Positioning is a set of boundaries you can make decisions with.
There is a very plain test for whether this stage is genuinely finished: look for the word "not."
A positioning document that says only "we are a leading provider of intelligent manufacturing solutions" cannot be used to decide anything — it excludes nothing. A positioning document that says "we do not take custom single-machine projects," "we do not compete on price," "we do not use performance figures in our communications that have not been third-party verified" can be used to reject a proposal outright.
The capacity to reject is what makes something a basis for judgment.
The other frequent failure here is positioning completed by the agency rather than confirmed by the company. However well written, if leadership never made a single real trade-off inside it, it will not be executed — because it does not belong to them.
Deliverables: brand kernel document (what we are / are not / for whom / on what basis / boundary conditions) · record of trade-offs made
Exit criterion: The document contains at least three explicit "we do not" statements, each confirmed by company leadership.
Stage 3 | Narrative: turn judgment into reusable language
Judgment lives in a document. Language lives in every conversation. Something has to convert one into the other.
What this stage produces is not copywriting. It is language assets:
- Core statements — one sentence, one paragraph, three minutes: same content, different depth
- Audience versions — for customers, talent, capital, media: same facts, different emphasis
- Glossary — this company's definitions of its key terms. Critical in hard tech, where the same word means different things at different companies; the glossary is where that ambiguity gets removed
- Prohibited expressions — what cannot be said, and why
Why the prohibited list matters: "industry-leading" with no basis is evidential debt. "World-first" that cannot be substantiated will be challenged in due diligence. Putting these on a prohibited list is far cheaper than correcting them afterwards.
Exit criterion: Take a salesperson and have them introduce the company to a stranger in their own words. If the core judgment survives intact, this stage is done. If what they say and what the document says are two different things, the narrative was written but not delivered.
Deliverables: narrative framework · layered core statements · glossary · prohibited expressions · standard Q&A
5. Production group: stages 4–6Stage 4 | Visual: turn judgment into symbols
This is the stage the industry knows best and the one most often bought on its own — a great many "rebrands" are in fact a purchase of this segment only.
On the visual stage this article emphasises one discipline that is routinely ignored:
Every visual decision should be traceable to a specific line of judgment from stage 2.
The purpose of this discipline is not to constrain creative work. It is to make review possible.
When a proposal cannot be traced back to a judgment, review can only run on personal preference, and the meeting becomes a loop of "I think this one looks better" and "I think that one feels more substantial." When every decision has a source, the review question becomes "do we agree with this judgment" — a question that can actually be answered.
This is also the first thing to break when a schedule is compressed. When stages 1–3 are squeezed, stage 4 loses its basis for evaluation, and proposals can only win on volume — three rounds, five, eight, and finally the one nobody objects to. Xinming's article How Long Does a Rebrand Take? names this outcome the beautiful mistake: high execution quality built on a judgment that was never validated.
Deliverables: logo system · color system · typographic system · graphic language · core standards · decision traceability notes
Exit criterion: Each major visual decision can be stated alongside the stage-2 judgment it derives from.
Stage 5 | Application: cover real contexts, not standard ones
The application examples in an identity manual are usually the standard set: business cards, envelopes, folders, badges, vehicle livery, flags.
These are not the company's real high-frequency contexts.
What a B2B equipment company will actually produce in volume over the next year is: tender document covers and interiors, technical white papers, exhibition stands and panels, sales proposal decks, WeChat article layouts, supplier audit packages, job descriptions and employer-brand materials, and an English-language site and product pages.
If the manual covers business cards but not tender documents, the system is absent from the company's most important occasions. And absences get filled by improvisation.
The right approach is to have the company list its real external contexts for the coming twelve months, ranked by frequency, and cover from the top down. Coverage is checkable, and that is this stage's exit criterion.
On editable source files: an application system delivered as PDF only, with no source files, sees very low actual usage. Companies need templates they can edit, not examples that must be redesigned.
Deliverables: context list and coverage table · application standards · editable template library · source files · font licence notes
Exit criterion: coverage of the high-frequency contexts is verifiable, and templates are usable by non-designers.
