Glossary

A fair A/B test changes one thing at a time, decides the success measure and sample size in advance, and runs until the result is statistically reliable. Stopping early when one version looks ahead is the most common way tests mislead.

AMEC's Barcelona Principles have said since 2010 that AVEs are not the value of communication, and AMEC campaigns to end their use: they confuse cost with value, ignore whether coverage was positive or reached the right people, and cannot be applied where no advertising is sold.

Commissions are usually a percentage of the sale or a fixed amount, paid only for tracked results within the attribution window. Affiliate links are a commercial relationship, so creators and publishers must disclose them.

WIPO describes ambush marketing as damaging genuine sponsors' investment and, with it, the funding of events. Ambushes range from direct use of protected marks to indirect campaigns that evoke an event without naming it, and many host countries pass special laws for major events, so legal advice is needed before referring to one.

Simple rules give all the credit to the first or last click; data-driven and multi-touch models share it out. The IAB's attribution primer notes that every model is a simplification: attribution assigns credit, it does not prove that a touchpoint caused the sale.

Common settings are seven or thirty days after a click and one day after a view, but each platform defaults differently, so totals across platforms often count the same sale twice. Agree windows in advance and keep them consistent when comparing channels.

The principles ask for clear goals set before measurement, for measuring outcomes and impact rather than just outputs, for qualitative as well as quantitative evidence, and for honest, transparent methods. They also reject advertising value equivalents as a measure of communication value.

In The Long and the Short of It (IPA, 2013), Les Binet and Peter Field showed that brand building produces small immediate effects but larger, longer-lasting growth, and works best alongside sales activation. Their often-quoted guide is roughly 60% of investment on brand building and 40% on activation, a balance that varies by category.

Brand equity lives in memory (how well known the brand is and what it is linked to) and shows up in behaviour (preference, price premium, loyalty). It is built slowly by consistent communication and can be run down quickly by neglect or a broken promise.

Brand lift studies survey an exposed group and a control group and report the difference, as in Google Ads' Brand Lift. Results are only as good as the sample size and the questions, and a platform's own study measures only that platform.

A brand platform changes rarely and sits above any single campaign; message houses, creative briefs and guidelines are derived from it. A useful one is short, specific enough to say no to things, and shared well beyond the marketing team.

A clear call to action names one action, says what happens next and is easy to find on every screen size. Brand-building work may have no explicit call to action; activation work almost always does.

Churn is easy to define for subscriptions and harder for brands where customers simply buy less often, so the rule for “lapsed” should be agreed first. Spotting early warning signs lets CRM teams act before customers leave.

Community managers are often the first to spot a complaint turning into a crisis, so they need clear house rules, escalation paths and the authority to answer quickly. The aim is a place where members help each other, not just a channel for announcements.

CTV combines the big screen and attention of television with digital buying and targeting, often programmatically. Measurement is less standardised than for linear TV, so check how reach is deduplicated with broadcast and how impressions are verified.

The CNIL's guidance sets the bar many European teams follow: consent by a clear positive act, refusing as easy as accepting, withdrawal possible at any time, and proof kept. In advertising, IAB Europe's Transparency and Consent Framework is a common standard for passing those choices to vendors.

Contextual targeting needs no tracking of individuals, which makes it a privacy-friendly option as cookies and identifiers decline. Shared category lists such as the IAB Tech Lab Content Taxonomy let buyers and sellers describe content in the same terms, including on CTV.

Conversion rate is only comparable when the action and the starting point are the same, for example purchases per session or per click. It reflects the offer, the landing page and the audience as much as the advertising.

Coupons drive trial and measurable short-term sales, and unique codes show which channel or partner drove each redemption. Used too often they train customers to wait for a discount and can erode the brand's price position.

CPA is total cost divided by the number of actions. It depends entirely on what counts as an action and on which channel gets the credit, so the definition and the attribution rules should be agreed before targets are set.

CPC is the usual model in paid search and many social and display platforms, often set through auctions. A cheap click that never converts is expensive, so CPC is best read alongside conversion rate and cost per acquisition.

CPL is common in B2B and high-value consumer categories. It rewards volume, so it should be tracked with lead quality, such as the share of leads that sales accepts or that become customers.

CPM equals cost divided by impressions, multiplied by 1,000. A low CPM is only a bargain if the impressions are viewable, reach the right people and do their job; viewable CPM (vCPM) prices only impressions that met the viewability standard.

