Methodology
Measure both sides.
A company can be extractive and valuable at the same time. CHI separates those questions instead of forcing a simplistic verdict. A price increase, advertising, a subscription, a paywall, or an upgrade is not inherently hostile: the question is what happens to the relationship over time.
Is the company extracting value from customers faster than it is delivering value back?
CHI Methodology v2.0 is the current, frozen scoring methodology used across active CHI assessments. It supersedes v1.0 for current scoring. v1.0 is preserved further down this page as historical/original methodology, not deleted.
CHI: Customer Hostility Index
Degree to which company behavior extracts money, time, data, flexibility, attention, or control. Scored across five dimensions totaling 100 points under v2.0.
CVI: Customer Value Index
Usefulness, quality, reliability, innovation, access, and benefit the customer receives. Scored across five dimensions totaling 100 points under v2.0.
CFS: Customer Fairness Score
CFS = CVI − CHI. Measures whether the company or the customer benefits more from the overall relationship.
Value alone is not enough. Extraction alone is not enough. The relationship between them is what matters.
How CHI works
CHI begins with observable behavior: pricing history, product changes, advertising, access restrictions, subscription structures, customer policies, and public disclosures. Actions are evaluated against the CHI Lexicon, a defined taxonomy of recurring patterns. Under v2.0, patterns describe behavior; the five CHI dimensions score the underlying customer harm. Value delivered is scored separately through CVI's five dimensions. The result is an assessment of the exchange between company and customer, not simply a hostility score.
Core principle
Direction matters more than one decision.
One price increase does not establish Price Creep. One ad does not establish Advertising Creep. What matters is direction, repetition, magnitude, and accumulation over a documented historical baseline.
A company can grow more expensive while becoming substantially better. Another can grow only slightly more expensive while quietly removing access. Those are not equivalent relationships. CHI measures the trajectory, not the screenshot.
The scoring model · CHI v2.0
Five dimensions, 100 points.
What each CHI dimension measures · v2.0
The measurement architecture.
Revenue Extraction — 25 pts
Measures how aggressively a company increases the amount of customer value it captures, through mechanisms including price increases, new fees, surcharges, advertising inserted into paid relationships, paid restoration of previously included value, paywalls, subscription/tier escalation, and other monetization expansion.
Patterns provide evidence but are not separately added together as independent penalties.
Behavioral Manipulation — 25 pts
Measures commercial use of mechanisms involving attention, uncertainty, scarcity, curiosity, urgency, behavioral dependence, psychologically salient signals, and engagement mechanics used to increase monetization or commercial outcomes.
Ordinary product engagement is not automatically customer hostility; there must be a meaningful customer cost or commercial manipulation component.
Customer Restriction — 20 pts
Measures reduction in customer choice, access, functionality, portability, configuration authority, ability to use previously available capabilities, or ability to leave or switch.
Relevant patterns can include Feature Erosion, Access Downgrading, Platform Lock-In, Customer Lock-In, Choice Illusion, and Platform Creep where customer control is materially reduced.
Information & Privacy — 15 pts
Measures expansion in data collection, tracking, profiling, behavioral inference, cross-context matching, third-party sharing, retention, secondary use, or surveillance/measurement practices.
Do not equate the mere creation of useful customer data or product artifacts with hostility. Customer consequence, consent, control, persistence, and use matter.
Trust & Transparency — 15 pts
Measures whether the company creates material gaps between customer expectations and actual treatment, stated promises and later behavior, disclosed terms and practical consequences, apparent choice and actual choice, or advertised simplicity and material commercial conditions.
Relevant evidence can include Promise Reversal, opaque pricing/conditions, hidden or poorly disclosed restrictions, misleading defaults, material non-reversible changes, or governance failures affecting customer trust. Disclosure and corrective action count as mitigating evidence where appropriate.
25 + 25 + 20 + 15 + 15 = 100 points.
Anti-double-counting rule
Patterns describe behavior; dimensions score customer harm. A single underlying company action may support multiple CHI pattern labels, but its customer harm is scored in only one primary scoring dimension, unless the evidence establishes genuinely distinct harms in separate dimensions.
