Skip to content
flayv.
Glossary

54 terms, defined without the hand-waving.

Every entry opens with a single self-contained sentence, then says the thing the definition alone leaves out. No figures here — anything numeric belongs in an article where it can carry a source.

Unit economics

Blended CAC
Blended customer acquisition cost is total acquisition spend across every channel divided by total new customers, regardless of which channel gets the credit.
Its value is precisely that it ignores attribution. When platform-reported numbers disagree, blended CAC is the figure the bank account agrees with.
SeeCustomer acquisition cost,Marketing efficiency ratio
Cohort analysis
Cohort analysis groups customers by when or how they were acquired and tracks each group's behaviour over time, rather than averaging all customers together.
It is the only way to see whether newer customers are behaving worse than older ones — a deterioration that a blended average hides completely.
SeeLifetime value,Retention rate
Contribution margin
Contribution margin is revenue minus the variable costs of delivering it, before fixed overhead.
It sets the ceiling on what can be paid to acquire a customer. Acquisition planning that ignores it optimises toward growth that loses money per unit.
SeePayback period,Lifetime value
Cost per acquisitionCPA
Cost per acquisition is the advertising cost of generating one defined conversion event, which may be a sale, a lead, an install or any other chosen action.
The term is only meaningful once the conversion event is named. A CPA on form fills and a CPA on closed revenue describe different worlds.
SeeCustomer acquisition cost,Cost per lead
Cost per leadCPL
Cost per lead is the advertising cost of generating one enquiry, before any judgement about whether that enquiry is worth working.
Trivially improved by making the lead worse. Cost per accepted lead is the version that tracks revenue.
SeeAccepted lead,Cost per acquisition
Customer acquisition costCAC
Customer acquisition cost is the total cost of acquiring one new customer, calculated by dividing all acquisition spend in a period by the number of new customers gained in that period.
Whether it includes salaries, agency fees and creative production changes the number substantially. Agree the inclusions before comparing it to anything.
SeeBlended CAC,Payback period,Lifetime value
Lifetime valueLTV or CLV
Lifetime value is the total contribution margin a customer is expected to generate across their entire relationship with a business.
In categories with skewed value distributions, the mean is dominated by a small tail. A conservative percentile is a safer basis for setting acquisition budget.
SeeContribution margin,Cohort analysis
LTV:CAC ratio
The LTV:CAC ratio compares expected lifetime value against the cost of acquisition, expressing how many times over a customer repays what they cost to win.
It says nothing about timing. A healthy ratio with a long payback period can still starve a business of working capital.
SeeLifetime value,Payback period
Marketing efficiency ratioMER
Marketing efficiency ratio is total revenue divided by total advertising spend over the same period, measured across the whole business rather than per campaign.
Unlike return on ad spend, it cannot be inflated by attribution choices, because both numbers come from your own accounts.
SeeReturn on ad spend,Blended CAC
Payback period
Payback period is the time it takes for the contribution margin from a customer to equal what it cost to acquire them.
It governs how fast a business can grow without external funding. A shorter payback recycles cash into the next cohort sooner, which can beat a lower acquisition cost outright.
SeeCustomer acquisition cost,Contribution margin
Return on ad spendROAS
Return on ad spend is revenue attributed to advertising divided by the cost of that advertising, usually as reported by the advertising platform itself.
Useful for comparing two campaigns inside one platform on one day. Close to meaningless as a business target, because the platform chose the base it is measured on.
SeeMarketing efficiency ratio,Attribution

