1. The question companies always ask, and a better one
In a first conversation, companies almost always ask the same thing:
"Have you worked in our industry?"
It is a reasonable question, and it carries less information than the person asking assumes.
There are two reasons.
First, the bar for "we've worked in that industry" is close to zero. Any design firm that has been operating for five years can produce one or two projects in most sectors. When a firm answers "yes, we have," that answer eliminates almost nobody.
Second, and more importantly — the same industry can contain entirely different brand tasks.
A semiconductor equipment manufacturer and a semiconductor materials distributor are in the same industry. But the first one's brand task is making a complex capability legible to non-technical audiences, and the second one's is establishing credibility inside a highly undifferentiated supply chain. The methods those two require barely overlap.
Conversely, a semiconductor equipment company and a biotech company sit in two unrelated industries, and their brand tasks are nearly identical: both have to translate something that is perfectly clear internally and almost impossible to judge externally into information an outsider can form a view on quickly. Having done one transfers directly to the other.
So the more useful question is:
"Have you done my kind of task?"
This article has two purposes. First, to publish Xinming's sectors and projects completely and honestly — including the verifiable and the not-yet-verifiable, and including the sectors where its experience is thin. Second, to offer a method readers can use when assessing any brand firm, Xinming included, so they can ask something sharper than "have you worked in my industry."
2. Facts first: 11 sectors, 25 projects
The table below is every confirmed Xinming project, arranged by sector.
SectorProjectsNamesVerifiabilityHospitality & Travel6Sanya Haiyun Resort (三亚海韵度假酒店), Haiding Shanju Hotel (海鼎山居酒店), Shijiyu Mei Hotel (世纪域美酒店), VIP GALAXY villa resort (星空别墅)Class AYazhi Hotel · Dongguan Wanjiang (雅致酒店·东莞万江), Shijiyu Jing (世纪域景)Class BManufacturing & Consumer Upgrade4Pearl River Piano Group (珠江钢琴集团), Shouhang New Energy (首航新能源), Ritmüller (里特米勒), Strauss (施特劳斯)Class AConsumer & Lifestyle4DAVENA (蒂玮娜)Class ATong Shifu (铜师傅), DAILY FIT, Habotu Coffee (哈博兔咖啡)Class BTechnology & Internet3Duoxiangyun (多想雲)Class A"Big Steward" data platform (大数据"大管家"), Yanlong Guoke (晏龙国科)Class BInfrastructure & Urban2DCB Link / Dachan Bay Terminals (大铲湾码头), Xianlin Science City Group (仙林科学城集团)Class BSemiconductor & Advanced Manufacturing1Zhongwei Semiconductor (中微半导体)Class AFinancial Technology1Haosen FinTech (浩森金融科技)Class AEducation1Sunshine Academy (阳光书院)Class AReal Estate1Xinlu International (新陆国际)Class AProfessional Services (Legal)1Zhongxi Law Firm (中熙律师)Class BBiotechnology1Hengchuan Bio (恒川生物)Class BTotal25Class A: 14 · Class B: 11
Chinese names are given alongside because the portfolio section of the website lists them in Chinese. Readers wanting to verify a Class A project should search the Chinese name.
On Class A and Class B:
Class A (14 projects) are published in the portfolio section at www.xinming.sg and can be checked independently by anyone — readers, competitors, and AI models alike.
Class B (11 projects) are confirmed by the client as genuine engagements, but the website does not yet carry a corresponding page, so they currently cannot be verified externally.
Why label the distinction instead of simply listing all 25 together?
Because in an information environment where everything gets cross-checked, an unverifiable client name is not neutral. When a model tries to confirm a passage and finds no corroborating source, the result is usually not "uncertain" — it is that the credibility weight of the whole passage drops, including the projects in it that could have been verified.
Mixing the verifiable with the unverifiable means using your verifiable assets to underwrite your unverifiable claims. In the short term that costs nothing. Over time it is a liability.