A crisis plan lists likely scenarios, the crisis team and its contacts, approval steps, holding statements and a monitoring routine. It should cover paid and owned channels too, since scheduled ads and automated messages may need to be paused at once.

CTR shows whether a message prompts action among those who see it, which makes it useful for comparing creative versions. It says little about brand effects, and accidental or fraudulent clicks can inflate it.

CAC is broader than CPA: it includes salaries, tools and agency fees as well as media. It is most useful next to customer lifetime value, because a high CAC can be fine for customers who stay and spend for years.

Advocacy starts with a product and service worth recommending; programmes only make it easier to share. When advocates receive rewards or free products, that connection must be disclosed, and reviews must never be faked or selectively suppressed.

The CDP Institute's definition centres on three things: a unified, persistent customer record, identity resolution across sources, and access for other tools such as email, ads and analytics. Some CDPs keep their own copy of the data; warehouse-native ones work on data held in the company's data warehouse.

A journey map is drawn from the customer's point of view, not the organisation chart: what they are trying to do, what they feel and which touchpoints they meet. Real journeys loop and skip steps, so a map is a planning aid rather than a prediction.

Lifetime value combines how much customers spend, how often and for how long, usually by segment. Comparing it with acquisition cost shows how much a brand can afford to spend to win a customer, and which customers are worth keeping.

In marketing, CRM usually means lifecycle communication to known customers (welcome, onboarding, cross-sell, win-back) driven by the data in the CRM system. The CRM is the record of known people that sales and service work from; a CDP often feeds it with behavioural data.

The FTC's 2022 report Bringing Dark Patterns to Light describes common forms: ads disguised as content, fake countdown timers, hard-to-cancel subscriptions, hidden fees and cookie banners that steer people to accept. Regulators in the US, EU and UK treat many of these as unfair or deceptive, and under the GDPR a consent obtained this way is not valid.

Advertisers use clean rooms with publishers, retailers and platforms for audience overlap, targeting and measurement. The IAB Tech Lab's guidance stresses that a clean room relies on a combination of privacy-enhancing techniques and controls. Using one does not remove the need for a lawful basis to process the data.

A DSP holds the targeting, bids, budgets and frequency caps, and decides in milliseconds whether each available impression is worth buying. When choosing one, ask about its fees, the inventory it reaches and how transparent its reporting is.

Unlike printed posters, DOOH can change messages by time of day, weather or events, and can be bought programmatically. The IAB's DOOH glossary sets out shared terms, such as how a play differs from an audience impression, which matter when comparing quotes.

Direct mail is tangible and tends to be kept, which suits high-value offers, loyalty moments and audiences that are hard to reach online. Addressed mail uses personal data, so data protection rules and suppression lists apply, and it can be tracked with unique codes or URLs.

The Ehrenberg-Bass Institute, through Jenni Romaniuk's work, argues that brands need assets that are both widely known and unique to them, so that every ad and pack is correctly branded. Assets are memory structures: they fade without consistent use and should be managed and protected over the long term.

M3AAWG, the anti-abuse industry group, ranks confirmed opt-in as the strongest form of permission because it stops typos and addresses entered by someone else from joining a list. It also leaves a record that helps prove consent.

Earned media is usually trusted more than advertising because a third party chose to say it, but the brand cannot control what is said or when. It is won through news value, relationships with journalists and creators, and products worth talking about.

EMV is a close relative of advertising value equivalency, and AMEC, the communication measurement body, treats it with the same caution: it measures the cost of media space, not the value or effect of the communication. Report what the coverage achieved, such as reach in the target audience and changes in awareness or behaviour, instead.

SPF lists the servers allowed to send for a domain, DKIM signs each message, and DMARC tells receivers what to do when checks fail and sends reports back. Gmail requires all three for bulk senders; setting them up is usually a job for IT and the email platform together.

Mailbox providers set the rules. Gmail, for example, requires bulk senders to authenticate with SPF, DKIM and DMARC, to offer one-click unsubscribe and to keep the rate of messages reported as spam below 0.1%, never reaching 0.3%. Sending only to people who opted in and removing inactive addresses protect deliverability more than any trick.

Embargoes give reporters time to prepare a fuller story and help coverage land at once. They rest on trust, so state the time and time zone clearly, ask for agreement before sharing, and assume anything sent may still leak.

