Marketing measurement glossary — terms, explained

What this category covers

Every number on a dashboard is the answer to a question somebody chose. This category is the vocabulary for saying which question, over what period, from how many observations — the parts that decide whether a metric means anything.

Impressions, clicks and positions
Share of voice and citation share
Coverage and sampling — how much of the ground was actually looked at
Windows, baselines and seasonality
Attribution and its limits
Confidence: when a change is a change and when it is noise

Why these words matter

A metric with no definition is decoration

Two dashboards showing different numbers for "visibility" are usually both right and measuring different things. The definition is the only way to tell.

Most disagreements are about the window

Last seven days against last month, rolling against calendar. Nothing changed except the frame, and the conclusion reversed.

It is what keeps a report honest

The terms here include the ones that limit a claim: sample size, coverage, margin. A report without them is not more confident, only less checkable.

Who needs this vocabulary

Growth teams

Deciding what to do next from numbers that arrive weekly and move slowly.

Agencies

Reporting to a client who will compare your figure to a different tool's.

In-house SEO

Defending a flat month that is, correctly, flat.

Product marketing

Category share and competitive presence, where the denominator is the whole argument.

Founders

Needing one honest number rather than nine plausible ones.

Analysts

Building the definition everybody else will quote.

Terms in this category232 terms

A31

B12

Baseline4 min read Bayesian Statistics6 min read Benchmarking7 min read Bias6 min read Bid Adjustment7 min read Bid Strategy6 min read Bidding6 min read BigQuery6 min read Bootstrapping6 min read Brand Hallucination5 min read Brand Lift6 min read Brand Safety6 min read

C34

D12

Data Harmonization8 min read Data Lake7 min read Data Warehouse6 min read Data-Driven Attribution5 min read Demographic Targeting7 min read Dependent Variable6 min read Device Targeting5 min read DiD6 min read Dirty Data6 min read DOE6 min read Double-Blind Experiment7 min read DSA6 min read

E8

EDA6 min read Effect Size6 min read Enhanced Conversions6 min read Event Tracking5 min read Exact Match5 min read Experiment Design6 min read Experimental Group6 min read External Validity7 min read

F8

Facebook Ads Manager7 min read Feed Optimization7 min read First-Click Attribution5 min read First-Party Cookie6 min read Floodlight5 min read Fraud Detection6 min read Frequency Capping6 min read Funnel Analysis7 min read

G7

GA46 min read Goodhart's Law14 min read Google Ads Editor5 min read Google Marketing Platform6 min read Google Optimize6 min read Grounding Gap5 min read Guardrail Metric6 min read

H3

Hallucination Monitoring5 min read Heatmap Analytics7 min read Hypothesis Testing8 min read

I6

Impression Share6 min read Impressions4 min read In-Market Audience5 min read Incrementality6 min read Independent Variable5 min read Internal Validity6 min read

K2

Keyword Match Type7 min read Kochava5 min read

L8

Last-Click Attribution7 min read Lift6 min read Linear Regression6 min read LLM Visibility Score4 min read Location Targeting6 min read Logistic Regression6 min read Lookalike Audience6 min read Lucas Critique7 min read

M14

Machine Learning Attribution6 min read Manual Bidding6 min read Matomo7 min read Max CPC7 min read Measurement Error6 min read Measurement Protocol5 min read Meta-Analysis6 min read Microsoft Advertising6 min read Mixpanel7 min read MMM5 min read Multi-Armed Bandit7 min read Multi-Channel Funnel6 min read Multiple Comparisons Problem7 min read Multivariate Testing6 min read

N5

Negative Keyword5 min read Non-Parametric Test8 min read North Star Metric7 min read Novelty Effect5 min read Null Hypothesis7 min read

O7

Observational Study6 min read Offline Conversion Tracking7 min read One-Tailed Test7 min read Optimized Targeting6 min read Optimizely6 min read Optional Stopping6 min read Outlier6 min read

P16

P-Value4 min read Paid Media5 min read Parametric Test7 min read Performance Max8 min read Pixel7 min read Placebo Effect3 min read Placement Targeting7 min read Population3 min read Post-Click Tracking6 min read Post-Hoc Analysis4 min read Power Analysis5 min read Privacy Sandbox6 min read Product Feed5 min read Prompt Share5 min read Prompt Volume4 min read Proxy Metric6 min read

Q1

Qualtrics4 min read

R12

Random Assignment4 min read RCT5 min read RDD6 min read Reach6 min read Regression Analysis5 min read Regression to the Mean7 min read Reliability3 min read Repeated Measures3 min read Retrieval Fit5 min read Retrieval Gap5 min read ROAS6 min read ROI6 min read

S23

T13

T-Test4 min read Target CPA7 min read Target ROAS7 min read Test Group5 min read Third-Party Cookie4 min read Time Decay Attribution6 min read Time Series Analysis5 min read Tracking Prevention5 min read Traffic Quality5 min read Treatment Effect5 min read Two-Tailed Test4 min read Type I Error5 min read Type II Error5 min read

U2

UA5 min read Unbiased Estimator6 min read

V5

Validity5 min read Value-Based Bidding6 min read Vanity Metric6 min read Variance4 min read VTC7 min read

W2

Web Analytics6 min read Website Tag7 min read

Z1

Z-Test4 min read

How to use this category

1

Learn the denominator

Start with share of voice and coverage. Almost every misread metric is a numerator without its denominator.

2

Fix your window before you read anything

Choose the period first, then look. Choosing it after seeing the data is how a flat quarter becomes a good week.

3

Ask what would falsify it

A number that cannot be wrong is not a measurement. Each term here says what its own limits are.

Common mistakes

Comparing two windows of different length and reading the difference as growth.
Reading a share as an absolute. Share of voice can rise while your mentions fall.
Averaging away the days with no data, which makes a gap look like a low reading.
Celebrating a jump smaller than the day-to-day noise of the same series.

Questions about this category

What is share of voice?

Your share of a defined set of mentions — yours plus the competitors you named — over a period you chose. Every part of that sentence is a decision, which is why two tools rarely agree.

How many observations before a number means something?

It depends on how much the thing being measured moves. The honest answer is on each metric's own page, with the sampling terms that describe it.

Why do two tools report different numbers for the same thing?

Usually a different set of questions, a different window, or a different denominator — not an error in either. The comparison worth making is of definitions, not of figures.

What window should I use?

One long enough to contain several observations of the thing you care about, and the same one next time. Consistency matters more than the exact length.

The other categories

Search engine optimisationThe words that describe how a page is found in a search engine — crawling, indexing, intent, structure, links. This is the ground the AI-search vocabulary is built on, not a separate subject.GEO and AI searchThe vocabulary of being found inside an answer rather than in a list of links: how a system retrieves, what it quotes, and what any of that can be measured with.Trust and E-E-A-TExperience, expertise, authoritativeness and trust — the signals that decide whether a page is treated as a source or as noise, by a search engine and by a system writing an answer.Marketing and growthThe commercial vocabulary around search — what an auction actually reports, who your real competitors are, how attention turns into revenue, and the frameworks used to decide what to write next.

See whether assistants quote you

Reading a definition tells you what the word means. The Vectorscope tells you what your own site looks like to the systems that use it.

Open the Vectorscope