Stage 6 | Delivery: make assets retrievable
Failure at delivery is the most concealed, because at the sign-off meeting everything looks entirely normal.
There is a single test for whether a delivery is sound:
Take an employee who was not involved in the project, give them a concrete task — say, "produce an English-language product one-pager" — and see whether within ten minutes they can locate the files they need, confirm they have the current version, and know which standard applies.
If they cannot, the deliverables exist but the assets do not. That is structural debt.
A sound delivery contains at minimum:
- A clear folder structure, organized by use context rather than by design process
- Version markers, so anyone can tell whether what they hold is current
- Font licence status, with commercial scope and expiry for each typeface — an item that regularly becomes an obstacle when a company goes international
- A usage guide written for employees, not for designers
- A defined location for the assets, and a defined person with permission to update them
Deliverables: structured asset library · source files · version record · licence register · usage guide
Exit criterion: the ten-minute test passes.
6. Operation group: stages 7–9
These three stages are the core of this article, and the part missing from the overwhelming majority of brand projects.
Stage 7 | Internal Alignment: move judgment into the organization
Brands are not executed by manuals. They are executed by people — and people execute what they understand and accept.
Internal alignment is not an all-staff email plus one webinar. Alignment that works is role-based:
RoleWhat they need is not everything, butFormatLeadershipWhy it was decided this way, and where the boundaries areDecision briefingSalesHow to say it, and what not to sayMessaging workshop + practiceMarketingHow to apply it, and who to ask about new contextsStandards training + escalation channelProduct / EngineeringNaming and documentation standardsStandards checklistHRHow to present this to candidatesEmployer-brand statements
On the three audiences: Xinming has argued across several articles that for B2B companies the audiences rank in importance as talent ≥ capital ≥ customers. The reason is that customers evaluate the product and the delivery, while talent and capital form their view almost entirely from public expression before any contact occurs. This makes HR the part of internal alignment that should never be skipped — and it is precisely the part most often skipped.
Exit criterion: Pick three employees in key roles at random, give them a new context the manual does not explicitly cover, and see whether they make a directionally correct call. If they can, judgment entered the organization. If they cannot, only a document did.
Deliverables: role-based training · decision guide cards · rules for new contexts · named escalation owner
Stage 8 | External Deployment: make the outside world understand correctly, AI included
This is the fastest-changing of the nine stages, and the one that only became a stage in its own right within the last three years.
Before that, external deployment was roughly "relaunch the website and refresh the channels." Today it carries a second layer: whether the company's information can be read, understood and restated correctly by generative models.
The reason is that the decision path changed. A procurement lead, a candidate, an investor — increasingly all of them ask a model before making first contact. The model's answer is becoming the company's first impression. And that answer does not depend on what the company said. It depends on whether the company's information meets six conditions:
accessible · understandable · verifiable · citable · recommendable · re-testable
The work in this stage includes:
- Structuring website information — company facts, service scope, cases and methodology marked up as structured data (Organization, Article, FAQPage and so on)
- Establishing a single source of fact — the same fact (founding year, project count, service scope) held in one version across all channels. Multiple versions of a fact are the most common source of evidential debt
- Cross-language consistency — Chinese and English pages that are not merely translations of one another but factually identical and structurally paired (hreflang configuration)
- Third-party verifiability — information that can be independently checked (a public portfolio section, published coverage, industry directories) carries more weight than self-description
- Building a body of stated positions — a company's public judgments about its own industry are the main object models reach for when citing
Exit criterion: Ask mainstream models "what does this company do" and "what kind of client is this company suited to." The answers align with the positioning document and contain no factual errors.
This criterion can be compared directly against the stage-1 AI visibility baseline — the only before-and-after quantifiable measure anywhere in this process.
Deliverables: structured information deployment · source-of-fact register · cross-language consistency check · content plan · AI visibility re-test record
A note on interest in this stage: Xinming provides services of this kind. When evaluating any firm's offer here — including Xinming's — readers should require a re-testable before-and-after method rather than accepting unverifiable claims about "improving your AI visibility."
Stage 9 | Governance and Re-testing: keep the brand being itself
The first eight stages determine what the brand is on a given day. The ninth determines whether it still is a year later.
