Incorporating the Non-Substitution Principle
The Six Pillars of Human Loyalty is a framework that views loyalty as a human relationship sustained by six non-substitutable relational conditions: Recognition, Dependability, Security, Connection, Respect and Understanding.
Each pillar reflects a fundamental question the customer is continually answering through the relationship.
The framework rests on one proposition, stated here as the Non-Substitution Principle:
No single pillar can substitute for a fundamental failure in another.
Two properties follow, and they are distinct. The six are simultaneous: they are conditions that coexist, not stages a relationship passes through. The six are non-substitutable: strength in one does not compensate for the collapse of another. The first describes how the framework is arranged. The second is the claim it makes.
Loyalty is not repeat purchase, retention or reward behaviour. It is the result of a relationship the customer repeatedly chooses to maintain.
Repeat purchase is evidence that a relationship may exist. It is not the relationship itself, and it is not reliable evidence. A customer can buy repeatedly out of convenience, contract, switching cost or absence of an alternative. None of those are loyalty, and all of them look like loyalty in the data.
The framework therefore starts from a different question. Not how do we make the customer come back, but what does this customer need from the relationship in order to keep choosing it.
That shift has one consequence worth stating plainly. Loyalty is not something a business gives a customer. It is something the customer decides, and keeps deciding, based on what the relationship repeatedly makes them feel.
Each pillar is named as a condition of the relationship, and expressed as the question the customer is silently answering.
| Pillar | The question the customer is answering | What it is not |
|---|---|---|
| Recognition | Do I matter here? | Being identified. A system knowing your name is not the same as a relationship knowing you count. |
| Dependability | Can I count on you? | A single good experience. Dependability is a judgement formed across time, not within one transaction. |
| Security | Can I be myself without hesitation? | Data security. This is psychological: whether the customer can ask, admit, change their mind or push back without cost. |
| Connection | Do I want to keep coming back? | Frequency. A customer can return often without wanting to. |
| Respect | Do I value what you stand for? | Politeness. This is about whether the relationship holds something the customer is willing to be associated with. |
| Understanding | Do I even need to explain? | Personalisation. Personalisation says you know me. Understanding says I should not have to explain myself again. |
The final row carries the most weight. Understanding is the only pillar whose test is the absence of something — the effort the customer no longer has to make. It is also the pillar that is hardest to evidence, because what it produces is a question nobody asks.
The six pillars are not six scores that can be added together.
Consider a relationship in which Recognition, Understanding and Connection are all exceptionally strong, and Dependability has effectively failed. The arithmetic says the average is high. The relationship says otherwise. A customer who feels deeply understood does not stop caring whether the supplier can be relied on. Someone who trusts a business completely still notices when they are no longer treated as someone who matters.
Strength in one pillar does not buy forgiveness in another.
This is the framework's central claim, and it is deliberately stated in a form that can be argued with:
When any one of the six conditions falls below what the customer requires, the remaining five cannot restore the relationship to what it was. The relationship does not become slightly weaker. It becomes a different relationship.
This does not mean the relationship cannot recover. It means recovery must address the failed pillar itself; strength elsewhere cannot substitute for that repair.
Most ways of assessing a commercial relationship permit the opposite. A strong result in one area quietly offsets a weaker one elsewhere, and the aggregate still reads as healthy. That logic is reasonable for performance measurement. It is a poor description of how people decide whether to stay.
Two points of scope, so the claim is not read as wider than it is.
First, this is a proposition about relationship integrity, not about choice. There is an established tradition in consumer research describing non-compensatory decision rules — screening mechanisms by which an option failing any one threshold is eliminated regardless of its other merits. That work concerns how people filter alternatives before choosing. The Non-Substitution Principle concerns what happens inside a relationship that already exists. The resemblance is real and worth knowing; the claim is not borrowed from it.
Second, no claim is made here that existing relationship-quality models assert the opposite. They largely do not address substitutability at all. What is asserted is that this property has not been made explicit, and that making it explicit changes what the framework is for.