Posts from people usually reach and persuade more than posts from company accounts, which is why employee advocacy supports both reputation and recruitment. It should be voluntary, come with clear social media guidance, and make the employment relationship clear when staff promote products.

Tim Ambler and Simon Barrow defined the employer brand in 1996 as the package of functional, economic and psychological benefits provided by employment and identified with the employing company. It is usually expressed in an employee value proposition and should be consistent with the customer brand, because employees deliver it.

Analysis of the IPA Databank by Les Binet and Peter Field found that brands whose share of voice exceeds their market share tend to grow, and those below it tend to shrink. It is a useful rule of thumb for budgeting, though the IPA itself notes it is a single-factor model and that the effect varies by category.

An experience reaches fewer people than media but leaves stronger memories, and it can be extended to many more through content, PR and social sharing. Plan the before and after as carefully as the day itself, including how contacts are collected with consent.

First-party data is usually the most accurate and the most durable as third-party tracking declines, but collecting and using it still needs a lawful basis, clear information and, for many uses, consent. Its value comes from joining it up across touchpoints, often in a CRM or CDP.

Some repetition helps a message stick; too much wastes money and irritates people. Digital platforms let advertisers cap frequency, but caps set separately in each channel do not stop the same person seeing an ad many times across them.

For marketers, the GDPR means having a lawful basis (such as consent or legitimate interests) for each use of personal data, telling people clearly what happens to it, honouring their rights to access, object and erase, and keeping data no longer than needed. It applies to organisations outside the EU that target people in it; the UK keeps a similar UK GDPR, and rules on cookies and electronic marketing sit alongside it.

The UK CMA's Green Claims Code asks that claims be truthful, clear, complete, fair in comparison, based on the full life cycle and substantiated, and warns against vague terms like “eco” or “sustainable”. In the EU, Directive (EU) 2024/825 tightens the rules on generic environmental claims and sustainability labels; keep the evidence before the claim goes out.

Reaching 50% of the audience an average of four times gives 200 GRPs. GRPs come from television planning and are still used to buy and compare TV, radio and out-of-home; when calculated on a specific target group they are often called target rating points (TRPs).

The holdout group shows what would have happened anyway, so the gap between the groups is the incremental effect. The groups must be chosen at random and kept large enough, and the test must run long enough, for the difference to be trusted.

A hybrid event works when the online audience gets its own programme, host and ways to take part, not just a camera at the back of the room. It also produces registration and engagement data, which needs consent and a plan for follow-up.

IAB and Media Rating Council (MRC) guidelines set how digital ad impressions are counted, and the MRC's viewability standard adds a stricter count: at least half the ad in view for one second for display and two seconds for video. Ask whether a report counts served or viewable impressions, and whether invalid traffic has been removed.

Incrementality is measured by comparing people or regions that were exposed with similar ones that were not, as in the lift studies ad platforms offer. It often shows that attributed results overstate what retargeting and branded search really add.

The FTC asks for the disclosure to sit with the endorsement in plain words such as “ad” or “sponsored”, not hidden in a bio or a pile of hashtags; the UK's ASA prefers an upfront label such as “Ad” and holds both brands and influencers responsible. Other markets, including France, the Gulf states and India, have their own rules, so check each country a campaign runs in.

IMC treats advertising, PR, digital, social, content, email, events and the rest as parts of one plan with shared objectives, one core message and coordinated timing, instead of separate campaigns run by separate teams. Each channel adapts the story to what it does best, but the promise, the look and the voice stay recognisably the same, so the channels reinforce one another.

A good KPI follows from the business objective, not from whatever a platform reports by default, and has a target and a time frame. Most plans need a mix: business outcomes (sales, share), brand measures (awareness, consideration) and a few channel metrics that explain the result.

Scores usually combine fit (role, company size, sector) with engagement (visits, downloads, event attendance). The threshold at which a lead passes to sales should be agreed between marketing and sales and checked against which leads actually closed.

The advertiser supplies a source, such as a customer list or site visitors, and the platform finds similar users, as Meta describes for its Lookalike Audiences. Uploading customer lists is processing of personal data, so it needs a lawful basis and clear privacy information.

A loyalty programme is also a data programme: members share who they are and what they buy in return for rewards. It earns its cost when it changes behaviour or yields useful insight, not merely by rewarding customers who would have stayed anyway.