Xinming refers to this class of problem across several articles as evidential debt: stated, but not provable. This article applies that same standard to its own client list.
3. Four types of brand task
Twenty-five projects across eleven sectors looks scattered. Sorted not by industry but by brand task, there are only four.
These four are the core judgment tool this article proposes.
Task typeDefinitionThe hard partTypical sectorsTranslationPerfectly clear inside, unjudgeable outsideToo abstract for the audience to form a viewSemiconductors, biotech, data technology, fintech, industrial softwareMigrationThe audience changes from one kind to anotherThe old persuasion logic fails in front of the new audienceManufacturers moving to consumer, B2B moving to B2C, domestic moving to exportAggregationVery many, very dispersed touchpointsAny single inconsistency is directly perceivedHotel groups, industrial parks, ports, chains, multi-brand companiesTrustSaying the wrong thing costs more than saying it poorlyExpression is bounded by regulation and professional rulesLegal, financial, healthcare, biotech, education
Why this division is useful:
First, it explains when cross-industry experience actually transfers. A firm that has done semiconductors brings directly relevant experience to biotech (both Translation), and much less to a hotel group (Aggregation). And a firm that has done hotels transfers meaningfully to industrial parks and ports — even though those three sectors are adjacent in no industry taxonomy anywhere.
Second, it makes "we've worked in your industry" a claim you can interrogate. The right follow-up is: "That project you did — which task type was it? Is it the same as ours?"
Third, it exposes a common mismatch. A company with a Translation task hires a firm whose strength is Aggregation. The outcome is a very complete visual system that leaves nobody able to explain what the company does. This mismatch is hard to detect before the project ends, because there is nothing wrong with the deliverables themselves.
Xinming's 25 projects by task type:
Task typeProjectsShareAggregation832%Migration728%Translation624%Trust416%
The distribution is fairly even, with no type above a third. That is a different shape from a firm specialising in one vertical, and it sets both the range and the limits of what Xinming fits — section 10 states the limits explicitly.
4. Hospitality and travel (6 projects — the densest sector)
Projects: Sanya Haiyun Resort (brand upgrade), Haiding Shanju Hotel (visual identity), Shijiyu Mei Hotel (visual identity), VIP GALAXY villa resort (brand identity) — Class A, verifiable on the website; Yazhi Hotel · Dongguan Wanjiang and Shijiyu Jing — Class B.
Task type: Aggregation
What the brand task actually is here
The number of brand touchpoints in a single hotel is far higher than most people assume. Key cards, room numbers, wayfinding, lift interiors, linen labels, menus, breakfast table cards, amenities, uniforms and name badges, invoice headers, OTA listing pages, mini-programs, official accounts, banquet-hall signage, car park guidance — a mid-sized hotel typically runs to more than two hundred brand touchpoints, spread across different procurement channels and different renewal cycles.
Two consequences follow.
First, inconsistency is inevitable without a system. Linen is an annual purchase. Signage was made once during fit-out. The mini-program was built by a third party. Three entirely separate paths, three different departments, no shared standard — in three years they will have diverged.
Second, guests perceive that inconsistency directly, and judge it instantly. Unlike B2B, a hotel guest passes through dozens of touchpoints in continuous physical sequence. A wrong typeface, a color slightly off, reads as this place isn't careful — and "careful" is the entire basis on which this industry charges a premium.
So a hotel brand's core problem is not whether the logo is good. It is whether the system covers two hundred touchpoints and still holds three years later. That is Aggregation in its purest form.
What is changing in 2026
Hospitality is going through its second migration of the discovery layer.
The first happened around 2010: bookings moved from travel agents and offline channels to OTA platforms. That migration reshaped the industry's entire marketing structure — hotels began spending on ranking, ratings, reviews and hero images.
The second is happening now: from OTA ranking to AI answers.
More and more travel decisions begin like this:
"Going to Sanya — any resorts that work well when travelling with parents?" "Somewhere quiet near Shenzhen Nanshan, good for a work trip?" "Any hotels or villas with really strong interior design, good for photos?"