The Non-Substitution Principle produces two states of a relationship, and they differ in kind rather than degree.
The Support Threshold is the minimum level a pillar must cross before it begins supporting the relationship. It is set by the customer, not by the business, and where it sits for one customer is not where it sits for another. Drawn as a single line across all six conditions, it represents the support criterion in visual form; it is not a claim that the six are psychologically identical or measured on the same scale.
Optimised Loyalty is the state in which every pillar has crossed the threshold. The relationship is fully supported. The lowest pillar may rest barely above the line; that is sufficient. "Optimised" here does not mean maximised. It means every condition is sufficient to support the relationship.
Conditional Loyalty is the state in which at least one pillar has crossed but not all. The relationship exists and continues, but it stands on incomplete support.
A pillar below the threshold is not absent. It is present, and the customer can see it, but it is not yet carrying any weight.
The six pillars are not expected to be equal in height, and this is the point most easily misread.
Tolerance is the distance between the highest and the lowest pillar when all six sit above the threshold. A relationship can be far stronger in Recognition than in Connection and remain Optimised. Unevenness is permitted; falling below the line is not.
Conditional Range is that same distance when the threshold falls inside it — some pillars above, some below. The spread looks similar. The state is not.
This distinction matters because non-substitution is otherwise read as a demand that all six be maximised, which would make the framework an ideal rather than a usable standard. It is not a demand for balance. It is a demand that nothing falls below what the customer requires.
Conditional Loyalty is not weaker loyalty. It is a relationship held together by a different set of reasons.
The commercial signals are often identical. The customer buys, reports satisfaction, renews. What has changed is that part of what was holding the relationship together is no longer doing so, and something else — convenience, contract, switching cost, habit — has taken up the load without anyone deciding that it should.
That is why a Conditional relationship is described as present but vulnerable. It is not deteriorating. It is dependent on conditions that were never chosen and are not being managed.
A deliberate departure from the existing literature should be noted here. Established typologies of loyalty — monopoly loyalty, cost-of-change loyalty, habitual loyalty, and the captive and convenience-seeking categories that follow from them — generally exclude such relationships from loyalty altogether, on the grounds that the customer is not choosing. This framework does not exclude them. It treats Conditional Loyalty as a state within loyalty, because from the customer's side the relationship continues and the behaviour continues; what has changed is what holds it in place.
The reason for naming the state rather than disqualifying it is practical. A relationship excluded from the category is no longer tracked, and the business loses the ability to see when, and from which condition, the change began. Those typologies describe what a relationship has become. This framework is concerned with how it got there, and which pillar stopped carrying weight.
The six pillars have no order. They are not a sequence a relationship advances through, and there is no pillar that must be satisfied before another becomes available.
This distinguishes the framework from stage models of loyalty, which describe loyalty as a progression — cognition before affect, affect before intention, intention before action — and from maturity models that treat relational depth as something a business climbs toward.
The distinction is not cosmetic. A stage model implies that a weak early stage blocks the later ones, and that the remedy is to go back and fix the foundation. This framework implies something different: a relationship can be strong in what a stage model would call an advanced condition while failing in what it would call a basic one, and the failure is no less serious for sitting out of order.
Simultaneity on its own is a modest claim. Most multidimensional frameworks could assert it, and few would dispute it. It is included here because it is the structural arrangement the Non-Substitution Principle requires: conditions that coexist can fail independently, and independent failure is what makes substitution impossible.
Simultaneity describes the shape. Non-substitution is the claim.
The relationship does not weaken by degree. It changes in kind.
This is the part most easily missed, because the commercial symptoms often arrive late or not at all. The customer may still buy. They may still report themselves satisfied. Volume may hold for a year. What has changed is the reason they stay, and that reason is not recorded anywhere.