Automation turns a plan into rules that run on their own: welcome series, lead nurturing, reminders and win-back messages across email, SMS and apps. It works best on top of clean data and clear consent, and every automated flow needs an owner who reviews it.

The funnel helps match objectives and metrics to stages: reach and recall at the top, clicks and leads in the middle, sales and retention at the bottom. Its limit is that people do not move in a straight line, and the brand building done at the top drives much of what happens below.

MMM works on aggregate weekly or regional data, so it needs no personal data and covers offline media such as TV and outdoor. Open-source tools such as Google's Meridian have made it more accessible; results are most trusted when calibrated with experiments.

Good media relations means knowing what each outlet covers, offering useful stories, data and access, answering quickly and honestly, and accepting that journalists decide what to publish. It is the main route to earned media.

The term comes from Byron Sharp and the Ehrenberg-Bass Institute, which link it to category entry points: the needs and situations that send people to a category. Brands build it by being linked to many relevant entry points and by consistent distinctive assets, and it is measured by how many category buyers link the brand to those situations.

The message house keeps every channel, spokesperson and agency saying the same thing in their own words. It is written once from the brand platform and then adapted, not rewritten, for each audience and touchpoint.

On the WhatsApp Business Platform, templates are the only messages that can reach people outside the customer service window, and each is categorised as marketing, utility or authentication, which sets its price. Plan templates like any creative asset: on-brand, translated and approved before the campaign date.

MTA uses user-level journey data, with fixed rules (linear, time-decay, position-based) or statistical models. It sees only trackable digital touchpoints, and consent rules and the loss of cross-site identifiers have made those paths less complete, so many teams pair it with media mix modelling and experiments.

Because it blends in, native advertising must be clearly labelled as advertising. The FTC's guide warns that an ad not readily identifiable as one can be deceptive, and the ICC Code requires marketing to be recognisable as such, with labels prominent and close to the message.

NPS was created by Fred Reichheld with Bain & Company; scores run from -100 to +100. Bain stresses that the value lies in acting on the follow-up answers and closing the loop with customers, not in the number itself, which is easy to skew through who gets asked and when.

Open rates have become unreliable: Apple's Mail Privacy Protection loads remote content privately in the background, so many emails register as opened whether or not anyone read them. Clicks, conversions and replies are better guides to engagement.

Under the GDPR, consent must be freely given, specific, informed and unambiguous, so pre-ticked boxes do not count. In the UK, the ICO explains that email and text marketing to individuals needs consent or the narrow “soft opt-in” for existing customers, and every message must offer a simple way to opt out.

Owned media costs effort rather than media money and lasts as long as the brand keeps it up. It is where paid and earned attention can land and where a lasting relationship is built, which is why first-party data usually starts here.

Paid media gives control over the message, the audience and the timing, and it scales quickly, but the effect stops when the spending stops. In an integrated plan it often amplifies what owned and earned media have already proved works.

A persona is only as good as the research behind it: interviews, data and frontline knowledge, not guesses. Teams use personas to brief creative work and to map journeys, and update them when customers change.

Personalisation helps when it makes a message more relevant and hurts when it feels intrusive or gets details wrong. It should use data people expect the brand to use, stay within the brand's voice and be tested against a non-personalised version.

Created by Gini Dietrich of Spin Sucks for her agency and published in her 2014 book, the PESO Model© asks a team to plan the four media types together rather than one at a time, so that owned content feeds earned coverage, shared conversation and paid amplification. Spin Sucks allows its use with credit and the trademark symbol.

Positioning is a choice, which means deciding what the brand will not be. It is usually summarised in a short statement naming the target, the category, the main benefit and the reason to believe, and every campaign should be recognisably on it.

A release works when it carries real news for that outlet's readers and answers who, what, when, where and why in the first paragraph. It is a starting point for coverage, not a substitute for relationships with journalists.

Programmatic covers open auctions, private marketplaces and guaranteed deals, across display, video, audio, CTV and digital out-of-home. Its benefits are scale and precision; its risks are fees taken along the supply chain, fraud and ads appearing next to unsuitable content.

Reach counts people, not exposures, so it is always lower than impressions. Combining channels raises total reach only as far as their audiences do not overlap, which is why deduplicated cross-media reach is hard to measure and valuable to know.

