0. Glossary - MarSci Study Guide (Facebook)

페이스북에서 정의한 주요 마케팅 과학 용어 일람

Acronym	Term	Definition
	A/B Test	A test that allows you to conduct experiments that compare multiple targeting groups or campaigns by splitting audiences into randomized and mutually exclusive groups.
	Advanced matching	A Facebook pixel feature that allows websites to pass additional site-visitor information (such as email addresses or phone numbers).
	Alternative hypothesis	A hypothesis contrary to the null hypothesis, that there's a relationship between two measured phenomena or an association among groups
API	Application programming interface	Defines how software components communicate.
	Attribution	The practice of assigning value to various marketing efforts.
	Attribution model	Determines how credit is given to touchpoints for a conversion. The attribution model logic can be based on a rule or set of rules, or a statistical model.
	Attribution window	The finite period of time during which conversions can be credited to a particular ad.
	Auction	An ad buying methodology in which the ads run based on the maximum bid and performance. Ads compete against each other in the process and the system determines the ones most likely to be successful to be displayed.
	Bid	The amount an advertiser is willing to pay to achieve their desired outcome.
	Bid cap	A bid strategy that enables advertisers to set a maximum bid Facebook can use in each auction
	Bid pacing	Adjusts a bid or which auctions to enter based on how much budget and time are left for an ad set.
	Bid strategy	A setting in Ads Manager and Reporting that helps Facebook determine how to spend a budget to compete in a Facebook auction.
	Brand lift	Measurement product that uses experimental design (randomized control trials) to detect brand impact that might be caused by ads run on Facebook.
	Budget allocation	The amount of marketing expenditure allotted for each marketing activity.
	Buying type	Options in Ads Manager that determine the method by which you purchase ads on Facebook, either through auction or reach and frequency.
CTR	Click-through rate	The number of times an ad or a link to a web page is clicked, compared with the number of times it's displayed.
	Control group	The group in an experiment or test for which none of the factors in the test are variable. It's used as a benchmark to measure the effect of the test.
	Conversion efficiency	How effectively your ads drove the interactions you're measuring.
	Conversion lift	A Facebook product that uses randomized control trials (RCTs) to measure the number of incremental conversions that result from Facebook ads. A marketing metric that quantifies the number of additional conversions that result from Facebook ads.
	Conversion lift percent	The percent difference in conversions between the people who did and didn't see your ads during a test.
	Conversion rate	The (estimated) number of times a link on a web page is clicked, compared with the number of times it's displayed.
	Conversion window	Period of time considered between seeing and acting upon it toward the ad's main goal, be it generating a lead or making a purchase.
	Cost cap	A bid strategy that enables advertisers to set a maximum average cost per optimization event for an ad set. Facebook will keep the average cost as far below that amount as possible, but keep showing ads until it's reached.
CPA	Cost per action	The cost to the advertiser each time an ad prompts an action.
CPC	Cost per click	The cost to the advertiser each time an ad is clicked.
CPM	Cost per thousand(Mille) impressions	Also known as "cost per mille", it's the average cost for 1,000 impressions of an ad, or the average revenue received for 1,000 impressions of an ad on apps or websites.
	Cross-channel measurement	Indicates outcomes related to advertising across numerous online or offline channels, such as Facebook and television.
	Custome event	A logged action based on a specific action you want audiences to take on your website, app or offline.
	Customer relationship management	A tool the enables businesses to manage customers' contact information and interactions through the customer life cycle.
	Daily budget	The average amount you're willing to spend on an ad set or campaign every day.
	Data analysis	A process of inspecting, cleansing, transforming, and modeling data with the goal of discovering useful information, informing conclusions, and supporting decision-making. Data analysis has multiple facets and approaches, encompassing diverse techniques under a variety of names, and is used in different business, science, and social science domains. In today's business world, data analysis helps businesses to operate more effectively and make decisions more scientific.
	Data source	A tool, connection, piece of code or another object that collects information, such as Facebook pixel, Facebook SDK, or offline conversions. The information can be used for measurement and analysis.
	Dependent variable	A variable whose value depends on that of another.
	Event	A logged action that people take on your website, app, or offline, usually used for capturing and measuring ad performance.
	Exact matching	In marketing research, an observational method that examines a group of people exposed to an ad campaign and tries to find a match for each person in the non-exposed group.