A relationship held together by Recognition, Connection and Understanding is one the customer chooses. A relationship that continues after one of those conditions has failed is often held together by something else — convenience, contract, switching cost, inertia. It is still a relationship. It is no longer loyalty in the sense this framework describes.
Three practical consequences follow.
The diagnostic question is therefore not how loyal is this customer. It is which of the six is no longer holding, and how long has it not been holding.
The six conditions are not equally visible, and that asymmetry is itself a finding.
| Pillar | Where the evidence sits | How visible in a CRM |
|---|---|---|
| Recognition | Whether past context is used without being re-supplied; whether what the customer once told you still shapes how they are treated | Partially — the record exists, its use does not |
| Dependability | Delivery, availability, response, consistency over time | Visible. The pillar most directly answered by transactional and operational data |
| Security | Whether the customer raises problems, changes their mind, or pushes back — and what happens when they do | Not visible. Lives in the exchange itself |
| Connection | Whether contact is initiated by the customer, and about what; whether the relationship survives a gap | Partially — contact is logged, its direction and reason often are not |
| Respect | Whether the customer is judged by what their business has become rather than what it used to be | Not visible |
| Understanding | What the customer no longer has to explain; requests anticipated before they are made | Not visible. Its evidence is an absence |
Only one of the six is directly answered by transactional data. Two are partially visible, and three are not visible in the CRM record at all.
This is not a complaint about CRM design. It follows from how systems are built: they record what has a historical trace and a recurring pattern. The conditions that matter most to a relationship frequently have neither.
The practical implication is that a loyalty assessment drawn only from system data will be an assessment of Dependability, reported as an assessment of loyalty.
If loyalty is a relationship sustained by six conditions, the operational question follows immediately: how does a business keep those conditions holding as the customer changes?
That question has produced three connected bodies of work.
Customer intelligence. Most of what would reveal the state of the six conditions is not transactional. It sits in a conversation, a salesperson's observation, a request that could not be fulfilled at the time, or a change in the customer's own business model that never existed as a field. Connecting those layers is not about collecting more data; it is about recognising when the evidence available cannot explain what changed.
Customer Interaction Management (CIM). The conditions are not sustained in a system. They are sustained in interactions. CIM addresses the moment the customer is actually in front of someone: what context is available, what judgement that person has to make, what support they have, and what is learned when the customer leaves without buying. CRM holds what is known about the relationship; CIM is where that knowledge becomes an interaction.
Adaptive intelligence. A customer's needs change, and the conditions that satisfied them change with them. Recognition is not a thing done once; it is the continuing question of what makes this customer feel seen now. The same holds across all six. Intelligence that can connect signals, interpret changing circumstances and anticipate what may become relevant moves the question from what will this customer do next to what do we need to become for them to keep choosing us.
The last of those is where the framework turns back on the business. Loyalty is not a one-way requirement placed on the customer. It is a standard the business has to keep meeting as the customer's world changes.
The decision remains human. Intelligence can surface possibilities, identify patterns and challenge assumptions. What matters most, what risk is acceptable and where to invest are judgements the business still has to make — including, sometimes, the judgement to do nothing.
This framework was developed by Gary Chitzu Yeung and set out publicly in the Loyalty Series, a continuing sequence of written essays published on LinkedIn beginning in mid-2026. This document is Version 1.0: the first consolidated statement of the framework, fixing its definition, its six conditions and its central proposition.
What is claimed. That this combination is original: loyalty understood as a human relationship, sustained by six named conditions that operate simultaneously and cannot substitute for one another, each expressed as a question the customer is continually answering.
What is not claimed. No claim is made to the individual terms. Recognition, dependability, trust, connection, respect and understanding are long-established in work on relationships and customer experience, and other six-part frameworks exist under similar names. The contribution is the architecture and the proposition, not the vocabulary.
Use and attribution. The framework may be used, taught, applied and built on freely. The request is attribution: that it be identified as the Six Pillars of Human Loyalty, developed by Gary Chitzu Yeung.