Measurement & attribution

Attribution
Attribution is the assignment of credit for a conversion to one or more marketing touchpoints that preceded it.
Every attribution model is an assumption about causation, not a measurement of it. Two defensible models can disagree by a wide margin on the same data.
SeeLast-click attribution,Incrementality,Multi-touch attribution
Attribution window
An attribution window is the period after an ad interaction during which a conversion will still be credited to that interaction.
Two platforms with different windows will report different conversion counts from identical activity. Comparing them without normalising the window is meaningless.
SeeAttribution,Last-click attribution
Holdout test
A holdout test measures incrementality by deliberately withholding advertising from a randomly selected or matched group and comparing outcomes against the exposed group.
It costs real revenue for the duration of the test. That cost is usually smaller than a year of scaling a channel that was taking credit rather than creating demand.
SeeIncrementality,Matched market test
Incrementality
Incrementality is the share of conversions that happened because of a marketing activity and would not have happened without it.
It is the only question attribution is really trying to answer, and the only one a controlled holdout test answers directly.
SeeHoldout test,Attribution
Last-click attribution
Last-click attribution assigns all credit for a conversion to the final marketing touchpoint before it.
It systematically favours channels that appear late in the journey — brand search, retargeting, coupon affiliates — over the ones that created the demand.
SeeAttribution,Incrementality
Marketing mix modellingMMM
Marketing mix modelling uses statistical regression on aggregate historical data to estimate how each marketing input contributed to sales.
It needs no user-level tracking, which is why it returned to favour under privacy restrictions. It also needs a long history and genuine variation in spend to say anything reliable.
SeeAttribution,Incrementality
Matched market test
A matched market test compares outcomes between regions with similar historical performance, where advertising is changed in one set and held constant in the other.
Useful where individual-level holdouts are impossible. Its validity depends entirely on how well the markets were matched before the test began.
SeeHoldout test,Incrementality
Minimum detectable effectMDE
Minimum detectable effect is the smallest difference between variants that a test is capable of detecting reliably, given its sample size.
Deciding it before the test starts prevents the most common failure in experimentation: running a test that never had the traffic to answer the question.
SeeA/B test,Statistical significance
Modelled conversions
Modelled conversions are conversions estimated statistically by an advertising platform when the actual event could not be observed, typically because of consent or tracking restrictions.
They are estimates presented alongside observed conversions, usually without visible distinction. Reporting that does not separate them overstates certainty.
SeeConsent mode,Server-side tracking
Multi-touch attributionMTA
Multi-touch attribution distributes credit for a conversion across several touchpoints according to a chosen rule or model.
Its accuracy depends on observing the full journey, which privacy restrictions and cross-device behaviour increasingly prevent.
SeeAttribution,Marketing mix modelling
Server-side tracking
Server-side tracking sends conversion data to advertising platforms from a server the advertiser controls, rather than from the visitor's browser.
It improves data completeness where browser restrictions block client-side collection. It does not remove the need for consent — it changes the transport, not the legal basis.
SeeModelled conversions,Consent mode
Statistical significance
Statistical significance is the degree of confidence that an observed difference between variants reflects a real effect rather than random variation.
Significance says nothing about size. A statistically significant improvement can be too small to matter commercially, and frequently is.
SeeA/B test,Minimum detectable effect

Channels & buying

Accepted leadSales accepted lead, SAL
An accepted lead is an enquiry that the sales team has reviewed and agreed is worth pursuing.
The correct optimisation target for lead generation. Optimising raw lead volume instead reliably produces enquiries nobody works.
SeeCost per lead,Lead scoring
Affiliate marketing
Affiliate marketing pays third-party partners a commission for driving a defined action, usually a sale or a qualified lead.
Its central risk is paying full commission for customers who were already buying. Differentiated commission rates are the structural fix.
SeeLast-click attribution,Incrementality
Frequency capping
Frequency capping limits how many times a single person may be shown a particular advertisement within a defined period.
Uncapped retargeting reaches the point of irritation faster than most reporting reveals, because the same reporting counts the annoying impressions as reach.
SeeRetargeting,Programmatic advertising
Lead scoring
Lead scoring assigns a numerical value to an enquiry based on attributes and behaviour, to estimate how likely it is to convert.
It is only as good as the outcome data behind it. A score never validated against closed revenue is an opinion with a number attached.
SeeAccepted lead,Cost per lead
Native advertising
Native advertising is paid placement designed to match the form and feel of the surrounding editorial content.
Performance networks such as PropellerAds and RollerAds operate largely in this space, reaching inventory the major platforms do not, at different prices and different quality variance.
SeeProgrammatic advertising,Paid media
Performance marketing
Performance marketing is marketing managed against measurable business outcomes such as acquisition cost or payback period, rather than against reach or awareness.
The defining feature is the target, not the channel. Any channel can be run this way if its contribution can be measured credibly.
SeePaid media,Customer acquisition cost
Programmatic advertising
Programmatic advertising is the automated buying and selling of ad inventory through real-time auctions rather than direct negotiation with publishers.
It buys reach efficiently and makes inventory quality hard to see. Placement transparency is worth insisting on.
SeeNative advertising,Frequency capping
RetargetingRemarketing
Retargeting serves advertising to people who have already interacted with a brand, typically by visiting its website.
The channel most prone to overstating its contribution, because its audience is by definition already interested. A holdout test is the honest way to size it.
SeeIncrementality,Frequency capping