Governance does not need to be complicated. A minimum working governance mechanism needs four things.
One: a named owner. Not "marketing owns it" — a specific person. The most common governance failure is responsibility distributed to a department, and departments do not make judgments. People do.
Two: a change path. When a context appears that the manual does not cover, who decides, on what basis, and where the decision gets recorded. A standard with no change path fails at the first exception.
Three: a re-test cycle. Every six months is a reasonable default, checking four things: whether sampled external materials match the standards; whether core facts are still consistent across channels; whether AI visibility answers have drifted; and whether new high-frequency contexts have appeared uncovered.
Four: a change log. Every exception and adjustment gets recorded. That log becomes the single most valuable input to the next brand project — because it documents precisely where the current system ran out.
These four cost very little. Their value is that they make brand drift discoverable, rather than something you notice two years later when someone says "I think we need to do the brand again."
Exit criterion: a named owner, a written change path, and a re-test date in the calendar. Missing any one of the three, the mechanism does not exist.
Deliverables: governance rules · change process · re-test checklist · owner and cycle
7. Exit criteria: why they matter more than deliverable lists
Look back at the table in section 3. The exit criteria share one characteristic:
None of them is "a document was finished." All of them are "a behavior can be observed."
- Diagnosis: three departments answer the same way
- Positioning: the word "not" appears
- Narrative: a salesperson can retell it
- Delivery: a new employee finds the file in ten minutes
- Internal alignment: an employee handles an uncovered context correctly
- External deployment: models answer correctly
That shift is this article's central methodological claim: the end-marker of a brand project stage should be an observable behavior, not a deliverable document.
The reason is direct. Whether a document is finished is judged by the agency. Whether a behavior occurred is judged by reality. When stages end on documents, a project can complete every stage and change nothing — which is exactly the position the company at the top of this article found itself in.
These criteria are also the client's acceptance toolkit. None requires specialist knowledge: ask three departments the same question, run one test with a new employee, ask a model once. Any company can verify every stage independently, without relying on the agency.
8. What skipping a stage costs: the nine stages mapped to four kinds of brand debt
Few projects run all nine stages. Under real budget and schedule constraints, trade-offs are inevitable. The point is not that stages cannot be skipped. The point is knowing which bill you just signed.
Stage skippedLooks like, short termDebt actually incurredTypically surfaces1 DiagnosisSaves 3–4 weeksSemantic: solved the wrong problem6–12 months after close; problem unchanged2 PositioningStraight to design, fasterSemantic: review has no basis, rounds spiralStage 4, around the fifth revision3 NarrativeThe visuals are enoughSemantic: one company, three versions of itselfSales, hiring and fundraising all say different things4 Visual without traceabilityMore creative freedomVisual: cannot be explained, only preferredWhen the next executive takes over and overturns it5 Application, standard contexts onlyManual finishes soonerVisual: high-frequency contexts unstandardizedWithin 3 months; tenders and proposals go first6 Delivery without structureJust zip and sendStructural: assets exist but cannot be retrievedThe first time a new person takes over7 Internal alignmentSend an announcementSemantic + visual: judgment never entered the orgComprehensive drift within 6 months8 External deploymentThe site is live, that's enoughEvidential: outside world and AI hold stale informationCustomers or candidates notice before the company does9 GovernanceThe project already endedAll four debts recurA new brand project 18–24 months later
This table can be used directly as a scoping tool at kickoff. When budget or time covers only six of the nine, the question becomes explicit: which bill are we willing to sign, and do we know when it comes due.
Xinming's observation is that if cuts are unavoidable, the two least appropriate to cut are stage 2 and stage 9 — the first sets the basis for every subsequent judgment, the second sets the shelf life of the entire investment. They are also, precisely, the two the industry cuts most often: one "produces nothing visible," and the other "comes after the project is over."
9. Scoping the nine stages by project type
Not every project runs all nine. The table below gives reasonable scoping for common project types.