Suggested citation:
Yeung, G.C. (2026). The Six Pillars of Human Loyalty: loyalty as a human relationship sustained by six non-substitutable relational conditions. Version 1.0.
Development record. The Loyalty Series essays published from mid-2026 onward constitute the development record of this framework, tracing its progression from an initial observation about customer behaviour, through the six conditions and their underlying questions, to the operational work on customer intelligence, CRM and interaction that follows from it. That record is deliberately preserved rather than rewritten.
Status. Version 1.0 is a formulation, not a validated measurement model. The Non-Substitution Principle is stated as a proposition open to testing and to challenge. Subsequent versions will be numbered and dated.
內含不可替代原則（The Non-Substitution Principle）
人本忠誠六支柱（The Six Pillars of Human Loyalty）是一套把忠誠視為人際關係的框架：這段關係由六個不可替代的關係條件所維繫——被看見（Recognition）、可依靠（Dependability）、安心（Security）、連結（Connection）、尊重（Respect）、被理解（Understanding）。
每一根支柱，都對應著客人在關係中持續回答的一個根本問題。
整套框架建立在一個命題上，在此稱為「不可替代原則」（The Non-Substitution Principle）：
任何一根支柱，都無法替另一根的根本失效作補償。
由此延伸出兩個性質，而且兩者不同。六者是同時存在的：它們是共同成立的條件，不是關係依序經過的階段。六者是不可替代的：某一根特別強，換不到另一根崩壞時的原諒。前者描述這套框架的結構，後者才是它提出的主張。
忠誠不是重複購買、不是留存率、也不是回饋行為。它是一段客人反覆選擇繼續維繫的關係所產生的結果。
重複購買只是「關係可能存在」的證據，它不是關係本身，而且不是可靠的證據。客人可能因為方便、合約、轉換成本或沒有替代方案而反覆購買。這些都不是忠誠，而它們在數據上全都長得像忠誠。
因此這套框架從另一個問題開始。不是「我們要怎麼讓客人回來」，而是「這位客人需要從這段關係裡得到什麼，才會繼續選擇它」。
這個轉換有一個必須講明的後果：忠誠不是企業給客人的東西。那是客人自己決定、而且持續在決定的事，依據的是這段關係反覆讓他感受到什麼。
每一根支柱都以「關係的條件」命名，並以客人心裡默默在回答的那個問題來表達。
| 支柱 | 客人正在回答的問題 | 它不是什麼 |
|---|---|---|
| 被看見 Recognition | 我在這裡重要嗎？ | 不是被辨識。系統知道你的名字，和一段關係知道你有份量，是兩回事 |
