A/B testing
Comparing two versions by randomly assigning participants and measuring a predefined outcome.
Illustrative example
Test two offer explanations against qualified inquiries.
100 terms · Plain-language definitions
Marketing accountability glossary
By Jay Mandel · Updated October 1, 2026
A plain-language guide to marketing measurement, data practices, and consumer trust, with definitions that can be read independently.
All examples are illustrative, not verified client results. Working definitions and accountability lenses describe practical uses of a term; they are not claims of a universal standard or original coinage.
A
A/B testing: definition, example, and related termsBack to index
Access control
Rules and technical measures restricting who can use information or systems.
Illustrative example
Only the assigned support team can open customer records.
Accountable owner
The person responsible for carrying an agreed outcome through delivery.
Illustrative example
One owner follows a complaint across departments.
AI disclosure
Clear information about material AI involvement in a service or communication.
Illustrative example
Tell customers when they are speaking with an automated assistant.
Anonymization
Transforming data so people are not reasonably identifiable in its context of use.
Illustrative example
Assess re-identification risk before publishing a research dataset.
Attribution
Assigning credit for an outcome to observed touchpoints using a stated rule or model.
Illustrative example
A last-click report credits search even though other channels influenced the decision.
Audit trail
A record of actions that supports review of what happened.
Illustrative example
Log who changed a permission setting and when.
B
Bait and switch
Attracting people with an appealing offer that the seller does not genuinely intend to provide, then steering them toward a different offer. A clearly disclosed introductory price is not, by itself, bait and switch; the issue is a misleading offer or substitution.
Illustrative example
A retailer advertises a low-priced product to bring shoppers in, refuses to sell that advertised product, and pressures them to buy a more expensive alternative instead.
Source reference
Related reading
Baseline
The starting reference against which change is assessed.
Illustrative example
Record inquiry quality before rewriting the landing page.
Behavioral debt
The accumulated relationship cost of tactics that manipulate people rather than help them make informed choices. A short-term gain in compliance can create resentment, complaints, and weaker loyalty over time.
Illustrative example
A subscription business makes cancellation deliberately difficult. Churn appears lower in the short term, while frustrated customers complain and become less willing to recommend it.
Related reading
Brand salience
The likelihood a brand comes to mind in a relevant buying situation.
Illustrative example
Ask which provider people think of when service fails.
C
Cancellation friction
Extra effort that obstructs leaving or stopping payment.
Illustrative example
An online signup requires a phone call to cancel.
Category entry point
A situation or need that prompts someone to consider a product category.
Illustrative example
Moving offices prompts a search for connectivity providers.
Causation
A relationship in which changing one factor produces a change in another.
Illustrative example
A controlled test estimates the effect of removing a fee.
Churn
The loss of customers or recurring revenue over a defined period.
Illustrative example
Distinguish canceled accounts from revenue lost through downgrades.
Claim substantiation
Evidence supporting an objective claim before it is communicated.
Illustrative example
Verify a performance claim under the conditions stated in the ad.
Clean data
Clean data combines quality with responsible use: accurate and verified where possible, collected for a clear purpose, used with appropriate permission, and maintained over time. Personal information is protected, and anonymous information is used where appropriate. A smaller, reliable audience can be more useful than a large list of suspect contacts.
Illustrative example
An assessment lets people voluntarily share preferences for a personalized recommendation. The company records the stated purpose and permission, checks the records, and keeps them current.
Related reading
The Clean Data Imperative
The principle that organizations should verify data quality, establish permission and purpose, and maintain the information used for decisions. Unverified records, decaying contacts, and bot-inflated activity are infrastructure problems that require better systems and ownership.
Illustrative example
A company discovers duplicate contacts, obsolete job titles, and records with no clear source or permission history in its CRM. It fixes the collection and maintenance process before launching the next campaign.
Related reading
Cohort analysis
Comparing groups sharing a starting event over equivalent time periods.
Illustrative example
Compare March and April signups at the same age.
Complaint ownership
Responsibility for resolving a complaint across internal handoffs.
Illustrative example
A representative stays accountable when billing needs to intervene.
Confidence interval
A range expressing uncertainty in an estimate under a statistical procedure.
Illustrative example
Report estimated lift with its interval rather than one precise number.
Confirmation bias
Favoring evidence that supports an existing belief.
Illustrative example
A team reads compliments but ignores cancellation feedback.
Consent architecture
The design of how an organization requests, records, honors, and updates people’s permission to use data. In a consent-based process, choices should be understandable, specific, and easy to withdraw. Clear explanations at collection need to connect to the systems that control how information is used.