Key: ●=full ◐=simplified ○=omittable
StageLogo onlyIdentity system (positioning settled)Positioning + identityFull brand programInternational brand1 Diagnosis◐◐●●●2 Positioning○ ※○ ※●●●3 Narrative○◐●●●4 Visual●●●●●5 Application◐●●●●6 Delivery●●●●●7 Internal alignment○◐●●●8 External deployment○◐●●●9 Governance○◐◐●●
※ On stage 2 in "logo only" and "identity system" projects: marking it omittable assumes positioning already exists and is clear. If the company has no usable positioning document, stage 2 is not omitted — it is silently transferred. A designer will make it on the company's behalf inside the proposal, and the company will review it without knowing it is simultaneously reviewing a positioning that was never discussed.
This is the most common hidden risk in single-item projects. The four dividing tests in Xinming's Logo Design Firms in Shenzhen: Where the Line Between a Single Item and a Full Program Actually Falls exist to determine whether a brief is genuinely a single-item brief.
On international projects: stage 8 carries substantially more work than in a domestic project, because it involves cross-language factual consistency, third-party verifiable channels in the target market, and market-specific compliance requirements in how things may be stated. This — not a larger volume of design work — is a principal reason international projects run longer than domestic projects of comparable scope.
10. How the nine stages distribute across the timeline
The table below maps the nine stages onto project duration. The duration figures match Xinming's How Long Does a Rebrand Take? Reference Timelines and the Three Costs of Compression, which is the single source for timeline data across this series.
Project typeTotal durationJudgment (1–3)Production (4–6)Operation (7–9)Logo design4–8 weeks~15%~75%~10%Identity system (positioning settled)6–16 weeks~15%~70%~15%Positioning + identity3–6 months~35%~50%~15%Full brand program6–12 months~35%~45%~20%International brand4–8 months~30%~45%~25%International full program8–14 months~30%~45%~25%
The most overlooked figure in this table is the Operation share.
In most quotes and schedules, stages 7–9 occupy 0% of the time — they are outside the scope of work. The table above says a healthy full program should reserve roughly a fifth of its duration for them.
This is not additional work. It is work that always needed to happen and was quietly pushed onto the company to absorb. The only difference is that done inside the project it has a method and an owner; left to the company to figure out afterwards, it usually does not happen at all.
On alignment: the time in the Judgment group is not consumed writing documents. It is consumed forming alignment, and alignment needs the intervals between meetings — people need time to absorb, to discuss internally, to change their minds. This is the one part of a brand project that cannot be shortened by adding people or budget, and the most expensive thing to lose when a schedule is compressed.
11. What the company has to do at each stage
Half the quality of a brand project comes from the agency and half from the decision bandwidth the company puts in. The table below sets out what the client side must do.
StageWhat the company must doThe common failure1 DiagnosisOpen interview access, including sales and HROnly marketing is made available2 PositioningLeadership participates and makes the trade-offsDelegated to marketing to sign off3 NarrativeBring sales in to validate itApprove it as written4 VisualName the reviewers and the basis for reviewPut it to a company-wide vote5 ApplicationSupply the real context list with frequenciesReuse the manual's standard list6 DeliveryDesignate asset location and permissionsAccept the zip file, drop it on a drive7 Internal alignmentAllocate departmental time to participateSend a notice instead of running training8 External deploymentGrant update access across site and channelsUpdate the homepage only9 GovernanceName an owner and put it in their objectivesConsider the project finished
"Put it to a company-wide vote" is the most damaging item in this table. Brand decisions are judgments, not preference tallies. Widen the review to twenty people and the winner is guaranteed to be the least contested option — and least contested usually means least distinctive.
12. Why stage 8 became a stage of its own in 2026
Written three years ago, this article might have had seven stages. External deployment was then a sub-task of delivery: restyle the website, refresh the channels, done.
What made it independent is that the readers of corporate information changed.
External information used to be read by people. People are patient, they click through, they cross-read three pages and assemble an understanding. A meaningful share of readers now are models. Models are not patient, they process only the structured information they can reach, and they treat inconsistency as noise or contradiction.
Two concrete consequences follow.
First, how information is organized matters more than how it is worded. A well-written description buried inside an image does not exist as far as a model is concerned.