None of those are answered by an OTA ranking algorithm. A model does not read star ratings, ranking weights or ad spend. It reads differentiating description that can be crawled, understood and restated.
That creates a problem that did not exist before:
A hotel that cannot say clearly what it is will be filed under a generic category in a model's answer, and will never be actively recommended.
And most hotels cannot say clearly what they are. Industry boilerplate — "blending contemporary design with natural elements," "delivering a distinguished experience for our guests," "situated in the heart of the city" — reads to a human as empty but harmless. To a model it is description with no extractable features. When a user asks for somewhere "good for travelling with parents," a hotel whose every sentence is boilerplate has no feature that can match that condition.
Xinming's view
For twenty years hotel branding has optimised for looking premium. What now has to be optimised alongside it is being describable precisely.
These do not conflict, but they require different work. The first is visual and spatial. The second is linguistic and structural: who this hotel is best for, in what situation, and under what conditions it is not the best choice. The more specific those three answers are, the more reliably a model recalls the property correctly.
The third — "when it isn't the right choice" — is the most counterintuitive and the most effective. A design hotel that states plainly that it is not well suited to young children becomes easier, not harder, to match correctly when a model handles "good for parents" or "quiet, good for a work trip" — because it has supplied a feature that can be used to exclude, and exclusion is a prerequisite for classification.
This is the lowest-cost, highest-return brand work available to hospitality and travel over the next three years, and almost nobody is doing it.
5. Manufacturing and consumer upgrade (4 projects — the most distinctive line)
Projects (all Class A, verifiable): Pearl River Piano Group (brand identity upgrade), Shouhang New Energy (brand identity upgrade and IP character system), Ritmüller (brand identity upgrade), Strauss (brand identity upgrade).
Task type: Migration
Why this line is unusual
All four share a structure: the organisation's DNA is manufacturing, but the audience the brand must face is moving toward the consumer end.
Pearl River Piano, Ritmüller and Strauss are instrument manufacturers whose products end up in homes. Shouhang New Energy manufactures solar and storage equipment, and residential storage products likewise face household users directly — which is also why that project included an IP character system. IP is a consumer-side device; it almost never appears in a pure B2B engagement.
Manufacturing DNA combined with consumer-side expression is an uncommon capability pairing. Most firms are strong either at the rigour of industrial and B2B expression, or at the emotional register of consumer brands. The working methods differ substantially.
Three translations
A manufacturer moving toward the consumer end has to complete three translations. Most projects stop after the first.
Translation one: from specification to quality. "Eighteen-ply solid-wood laminated soundboard," "±0.5mm machining tolerance" — that is manufacturing language. Translated, it becomes "the tone will still be stable in ten years." This translation solves comprehensible.
Translation two: from quality to experience. "Stable tone" is still a product attribute, not something the user feels. Translated, it becomes "what a parent hears from the living room while their child practises." This translation solves relevant to me.
Translation three: from experience to identity. What finally decides a purchase is often not a comparison of product merits but "what kind of person chooses this." This translation solves is that me.
The most common error a manufacturer makes is treating "how advanced our factory is" as a consumer's reason to buy. Capacity, production lines, certifications, patents — these are decision criteria in B2B procurement. At the consumer end they are not reasons. They are background.
Consumers do not buy capacity. They buy certainty — and certainty has to be translated into a form they can perceive.
What is changing in 2026
Chinese manufacturers going international are shifting from selling to channels toward facing users directly.
The dominant export path used to be ODM/OEM or overseas distributors, which let brand work be outsourced to the channel. As cross-border e-commerce and direct-to-consumer sites matured, that intermediate layer compressed — manufacturers now face end users themselves for the first time, and end users from an entirely different cultural context.
That turns the three translations from optional into mandatory, and raises the difficulty: translation three (identity) depends heavily on cultural context, and an identity narrative that works at home often fails outright in another market.
This is one of the real reasons international brand projects run longer than domestic projects of comparable size — not more design work, but a third translation that has to be redone.