Illustrative example
An email sign-up explains what will be sent and how the information will be used. The company records the choice, timestamp, and wording shown at signup, provides an easy unsubscribe route, and ensures that withdrawal reaches the mailing system.
Source reference
Consent withdrawal
A usable way to reverse a previously given permission.
Illustrative example
Provide a working unsubscribe without requiring login.
Content provenance
Information about the origin and editing history of content.
Illustrative example
Retain the creator and modification record for an image.
Contribution margin
Revenue remaining after specified variable costs.
Illustrative example
Subtract delivery and payment costs before evaluating a promotion.
Conversion rate
The share of eligible visits or people completing a defined action.
Illustrative example
Divide completed bookings by eligible visitors using the same period.
Correlation
A statistical association that does not by itself establish cause.
Illustrative example
Sales and ad spend rise together during a holiday.
Customer acquisition cost
The stated costs of acquiring customers divided by customers acquired.
Illustrative example
Include sales effort when reporting a fully loaded acquisition cost.
Customer autonomy
A person’s ability to make meaningful choices without coercion.
Illustrative example
Offer a clear way to decline personalization.
Customer effort
The work a person must do to get an outcome from a service.
Illustrative example
Count transfers and repeated explanations when reviewing support.
Customer lifetime value
An estimate of customer value over a relationship using explicit assumptions.
Illustrative example
Model retention and contribution margin rather than treating revenue as profit.
D
Dark patterns
Interface choices that steer people into actions they would not knowingly choose.
Illustrative example
A preselected add-on increases the price without a clear choice.
Dashboard dependency
Over-reliance on measurement tools as a substitute for strategic judgment. When teams optimize the most visible numbers without testing their meaning, the dashboard can improve while business or customer outcomes stagnate.
Illustrative example
A team adjusts spending in response to hourly click-through-rate changes but never asks whether the campaign reaches the right people or helps them make a useful decision.
Data accuracy
How closely information reflects the person or event it describes.
Illustrative example
Correct an outdated employer before using a job-title segment.
Data agency
A person’s practical ability to understand and influence uses of information about them.
Illustrative example
Let a customer inspect and correct their preferences.
Data completeness
The presence of information required for a specified use.
Illustrative example
A lead lacks the permission source needed for an email campaign.
Data freshness
How recently information was verified relative to its intended use.
Illustrative example
Recheck purchasing authority before a renewal conversation.
Data lineage
The path data takes through transformations and systems.
Illustrative example
Trace a dashboard total back through imports and deduplication.
Data minimization
Collecting and retaining only information needed for a stated purpose.
Illustrative example
A booking form does not request a birth date it never uses.
Data provenance
The record of where information came from and how it was produced.
Illustrative example
Keep the original source and collection date beside a contact record.
Decision rights
Explicit authority to make a particular decision.
Illustrative example
Name who can approve a refund without another escalation.
Deduplication
Identifying and resolving records representing the same entity.
Illustrative example
Merge duplicate accounts while preserving source history.
Dirty data
In data-quality work, dirty data is inaccurate, incomplete, duplicated, inconsistent, or outdated. Responsible data use also requires an established source, permitted uses, and permission history. Poor quality and unclear permission create different problems, and both deserve attention.
Illustrative example
A sales team buys a contact list with duplicate records, invalid addresses, and outdated job titles. The list also lacks a clear permission history, so the team cannot assume it is suitable for a campaign.
Related reading
Drip pricing
Revealing additional charges after an attractive initial price has drawn a buyer in.
Illustrative example
A ticket’s mandatory fees appear only near payment.
E
Emotional needs
Motivations such as safety, belonging, or autonomy that influence choices in context.
Illustrative example
Two buyers of the same product may want reassurance and independence.
Emotional resonance
The degree to which a message or experience connects with a relevant emotional need.
Illustrative example
Test reassurance against status language with the intended audience.
Extraction-based growth
Growth that prioritizes short-term value taken from customers, employees, or data without a fair reciprocal exchange. It treats people primarily as inputs to revenue rather than participants in an ongoing relationship.
Illustrative example
A platform monetizes detailed user behavior while giving people little understanding of the exchange or meaningful control over how their information is used.
Related reading
F
Fake urgency
Manufactured scarcity or time pressure that misrepresents a buyer’s actual choices. It pushes a quick decision through a deadline or shortage that is not genuine and can undermine trust in future offers.
Illustrative example
A store displays “Only three left” despite having ample stock, or restarts the same supposedly final countdown every time a customer returns.
Source reference
First-party data
Information an organization obtains through its direct interactions with people.
Illustrative example
A customer submits an order to the store they buy from.