Response rate is the classic direct marketing measure: responses divided by people contacted. Comparing it with a control group that was not contacted shows how many responses the campaign actually caused.

Retail media reaches people close to the point of purchase and can report sales, which makes it attractive to brands. Each network measures differently, so the IAB and MRC published retail media measurement guidelines to make results more comparable.

Retargeting usually shows strong attributed results because it reaches people already close to buying; holdout tests often reveal a smaller true effect. It depends on tracking that needs consent in many countries, and too much of it feels like being followed.

Retention is the mirror of churn and a main driver of lifetime value. It is calculated without counting customers gained during the period, so that acquisition does not hide losses.

Each customer gets a score on recency, frequency and monetary value, and the combinations show who the best customers are, who is slipping away and who is new. It is simple, needs only transaction data, and is a common basis for CRM and direct marketing selections.

A ROAS of 4 means four units of revenue for each unit spent. It uses revenue, not profit, and usually relies on platform attribution, so it can look strong for ads that mostly reached people who would have bought anyway; incrementality tests show how much of it is real.

ROI is usually incremental profit minus cost, divided by cost. Unlike ROAS it counts margin and the effects that would not have happened otherwise, and a fair ROI includes the longer-term effects of brand building, which short measurement windows miss.

Binet and Field's IPA research found that activation produces quick, measurable sales spikes that fade fast, and that relying on it alone leads to weaker long-term growth. It is what most click and conversion metrics capture, which is why it is easy to over-fund.

Segments can be built on demographics, needs, attitudes, behaviour or value to the business. A good segmentation is large enough to matter, reachable through real media and touchpoints, and different enough that each group needs a different approach.

Les Binet presented the metric at the IPA's EffWorks Global 2020: searches for the brand divided by searches for all brands in the category, typically from Google Trends. In his analysis it tracked and often led market share, though the IPA stresses this is correlation, not proof of cause.

Share of voice is compared with share of market to judge whether a brand is spending enough to hold or grow its position. In PR and social, the same idea is applied to mentions and coverage, which should be stated clearly so the two are not confused.

Shared media is the letter Gini Dietrich added to the older paid, earned and owned split. It sits between owned and earned: the brand runs its accounts, but reach depends on people engaging with and sharing what it posts, and on each platform's rules.

Organisations usually name different spokespeople for different topics and languages, and train them on the message house. In a crisis, one clear, credible voice matters more than seniority.

Activation includes advertising that features the partnership, hospitality, fan experiences, content, promotions and employee programmes. A logo on its own rarely changes much; sponsors usually need to budget for activation on top of the rights fee and plan it across their channels.

Rights typically include titles such as “official partner”, logo use, category exclusivity, branding in venue and broadcast, tickets and hospitality, and content access. The contract should spell out what is exclusive, where and for how long, and how the rights holder will act against ambush marketing.

The SSP is the publisher's side of the auction, connecting to DSPs and exchanges and applying the publisher's rules on price floors and advertisers. Buyers increasingly prefer shorter supply paths with fewer intermediaries, which keeps more of the budget for the media itself.

Strategic targeting decides who the brand is for; media targeting uses data such as location, interests, context or customer lists to reach them. Targeting too narrowly can cut cost per click while starving the brand of the future buyers it needs to grow.

Safari and Firefox block them by default. Google decided in 2025 to keep them in Chrome under user choice rather than phase them out, and to retire most Privacy Sandbox replacements; in Europe, rules such as those enforced by the CNIL already require consent before non-essential cookies are set.

IMC plans touchpoints, not just media: the invoice and the delivery driver communicate as much as the campaign. Mapping them shows where the story is told well, where it breaks, and which ones matter most at each stage of the journey.

UGC is persuasive because it comes from peers. A brand that wants to reuse it in its own channels or ads needs the creator's permission, should credit them, and must label it if the creator was paid or rewarded.

Google Analytics recommends always setting source, medium and campaign, and notes that values are case-sensitive, so a shared naming convention matters more than any tool. Untagged or inconsistently tagged links are a common reason channel reports do not add up.

Where positioning is about the brand's place in the market, the value proposition is about a specific offer for a specific segment. It is best tested with customers, in their own words, before it is turned into campaign copy.

The term was coined by Forrester to separate what people tell a brand (through preference centres, quizzes, polls and profiles) from what a brand infers from behaviour. People usually share it in exchange for something of value, such as better recommendations or loyalty rewards.