| 可依靠 Dependability | 我還能靠你嗎？ | 不是一次好的體驗。可依靠是跨時間形成的判斷，不是單筆交易裡的 |
| 安心 Security | 我可以毫不猶豫地做自己嗎？ | 不是資料安全。這是心理層面的：客人能不能發問、承認、改變主意或反對，而不需要付出代價 |
| 連結 Connection | 我想要繼續回來嗎？ | 不是頻率。客人可以常常回來，但並不想回來 |
| 尊重 Respect | 我認同你所代表的東西嗎？ | 不是禮貌。這關乎這段關係裡有沒有客人願意與之連在一起的東西 |
| 被理解 Understanding | 我還需要解釋嗎？ | 不是個人化。個人化說的是「我知道你」；被理解說的是「我不應該每次都要重新解釋自己」 |
最後一列份量最重。被理解是唯一一根，它的檢驗標準是「某樣東西的不存在」——客人不再需要付出的那份力氣。它也是最難拿出證據的一根，因為它產生的結果，是一個沒有人問出口的問題。
六根支柱不是六個可以加總的分數。
設想一段關係：被看見、被理解、連結都極強，而可依靠幾乎已經失效。算術說平均分很高，關係說的卻不是這樣。一個覺得被深深理解的客人，並不會因此就不在意這家供應商靠不靠得住。一個完全信任這家企業的人，還是會察覺到自己什麼時候不再被當一回事。
某一根強，買不到另一根的原諒。
這是整套框架的核心主張，而且它刻意用一種可以被反駁的形式陳述：
當六個條件中的任何一個掉到客人所需要的水準以下，其餘五個無法把這段關係還原成原本的樣子。關係不是變得稍微弱一點，而是變成了另一種關係。
這不代表關係無法恢復。它代表的是：恢復必須直接處理失效的那一根；其他支柱的強度，不能替代那項修復。
大多數評估商業關係的方式，允許的正好相反：某一塊分數高，可以悄悄抵銷另一塊的不足，整體看起來還是健康的。那個邏輯用在績效衡量上是合理的，但它不是對「人如何決定要不要留下」的好描述。
有兩個範圍要先講清楚，以免這個主張被讀得比實際更大。
第一，這是關於關係完整性的命題，不是關於選擇的命題。消費者研究裡有一支談非補償性決策法則的傳統，描述的是一種篩選機制：任何一個門檻未達標的選項，無論其他方面多優秀都會被剔除。那些研究處理的是人在「選擇之前」如何過濾選項。不可替代原則處理的是「已經存在的關係內部」發生什麼事。兩者的相似是真的，也值得知道；但這個主張不是從那裡借來的。
第二，這裡沒有主張既有的關係品質模型持相反立場。它們多半根本沒有處理可替代性的問題。這裡主張的是：這個性質一直沒有被明確提出，而把它明確提出來，會改變這套框架的用途。
不可替代原則會產生兩種關係狀態，而兩者是性質上的差別，不是程度上的差別。
支撐門檻（Support Threshold） 是一根支柱在開始支撐這段關係之前，必須跨過的最低水準。它由客人設定，不是由企業設定，而且它在某一位客人身上的位置，和在另一位客人身上並不相同。圖上把它畫成貫穿六個條件的同一條線，是支撐判準的視覺表現，而不是主張六根在心理上相同、或可以放在同一個量尺上比較。
完備忠誠（Optimised Loyalty） 是六根支柱全部跨過門檻的狀態。關係被完整支撐。最低的那一根可能只剛好跨過那條線，而那就夠了。這裡的「完備」不是指六根都達到最高，而是指六個條件都已足以支撐關係。
有條件忠誠（Conditional Loyalty） 是至少有一根跨過、但並非全部跨過的狀態。關係存在，也繼續著，但它站在不完整的支撐上。
低於門檻的支柱不是不存在。它存在，客人也看得到，只是它還沒有在承重。
六根支柱不被要求等高，而這是最容易被誤讀的一點。
寬容度（Tolerance） 是六根全部在門檻之上時，最高與最低那兩根之間的距離。一段關係可以在被看見上遠遠強過連結，而仍然是完備忠誠。不均等是被允許的，掉到線以下不是。
有條件區間（Conditional Range） 是同一個距離，但門檻落在它裡面——有些支柱在線上，有些在線下。跨距看起來差不多，狀態不是。