G
The grift economy
A critical label for business practices that rely on manipulation, opaque terms, or unnecessary friction to extract value. It asks whether the customer receives genuine value and can make a clear, informed choice.
Illustrative example
A subscription is easy to start but deliberately difficult to cancel, while unavoidable charges only appear at the final checkout step.
Related reading
H
Holdout group
A comparable group deliberately excluded from an intervention to estimate its effect.
Illustrative example
Keep a random customer group out of a discount campaign.
Human oversight
People with authority and information review consequential automated decisions.
Illustrative example
A reviewer can reject an AI-generated exclusion recommendation.
I
Identity resolution
Connecting records believed to refer to the same person or organization.
Illustrative example
Validate a match before joining an anonymous visit to a customer profile.
The Impression Delusion
Treating impression volume as proof of unique reach, attention, engagement, or revenue. Viewability measures an opportunity to see an ad; it does not establish attention or business impact.
Illustrative example
A brand reports record impressions while sales decline. The delivery count is real, but it does not show how many people paid attention or what the advertising contributed.
Source reference
Incentive corruption
When incentives reward the appearance of performance at the expense of actual performance. Teams learn to optimize whatever the compensation plan or scorecard rewards, even when that behavior weakens the underlying business goal.
Illustrative example
A marketing team rewarded for marketing-qualified lead volume lowers its qualification threshold. The lead count rises while the share of leads becoming customers falls.
Incrementality
The additional outcome caused by an activity compared with what would have happened without it.
Illustrative example
Compare purchases in a randomized campaign group with an untreated group.
Inferred data
Information estimated from other signals rather than supplied or directly observed.
Illustrative example
An age band guessed from browsing can be wrong.
Informed consent
A freely made choice based on understandable information about the proposed use.
Illustrative example
Explain partner sharing before asking a person to opt in.
L
Lagging indicator
A measure of an outcome after it has occurred.
Illustrative example
Annual customer retention reflects earlier experiences.
Leading indicator
A measure that may signal a later outcome and needs validation.
Illustrative example
Check if product activation predicts renewal in your own records.
M
Marketing accountability
Marketing accountability means respecting customers, delivering value consistently, and explaining marketing decisions with credible evidence. It connects investment and claims to business outcomes while being candid about attribution limits.
Illustrative example
A VP of Marketing connects brand investment to pipeline and closed revenue where the evidence supports it, explains what cannot be attributed confidently, and looks beyond impressions and click-through rates.
Related reading
Marketing mix modeling
Estimating contributions of marketing and other factors from aggregate results over time.
Illustrative example
A model includes price changes and seasonality alongside media spending.
Meaningful disclosure
Information presented clearly enough and early enough to affect a choice.
Illustrative example
Explain who provides the service before a buyer submits a referral form.
Model drift
Changes that reduce how well a model fits current conditions.
Illustrative example
A scoring model trained on last year’s buyers fails on a new segment.
O
Opportunity cost
The value of the best alternative forgone by a decision.
Illustrative example
A campaign consumes staff time that could repair onboarding.
P
Payback period
The time required to recover an investment through its returns.
Illustrative example
Estimate how many months of customer margin repay acquisition spending.
Performative marketing
Marketing that presents the appearance of progress, authenticity, or responsibility without the operating behavior to support it. Strong campaign results do not establish that a company is keeping the promise behind its message.
Illustrative example
A brand promotes substantial sustainability progress but cannot provide evidence of the operational changes behind that claim. The campaign draws attention, but its central promise remains unsupported.
Related reading
Permission record
A retained account of the scope, source, and timing of a permission.
Illustrative example
Store the exact opt-in wording with its timestamp.
Pipeline quality
How well opportunities match buyer needs, fit criteria, and realistic progress.
Illustrative example
Review stalled opportunities instead of celebrating their total value.
Positioning
The role an offer seeks to occupy for a particular audience relative to alternatives.
Illustrative example
Explain why a small business would choose your service over doing nothing.
Promise-delivery gap
The distance between a stated commitment and the experience delivered.
Illustrative example
A support promise says one owner while customers repeat their case.
Pseudonymization
Replacing direct identifiers with a code while allowing linkage through separately held information.
Illustrative example
Keep the lookup key outside the analyst’s dataset.
Psychographic segmentation
Grouping people by motivations, attitudes, or values rather than demographics alone.
Illustrative example
Compare people seeking control with people seeking belonging.
Purpose limitation
Restricting information use to defined and appropriate purposes.
Illustrative example
A support email is not automatically permission for marketing.
Q
Qualified lead
An inquiry meeting explicit criteria for fit and readiness.
Illustrative example
An email address alone does not qualify an executive-search buyer.
R
Retention
The share of a defined customer cohort remaining active after a period.