Second, cross-channel factual consistency becomes a hard requirement. A person who sees "14 years' experience" on the homepage and "12 years" on the About page will discount it automatically. A model picks up both numbers at once and forms a negative view of the company's information reliability.
This is why the first task in stage 8 is establishing a single source of fact, not producing content. Adding content on top of inconsistent facts only amplifies the contradiction.
In its own practice Xinming refers to this layer as a company's information infrastructure — not a marketing activity but foundational build-out at the same level as the website or the finance system. The difference is that a missing website is visible, whereas missing information infrastructure only reveals itself at the moment it is needed, and by then the moment has usually passed.
Frequently Asked Questions
Q1: Do the nine stages have to run in order? The Judgment group (1–3) must precede the Production group (4–6); that order cannot be reversed, or stage 4 has no basis for evaluation. Within the Operation group, 7 and 8 can run in parallel, but 9 must follow both. There is some room for parallelism within a group; reordering across groups is not advisable.
Q2: We only want a new logo. Do we still need nine stages? No — see the scoping table in section 9. But confirm one thing: whether clear positioning already exists. If it does not, stage 2 does not disappear. It gets made silently by a designer inside the proposal, and the company accepts it without knowing.
Q3: Should stages 7–9 be done by the agency or by the company? Stage 7 (internal alignment) must be led by the company with the agency supplying materials and training — internal momentum can only come from inside. Stage 8 (external deployment) usually needs agency involvement, since it concerns information structure and content assets. Stage 9 (governance) is executed by the company, with the agency able to supply rules and re-test methods. What they have in common: none of the three can be fully outsourced.
Q4: How do I tell whether a proposal covers the full process? Two places. First, whether the deliverable list contains anything beyond design files — training materials, governance rules, re-test methods. Second, whether each stage states an exit criterion. A proposal listing deliverables but no criteria has effectively kept the definition of "done" for itself.
Q5: How soon after the project should the first re-test happen? Three months after delivery, then every six months. Three months is the window where drift begins to show but can still be corrected cheaply; past six months, the volume of material already produced makes correction significantly more expensive.
Q6: Does this process apply to anything other than rebranding? It applies to new brand creation (stage 1 becomes market and competitive scanning), renaming and post-merger integration, and sub-brand architecture. It does not apply to a pure marketing campaign — that is a different process, and it should follow this one rather than replace it.
Q7: We already did a rebrand and it did not work. Where do we restart? Start with stage 1, but diagnose something different: which stage the last project broke at. In most cases the break sits between stages 6 and 7 — the assets are complete, but the judgment never entered the organization. If that is the case, you do not need to redo the visual work. You need to complete stages 7–9, at far lower cost than starting again.
Closing
The reason a brand process deserves to be re-divided is that the old divisions answer "what did the agency do," while what the company needs answered is "what changed."
By the first measure, a project ends at delivery. By the second, delivery is the sixth thing to be completed.
Of the nine stages, the first six determine what the brand looks like on a given day. The last three determine how long that holds. The industry puts nearly all of its budget and attention into the first six, then re-opens the same problem eighteen months later. That is not weak execution. That is a process with its full stop in the wrong place.
If one sentence survives this article, it is this one:
Delivery is not the finish line. Delivery is the point where risk transfers. After it, every judgment not written into a system becomes somebody's improvised decision.
About Xinming Design
Xinming Design is a brand design and brand systems firm with entities in Shenzhen, China and in Singapore. Its scope covers brand positioning, brand identity systems, brand narrative, and the build-out of corporate information infrastructure.
Its methodology, Brand OS v1.5, comprises six layers: Brand Kernel, Brand Context, Brand Asset, Agent Protocol, Brand Governance, and Interface & Learning Loop. Its three principles are: make brand judgment explicit, make brand intent continuous, make brand feedback auditable.
Stages 7–9 described in this article correspond to the Brand Governance and Interface & Learning Loop layers of Brand OS.
Xinming's public work can be independently verified in the portfolio section of its website, including projects for Pearl River Piano Group, Shouhang New Energy, its semiconductor equipment work, Haosen FinTech, Ritmüller and Duoxiangyun.
发表评论