這個區別很重要，因為不可替代否則會被讀成「六根都要做到滿分」，而那會讓這套框架變成一個理想，而不是一個用得上的標準。它要求的不是均衡，而是沒有任何一根掉到客人所需要的水準以下。
有條件忠誠不是「比較弱的忠誠」。它是一段由另一組理由維繫起來的關係。
商業上的訊號往往一模一樣。客人還是在買、還是表示滿意、還是續約。改變的是：原本撐著這段關係的某一部分已經不再在撐了，而別的東西——方便、合約、轉換成本、習慣——在沒有任何人決定要讓它承擔的情況下，接下了那個重量。
這就是為什麼有條件忠誠被描述成「存在，但脆弱」。它不是正在惡化，而是正依賴著一些從來沒有被選擇、也沒有被管理的條件。
這裡有一個刻意與既有文獻不同的地方，必須先說明。既有的忠誠度類型學——壟斷型忠誠、轉換成本型忠誠、習慣型忠誠，以及由此延伸的「被困住的」與「圖方便的」分類——通常把這類關係整個排除在忠誠之外，理由是客人並不是在選擇。本框架不排除它。它把有條件忠誠視為忠誠內部的一種狀態，因為從客人那一側看，關係還在繼續、行為也還在繼續；改變的是「什麼東西把它固定在原處」。
選擇為這個狀態命名、而不是取消它的資格，理由是實務上的。一段被排除在這個範疇之外的關係，就不再被追蹤，而企業也就失去了看見「改變從何時、從哪一個條件開始」的能力。那些類型學描述的是一段關係已經變成了什麼；本框架關心的是它怎麼走到那裡，以及哪一根支柱不再承重。
六根支柱沒有順序。它們不是關係依序推進的序列，也沒有哪一根必須先被滿足，另一根才會出現。
這一點讓它和忠誠度的階段模型區分開來——那些模型把忠誠描述成一種進程：先認知、再情感、再意圖、最後行動；也和把關係深度視為企業逐級攀登的成熟度模型區分開來。
這個區別不是修辭。階段模型隱含著：早期階段弱會擋住後面的階段，而解方是回頭把地基補好。這套框架隱含的是另一回事：一段關係可以在階段模型所謂的「進階條件」上很強，同時在它所謂的「基礎條件」上失效，而那個失效並不會因為位置不對就比較不嚴重。
「同時存在」本身是一個溫和的主張。大多數多維度框架都能這樣宣稱，而且幾乎沒有人會反對。它被寫進來，是因為它是不可替代原則所需要的結構安排：共同存在的條件可以各自獨立失效，而獨立失效正是替代之所以不可能的原因。
同時存在描述的是形狀。不可替代才是主張。
關係不是程度上變弱，而是性質上變了。
這是最容易被忽略的部分，因為商業上的症狀往往來得很晚，甚至不會來。客人可能還是在買，可能也還算滿意，量可能撐一整年。真正改變的是他留下來的理由，而那個理由沒有被記錄在任何地方。
一段靠被看見、連結與被理解維繫的關係，是客人主動選擇的。一段在其中一個條件失效之後仍然繼續的關係，往往是靠別的東西維繫的——方便、合約、轉換成本、慣性。它仍然是一段關係，但它已經不是本框架所描述的那種忠誠。
由此延伸出三個實務上的後果。
所以診斷的問題不是「這位客人有多忠誠」，而是「六個裡面哪一個已經撐不住了，而且撐不住多久了」。
六個條件的可見度並不相同，而這個不對稱本身就是一個發現。
| 支柱 | 證據在哪裡 | 在 CRM 裡的可見度 |
|---|---|---|
| 被看見 Recognition | 過去的脈絡有沒有被使用、而不需要客人重講一次；客人曾經說過的話，現在還有沒有在影響他被對待的方式 | 部分可見——紀錄存在，有沒有被使用則看不到 |
| 可依靠 Dependability | 交付、供貨、回應、跨時間的一致性 | 可見。最能由交易與營運資料直接回答的一根 |
| 安心 Security | 客人會不會提出問題、改變主意、表達反對——以及他這麼做的時候發生了什麼 | 看不到。活在互動本身裡 |
| 連結 Connection | 聯繫是不是由客人主動發起、為了什麼而發起；關係能不能撐過一段空白 | 部分可見——聯繫有紀錄，但方向和原因往往沒有 |
| 尊重 Respect | 客人是以「他的生意已經變成什麼」被評價，還是以「他過去是什麼」被評價 | 看不到 |