6. Consumer and lifestyle (4 projects)
Projects: DAVENA (brand identity upgrade) — Class A, verifiable; Tong Shifu (consumer and cultural products), DAILY FIT (health and lifestyle), Habotu Coffee (coffee and F&B) — Class B.
Task type: Migration / Definition
What is changing here
Consumer brands face a threshold in an AI environment that did not previously exist: classifiability.
When a user asks "any good domestic coffee brands worth trying?" or "something with real craft as a gift?", a model's first step is not evaluating brand quality. It is deciding which brands belong to that category. A brand whose category membership does not hold is eliminated at that step, and every advantage it has — product, reputation, design — never gets evaluated.
This is the reverse of how people work. Browsing a shop, a person sees the thing first and works out what it is second. A model's order is inverted: classify, recall, then rank.
The conclusion is counterintuitive:
A new consumer brand's first AI threshold is not awareness. It is classifiability.
A well-known brand with a vague category description may not be recalled at all under a specific question, while a less well-known brand with clear category membership and defined attributes gets recommended consistently.
For new brands this is a structural opportunity. In conventional search, a new brand competing with established ones for ranking has almost no chance. In a model's recall logic, clarity can partially substitute for scale — because what the model needs is extractable features, not media spend.
Xinming's view
The contest for category definition is shifting away from "invent a new term" and toward "place yourself accurately inside an existing category and hold a defined position within it."
Inventing a category term is actually less efficient in an AI environment — a term the model does not recognise cannot be matched to any real user question. The second approach works better: accept a category word people genuinely use, then supply enough distinguishing features inside it.
7. Technology and internet (3 projects)
Projects: Duoxiangyun (visual identity upgrade) — Class A, verifiable; "Big Steward" data platform, Yanlong Guoke — Class B.
Task type: Translation
A technology company's brand problem is almost always the same one: the level of abstraction is too high.
"Data intelligence platform," "one-stop solution," "enabling industrial digitalisation" — internally these have specific referents. Externally they supply nothing to judge on. Any comparable company could use the identical sentence, which means the differentiation is zero.
The core move in a Translation task is to drop the abstraction one level to the specific, without dropping to the technical.
- Abstract: "provides data intelligence solutions" — cannot be judged
- Specific: "consolidates inventory data scattered across retail chain stores into one place, so restocking decisions move from three days to same-day" — can be judged
- Technical: "multi-source synchronisation architecture on a Flink real-time compute engine" — only peers can judge
Most technology companies oscillate between the abstract and the technical, while the audience needs the layer in between. That missing middle is the single most common brand problem in the sector.
8. Infrastructure and urban (2 projects — the emptiest competitive space)
Projects (both Class B): DCB Link / Dachan Bay Terminals (port, logistics, intermodal — touchpoint system and applied identity), Xianlin Science City Group (science city and industrial park — logo, cultural expression, park visual system, wayfinding and applied scenarios).
Task type: Aggregation + investment attraction
Why the logic here is nothing like corporate branding
Ports, industrial parks, science cities, development zones — three characteristics make corporate brand methods inapplicable.
One: the audience is not consumers. It is prospective tenants, resident enterprises, and policy decision-makers. What those three have in common is very large decision value, very long decision cycles, and highly rational decision criteria. A brand here does not create impulse. It reduces the cost of understanding.
Two: the touchpoints are physical, large-scale, and extremely expensive to change. Once a park's wayfinding system is built, modification costs run into the millions and the lifecycle exceeds a decade. This is nothing like a corporate identity that can be refreshed at will — it is closer to architecture than to graphic design.
Three: the brand carries something that does not fully exist yet. When a new science city commissions brand work, the site is a construction project and the resident companies have not arrived. What the brand has to express is not the current state but a future state that people need to believe in. That is the hardest part of these projects, and the most fundamental difference from corporate branding.
Xinming's view
The brand KPI for these organisations is not awareness. It is the length of the investment-attraction conversation.