Creative & conversion

A/B testSplit test
An A/B test randomly divides traffic between two or more variants and compares their performance against a single chosen metric.
Valid only when assignment is genuinely random and the sample is large enough for the effect being sought. Most inconclusive tests were underpowered from the start.
SeeStatistical significance,Minimum detectable effect
Conversion rate optimisationCRO
Conversion rate optimisation is the systematic improvement of the proportion of visitors who complete a desired action.
A conversion rate improvement that lowers lead quality is not an improvement. The metric has to be tied to something downstream to mean anything.
SeeA/B test,Landing page
Creative fatigue
Creative fatigue is the decline in an advertisement's performance as the target audience sees it repeatedly.
Often confused with audience saturation. The distinction matters: new creative fixes fatigue, while a larger audience is the only fix for saturation.
SeeFrequency capping,Creative testing
Creative testing
Creative testing compares advertising variants under controlled conditions to establish which performs better against a defined outcome.
Testing many variants of the same idea measures execution. Testing genuinely different ideas measures strategy, and usually returns more.
SeeA/B test,Creative fatigue
Friction
Friction is anything in a conversion path that increases the effort required to complete it, such as additional form fields or verification steps.
Not uniformly bad. Friction placed after the value is understood filters mostly in the right direction; friction placed before it repels everyone equally.
SeeConversion rate optimisation,Accepted lead
Landing page
A landing page is the page a visitor arrives at after clicking an advertisement, built for one specific audience and one specific action.
Frequently the largest single lever on acquisition cost, and the one most often treated as a downstream detail of the media buy.
SeeConversion rate optimisation,Message match
Message match
Message match is the degree of consistency between the promise made in an advertisement and the content of the page it leads to.
A mismatch shows up as a high bounce rate that looks like a traffic quality problem and is actually a page problem.
SeeLanding page,Conversion rate optimisation

Lifecycle & retention

Churn rate
Churn rate is the proportion of customers who stop being active over a defined period.
The inverse of retention, and the more intuitive framing for subscription businesses. Both hide the same thing: churn is rarely uniform across cohorts.
SeeRetention rate,Cohort analysis
Reactivation
Reactivation is marketing directed at lapsed customers with the aim of restoring them to active status.
Usually far cheaper than new acquisition, and usually undercounted, because the revenue often lands in whichever channel touched them last.
SeeChurn rate,Attribution
Repeat purchase rate
Repeat purchase rate is the proportion of customers who buy more than once within a defined period.
In transactional businesses it is the single most powerful lever on acquisition budget, and almost always cheaper to move than media efficiency.
SeeRetention rate,Lifetime value
Retention rate
Retention rate is the proportion of customers who remain active over a defined period.
It sets the ceiling on lifetime value, and therefore on what can be paid to acquire a customer. Improving it raises the acquisition ceiling for every channel simultaneously.
SeeChurn rate,Lifetime value
RFM segmentation
RFM segmentation groups customers by how recently they purchased, how frequently they purchase, and how much they spend.
Crude, decades old, and still outperforms most elaborate segmentation schemes for deciding who to contact next.
SeeCohort analysis,Reactivation

Tell us the number you are trying to move.

Describe what you are spending and what it has to return, and we will tell you whether we are the right people.

Start a conversation