Illustrative example
Track first-year renewals separately from long-standing accounts.
Retention schedule
Rules defining how long different information is kept and when it is removed.
Illustrative example
Delete obsolete prospect records after the agreed period.
Return on ad spend
Revenue attributed to advertising divided by advertising spend.
Illustrative example
A high reported ROAS may exclude returns and organic sales.
S
Sample size
The number of observations used in an analysis.
Illustrative example
Twenty replies may be too few to compare several customer segments.
Second-party data
Another organization’s first-party data shared through a direct arrangement.
Illustrative example
A partner shares a permissioned dataset under a specific agreement.
Selection bias
Distortion caused by the way people or records enter a sample.
Illustrative example
A survey of newsletter fans cannot represent all former customers.
Sensitive data
Information whose misuse could cause heightened harm or requires special handling.
Illustrative example
Limit access to health information collected for support.
Service recovery
The actions taken to repair a failed experience and address its effects.
Illustrative example
Restore access, explain the cause, and review compensation.
Shadow AI
AI tools used outside an organization’s approved processes and oversight.
Illustrative example
A staff member uploads client files to an unapproved assistant.
Shadow IT
Technology used outside established organizational oversight.
Illustrative example
A team creates a spreadsheet service without access review.
Shrinkflation
Reducing a product’s size or quantity without a corresponding price reduction, so the customer pays more per unit. Reducing quality is a related but different practice. Whether the change is communicated clearly is a separate accountability question.
Illustrative example
A package falls from 500 grams to 450 grams at the same shelf price. A shopper who compares the price per gram can see the increase, even if the packaging looks familiar.
Source reference
Sponsorship conflict
A tension between funding relationships and independent judgment.
Illustrative example
An event must explain how it reviews a sponsor it also critiques.
Statistical power
The probability that a test detects an effect of a specified size when it exists.
Illustrative example
Plan enough participants before testing a small conversion change.
The Surveillance Economy
A business environment in which tracking, profiling, and monetizing people’s behavior become central sources of value. It raises questions about meaningful choice, transparency, and the value people receive for their information.
Illustrative example
Someone downloads a free app and later discovers that its location data supports advertising profiles across other services. The collection was buried in terms they did not understand.
Survivorship bias
Drawing conclusions only from cases that remain visible while ignoring those that disappeared.
Illustrative example
Study failed campaigns alongside award winners.
Synthetic data
Artificial records generated to resemble selected properties of real data.
Illustrative example
Use simulated records for a demo without claiming they are customer evidence.
T
The Template Trap
The pattern of choosing familiar campaign structures and best practices before understanding the problem. Templates can organize work, but they cannot replace original thinking about a specific buyer, situation, and goal.
Illustrative example
A team uses the same intake document for every launch. It asks for a target persona, budget, and timeline, but never asks what the buyer is trying to solve.
Third-party data
Information obtained from parties outside the direct relationship with the person.
Illustrative example
A broker sells audience records collected from multiple sources.
Trust debt
The accumulated cost of practices that erode customer confidence over time. Repeated gaps between a promise and an experience can make future claims harder to believe, relationships harder to sustain, and trust more costly to rebuild.
Illustrative example
A brand runs countdown timers that repeatedly reset. Customers learn to ignore them. When a genuinely time-limited offer arrives, those customers have little reason to believe it.
Related reading
Trust repair
Evidence and action that rebuild confidence after a failure.
Illustrative example
Correct a claim and publish a remedy people can use.
V
Value proposition
A specific account of the useful outcome an offer provides and why it is credible.
Illustrative example
State the problem solved, the benefit, and supporting evidence.
Vanity metrics
Numbers that look impressive but offer little evidence of progress toward a stated business or customer goal. Raw impressions, lead volume, and follower counts can become vanity metrics when reported without quality, context, or a connection to outcomes.
Illustrative example
A campaign generates two million impressions but no qualified pipeline. The dashboard is green. The team still cannot explain what changed for the business.
Vendor due diligence
Evaluating a supplier’s practices and fitness before relying on it.
Illustrative example
Check a research vendor’s sources, permissions, and sample controls.
Z
Zero-party data
Information someone intentionally shares with a brand, such as preferences, goals, or purchase plans. These are stated signals; they still need context and may change over time.
Illustrative example
A quiz asks someone about their goals and preferences. They knowingly provide answers in exchange for a personalized recommendation, with a clear explanation of how the answers will be used.
Source reference
Put the language to work
Which of these shows up
in your marketing?
Bring the decision, dashboard, or customer experience you need to make sense of.
Talk with Jay about marketing accountabilityExplore the Carousel of TruthExplore Jay’s writing