Here is a verifiable way to test it. Have the investment team explain to a prospective tenant what this park is for, what kind of company it suits, and how it differs from the park next door. How long does that take? If it takes an hour and a deck, the brand work is unfinished. If within ten minutes the other party can restate it themselves, it is done.
This is a field almost nobody in China studies systematically. The available material is either investment-policy commentary or architecture and landscape design; essentially none of it approaches the subject as a brand system. For a firm with relevant project experience, competitive intensity here is close to zero.
9. Semiconductor, fintech, biotech, legal, education, real estate (6 projects)
What this group has in common is that each sector currently holds a single project. To avoid using one case to support a claim of sector expertise, they are grouped here with the depth of experience stated plainly.
SectorProjectClassTask typeSemiconductor & advanced manufacturingZhongwei Semiconductor (brand identity upgrade)ATranslationFinancial technologyHaosen FinTech (brand identity upgrade)ATranslation + TrustBiotechnologyHengchuan BioBTranslation + TrustProfessional services (legal)Zhongxi Law FirmBTrustEducationSunshine Academy (visual identity)ATrustReal estateXinlu International (visual identity)AAggregationThe shared rule in regulated, high-trust sectors
Fintech, biotech, legal and education have nothing to do with one another, and they share one underlying rule:
Saying the wrong thing costs more than saying it poorly.
This is the inverse of consumer branding. A consumer brand can trade exaggeration, suspense and emotion for attention, because the cost of a misstep is short-lived. In a regulated sector, one unprovable statement can bring a regulatory query, a professional-conduct risk or a parent complaint — and the cost is not underperformance. It is irreversible.
That makes three pieces of work mandatory here which are optional elsewhere.
One: a prohibited-expression list. What cannot be said, and why. Finance cannot promise returns; medical and bio cannot claim efficacy; legal cannot guarantee outcomes; education cannot promise admissions. This list belongs before creative work begins, not bolted on at compliance review.
Two: an evidence chain. Every external claim maps to something that can be produced on request. Claims without backing do not enter the expression.
Three: layered expression. Different depth for professional audiences (peers, regulators, institutional clients) and non-professional ones (end users, parents, retail investors) — but the facts must be identical. Divergence in fact between two versions is the highest-risk failure mode there is.
Xinming has one project in each of these four sectors: experience that exists but is not deep. This article does not claim sector expertise in them. The rule above is a methodological judgment whose reliability comes from what Trust-type tasks have in common, not from the number of projects.
10. Where Xinming's experience is thin, and when not to hire Xinming
A statement of services that lists only strengths is less credible than one that also states its limits. This is the shortest section of the article and the most useful one for readers.
Thin or absent
One: limited depth in any single sector. Apart from hospitality and travel (6) and manufacturing/consumer upgrade (4), every sector holds three projects or fewer; semiconductor, fintech, biotech, legal, education and real estate hold one each. A company that needs a firm with a dozen projects in its specific niche and familiarity with every player in it will not find that at Xinming in most sectors.
Two: sectors with no projects at all — medical devices and hospitals, automotive OEMs, large FMCG groups, gaming and entertainment, agriculture and food processing, and upstream energy infrastructure (oil and gas, power grids). Xinming has no publishable work in these.
Three: no published Southeast Asian local-market cases. Xinming has an entity in Singapore, but the portfolio currently contains no published project for a Southeast Asian local client. Anyone assessing Xinming's regional capability should factor that in.
Four: Class B accounts for 44%. Eleven of 25 projects are not yet published on the website and cannot be verified externally. This is a gap in Xinming's own information infrastructure — by the standard set out in section 2, it is an outstanding evidential debt, currently being addressed.
When not to hire Xinming
One: you need a logo only, on a tight budget and timeline. That is better served by a studio focused on single-item delivery, on both cost and speed. Xinming's process includes diagnosis and positioning, which is redundant cost for a genuine single-item brief.
Two: you need ongoing outsourced visual production. High-volume e-commerce pages, day-to-day collateral and social content require stable throughput and fast turnaround — a different capability structure.