| 被理解 Understanding | 客人不再需要解釋的是什麼；在他開口之前就被預先想到的需求 | 看不到。它的證據是一種缺席 |
六根裡面只有一根能由交易資料直接回答；兩根只能部分看見；三根在 CRM 紀錄中幾乎完全不可見。
這不是在抱怨 CRM 的設計。它來自系統被建立的方式：系統記錄的是有歷史軌跡、會重複出現的東西，而對關係最重要的那些條件，經常兩者都沒有。
實務上的意涵是：一份只從系統資料得出的忠誠度評估，其實是一份「可依靠」的評估，卻被當成忠誠度的評估來報告。
如果忠誠是一段由六個條件維繫的關係，那麼營運上的問題立刻就來了：當客人在改變，企業要怎麼讓這些條件持續成立？
這個問題延伸出三塊相連的工作。
客戶情報。 能揭示六個條件狀態的東西，大多不是交易性的。它們坐落在一段對話裡、業務員的一個觀察裡、一個當時無法滿足的要求裡，或者客人自己商業模式的改變裡——而那從來不曾以欄位的形式存在。把這些層連起來，重點不是收集更多資料，而是認出手上的證據不足以解釋到底發生了什麼改變。
客戶互動管理（CIM）。 這些條件不是在系統裡被維繫的，是在互動裡被維繫的。CIM 處理的是客人真正站在某個人面前的那一刻：有什麼脈絡可用、那個人必須做出什麼判斷、他有什麼支援，以及客人沒買就離開時我們學到了什麼。CRM 保存的是我們對這段關係知道什麼；CIM 是那些知識變成一次互動的地方。
具適應性的智慧。 客人的需求會變，而原本滿足那些需求的條件也跟著變。被看見不是做過一次就算了的事，它是「現在什麼會讓這位客人覺得被看見」這個持續的問題。六根皆然。能夠連接訊號、詮釋變動中的情境、預先看出什麼可能變得重要的智慧，會把問題從「這位客人接下來會做什麼」推到「我們需要變成什麼，他才會繼續選擇我們」。
最後這一項，是框架轉回來面對企業自己的地方。忠誠不是單向加諸在客人身上的要求，而是企業必須隨著客人的世界改變而持續達到的標準。
決策仍然屬於人。智慧可以提出可能性、辨識模式、挑戰既有假設。什麼最重要、能承擔多少風險、資源投在哪裡，仍然是企業必須自己做的判斷——有時候，包括判斷什麼都不做。
本框架由楊濟慈（Gary Chitzu Yeung）發展，自 2026 年年中起透過 LinkedIn 上持續發表的「忠誠度系列」（Loyalty Series）文章公開提出。本文件為第 1.0 版：框架的第一次完整彙整，確立其定義、六個條件，以及核心命題。
主張的是什麼。 主張這個組合是原創的：把忠誠理解為一段人際關係，由六個被命名的條件所維繫，這些條件同時運作、彼此不可替代，而且每一個都被表達為客人持續在回答的一個問題。
不主張的是什麼。 不對個別詞彙提出任何主張。被看見、可依靠、信任、連結、尊重、被理解，在關係與顧客體驗的研究中早已存在，市場上也有名稱相近的其他六項框架。本文的貢獻在於架構與命題，不在詞彙。
使用與署名。 本框架可自由使用、教學、應用與延伸。所請求的是署名：請標示其為「人本忠誠六支柱」（The Six Pillars of Human Loyalty），由楊濟慈提出。
建議引用格式：
Yeung, G.C. (2026). The Six Pillars of Human Loyalty: loyalty as a human relationship sustained by six non-substitutable relational conditions. Version 1.0.
發展紀錄。 自 2026 年年中起發表的「忠誠度系列」（Loyalty Series）文章構成本框架的發展紀錄，呈現它如何從一個關於客戶行為的初始觀察，走到六個條件及其背後的提問，再走到由此延伸的客戶情報、CRM 與互動工作。該紀錄刻意保留原貌，不予改寫。
狀態。 第 1.0 版是一份論述，不是一個已被驗證的測量模型。不可替代原則以一個可供檢驗、也可供挑戰的命題形式提出。後續版本將標註版號與日期。