Example: a company auto-enabling an AI feature could plausibly be described as Platform Creep, Feature Activation Pressure, and Commercial Steering. That does not automatically create three separate penalties. If the same implementation independently creates a privacy harm, that separate harm may also score under Information & Privacy. The purpose of this rule is to prevent artificial score inflation.
The lexicon
18 patterns, three categories.
Revenue extraction
Increasing the amount or frequency of value taken.
Customer restriction
Reducing access, flexibility, ownership, or choice.
Behavioral manipulation
Design that shapes attention or decisions for gain.
CHI's named behavioral patterns, including Price Creep, Advertising Creep, Feature Erosion, Access Downgrading, Rentalization, Platform Lock-In, Choice Illusion, Platform Creep, Feature Activation Pressure, Data / Tracking Creep, and Promise Reversal, remain part of the analytical framework as diagnostic vocabulary. They identify and describe company behavior. They are not automatically independent scoring buckets: the five CHI dimensions above are the measurement architecture. Some patterns, like Platform Creep and Feature Activation Pressure, are provisional concepts used on individual assessments and not yet formal Lexicon entries; formalizing them is a taxonomy decision reserved for cross-company normalization.
Evidence standard
Evidence first → Classification second → Assessment third.
Assessments rely on pricing, product documentation, terms and policies, corporate announcements, and regulatory material wherever possible. Customer dissatisfaction alone does not establish a pattern. If the evidence changes, the assessment changes with it.
What CHI doesn't measure
Not whether a company is good or bad. Not whether it should make money. Not whether customers like it. CHI asks one narrower question:
The value side · CVI v2.0
Five dimensions, 100 points.
Core Product Value measures the fundamental utility delivered by the product or service. Feature & Capability Improvements measures meaningful additions that expand customer utility, choice, convenience, quality, or outcomes; mere feature quantity does not qualify automatically. Technology & Performance measures meaningful improvements in technology, performance, accessibility, usability, efficiency, or technical capability. Trust, Safety & Reliability measures reliability, safety, security, consistency, customer protections, and dependable delivery of the core product relationship. Innovation measures meaningful product or service innovation that creates genuine customer value.
30 + 25 + 20 + 15 + 10 = 100 points.
Reconciling the two sides
CFS = CVI − CHI.
CFS measures whether the company or the customer benefits more from the overall relationship. It does not net CHI and CVI against each other during scoring: hostility is scored in CHI, value is scored in CVI, and the two only reconcile through this formula.
These are the site's existing CFS classification ranges, preserved unchanged under v2.0.
Historical methodology
v1.0, preserved.
CHI originally scored companies against five more category-specific dimensions: Price Creep (25 pts), Advertising Creep (25 pts), Feature Erosion (20 pts), Data Collection Expansion (15 pts), and Customer Lock-In (15 pts).
v1.0 worked naturally for subscription and media businesses, where price, advertising, feature erosion, data, and lock-in map cleanly onto how those companies extract value. Cross-sector testing showed it did not consistently capture customer hostility across SaaS, marketplaces, platforms, airlines, devices, and other business models, where the same underlying harms (restricted access, manipulative defaults, opaque terms) don't sort neatly into those five category-specific buckets.
v1.0 was not wrong for the companies it was built around. v2.0 is an evolution toward a more universal, cross-sector scoring architecture: it separates behavior (patterns) from harm (dimensions) so the same five dimensions can score a streaming subscription, an airline, a dating app, and a workplace platform on a consistent basis. Company assessments published before v2.0 retain their original v1.0 scores; updating this methodology page does not retroactively rescore them. Sitewide score normalization, if it happens, is handled as a separate project.
Interpreting the result
Every assessment answers four questions: what changed? What patterns does it represent? What value remains? Which direction is the relationship moving?
The score summarizes the analysis. It does not replace it.
Methodology v2.0: current and frozen for active scoring. v1.0 preserved above as historical baseline.