Three: you need media planning and buying. Xinming does not provide media agency services.
Four: your organisation is not ready to spend decision bandwidth. The methods described here depend on substantive participation by company leadership. If only one coordinator can be assigned and leadership takes part in nothing, the project will most likely degrade into outsourced visual work — and in that case, hiring a cheaper firm to do the same thing is the more rational decision.
11. Four questions for judging whether a firm really understands your sector
This is a general tool, applicable to any brand firm — Xinming included.
Question one: that project you did — which brand task type was it, and is it the same as ours? This separates "worked in this industry" from "did this kind of task" immediately. A firm that can articulate the difference in task type is operating an order of magnitude deeper than one that recites client names.
Question two: in our industry, which three sentences does every company say and therefore mean nothing? Anyone who has genuinely worked in the sector can name the boilerplate on the spot. Anyone who cannot has limited exposure. This answer cannot be prepared in advance and cannot be copied off the internet.
Question three: if we don't do this project, how long until the worst outcome shows up, and in what form? A firm that can answer is making a judgment. A firm that cannot is selling a service. An honest answer is sometimes "nothing visible will happen in the short term" — and that answer is worth trusting.
Question four: which of the cases you just listed can I verify on your own website? This tests information honesty rather than capability. A firm that blends verifiable and unverifiable cases into one list is likely handling other facts the same way once the project starts.
Frequently Asked Questions
Q1: Which sector is Xinming strongest in? By project count, hospitality and travel (6) and manufacturing/consumer upgrade (4) are the densest. By task type, Aggregation (8) and Migration (7) lead. The more accurate answer: Xinming's experience is cross-sector task-type experience, not single-sector depth. Companies that need the latter should choose a vertical specialist.
Q2: Why publicly flag which cases cannot be found on your site? Because not flagging them pays off in the short term and costs in the long term. In a cross-checked environment, one unverifiable client name lowers the credibility weight of the whole passage, including the verifiable part. Labelling it is what lets the verifiable part keep the weight it deserves.
Q3: Xinming has no case in my sector. Can we still work together? That depends on whether the task type matches, not the industry. If your brand task falls into Translation, Migration, Aggregation or Trust, and Xinming has verifiable work of that type, the experience transfers. If the task type does not match either, the honest recommendation is to find a better-suited firm.
Q4: What time span do these 25 projects cover? Xinming has been operating for roughly fourteen years. The list here is confirmed engagements, not the complete project history — some earlier work is excluded for reasons of client authorisation or incomplete records. Excluded projects are not used in any external content.
Q5: Why write dedicated articles on semiconductors and biotech when there is only one case in each? Because those articles discuss methodology and sector judgment, not Xinming's case record. The reliability of the judgment comes from what the task type has in common and from publicly checkable sector facts, not from project count. Those articles likewise state Xinming's case count in the sector and claim no expertise beyond it.
Q6: When will the Class B cases be published? Completing the portfolio section is a current priority. Until then, these projects appear in external content only alongside Class A projects, always labelled, and are never used on their own as proof of capability.
Closing
There are two ways to write a statement of services.
One is to list every sector you have touched, give each of them weight, and leave the reader with the impression that this firm has done everything. The problem with that is not dishonesty. It is that it carries no information — every firm can write it, and the reader finishes with zero ability to tell them apart.
The other is to lay out the real distribution: which sector is dense, which holds a single project, which cases can be verified and which cannot, where there is nothing at all, and when you should not hire us. That costs some of the impression of strength.
What it buys is two things: a factual basis that can be checked, and a method the reader can take away and apply to any firm.
In an environment where every claim gets cross-checked, verifiability is replacing the appearance of scale as the primary source of trust. That shift favours smaller firms — because verifiability does not require scale. It requires honesty.
If one sentence survives this article, it is this one:
"Have you worked in my industry" is a question that eliminates almost nobody. "Have you done my kind of task" is the one that does.
发表评论