	Experimental method	In marketing research, a measurement method that shows different types of ads to separate groups of people in a controlled manner.
	Facebook attribution	A Facebook measurement product that provides advertisers with cross-platform, cross-channel, cross-device, multi-touch attribution performance reports for ad campaigns.
	Facebook's data-driven attribution	The process of determining credit for touchpoints on a conversion path based on their estimated incremental impact and statistical modeling.
	Facebook pixel	A piece of code installed on a website that captures website events.
	First-party data	Data a brand collects directly fro its customers (for example, website activity data, sales data, and so on).
	First-touch attribution	A rule-based attribution model that gives 100% of conversion credit to the first click or visit in a conversion path. If there is no click or visit, then it will credit the conversion to the first impression.
GRP	Gross rating point	A unit of measurement of audience size for TV advertisements. GRP is used to measure exposure to one or more programs or commercials, without regard to multiple exposures of the same advertising to individuals.
	Holdout test	Measures the total conversions caused by your Facebook ads.
	Hypothesis	A supposition or proposed explanation made on the basis of limited evidence. used as a starting point for further investigation.
	Impression	A single instance of an ad or piece of content (such as a post) appearing on screen.
	Independent variable	A variable whose variation does not depend on that of another. It is the factor that is purposely changed or controlled in order to see what effect it has.
ITT	Intent to treat	A randomization method that includes all randomized participants in statistical analysis and analyzes according to the group they were originally assigned, regardless of what treatment (if any) they received. cf. TOT(Treatment of the treated) https://blog.naver.com/snower2357/222191143358
	Inter-channel budget	The allocation of budgets across different channels.
KPI	Key performance indicator	A metric selected to evaluate the success of a campaign or ad.
	Last-touch attribution	A rule-based attribution model that gives 100% of credit to the last ad a person interacted with before a conversion, whether it's an impression or a click.
	Lifetime budget	The amount you're willing to spend over the entire runtime of your ad set of campaign.
	Lift solutions	A platform for running experiments on Facebook ads and measuring the incremental impact of ads on business outcomes.
	Liquidity	In the context of machine learning, the concept of allowing Facebook to determine the best placement, ad set budget, and bids for a campaign.
	Lowest-cost bid strategy	The bid strategy that provides the highest volume of conversions for the available budget
	Machine learning	A discipline that uses science, information, and computer code to automatically predict certain outcomes based on discovered patterns that are not explicitly programmed.
MMM	Marketing mix modeling	A regression-based analysis that quantifies how much sales and thus return on ad spend can be correlated to each media channel in the mix, for both offline and online channels.
	Multi-cell test	A test that runs multiple experimental designs at the same time. For example, in a two-cell multi-cell test, cell A has two groups (test and control) and cell B has two groups (test and control).
	Multi-touch attribution	The process of considering credit for each touchpoint on a conversion path.
	Natural experiment	An observational study in which subjects fall into either the exposed or non-exposed groups based on a naturally occurring event.
	Nested test	An experimental method in which the treatment group for one test is subdivided into a control group and a secondary testing group.
	Null hypothesis	A general statement or default position in inferential statistics that there in no relationship between two measured phenomena or no association among groups.
	Observational method	A way of collecting data through observing.
	Offline conversions	Allow you to measure how much Facebook ads lead to offline outcomes, such as in-store purchases, phone orders, bookings, and more.
	Opportunity logging	In an ad's test period, for the target audience that has an opportunity to see the ad at some specific stage along the delivery funnel (meaning, for a user's request, that the user has an ad passed at some stage), Facebook logs the opportunities (uniquely identified by tags like, userId, studyId, dynamicHoldoutId, and isContro) to scribe tables.
	Pacing system	A component of ad delivery whereby Facebook measures and projects a campaign's end date to ensure a budget is spent as evenly as possible over the lifetime of an ad set.
	P-value	The probability of obtaining test results at least as extreme as the results actually observed during the test, assuming that the alternative hypothesis is incorrect.
	Parameter	Included alongside events to measure additional information about the product or service type people engage with on an app or website, or offline.
	Propensity score matching	In marketing research, an observational method that assigns everyone in the exposed and non-exposed groups a probability of being exposed. Then people from each group are matched based on their similar probabilities.
	R²	Also known as the coefficient of determination, R² is the proportion of the variance in the dependent variable that is predictable from the independent variable(s).
	Randomized control trial	RCTs test a hypothesis by introducing a treatment, studying the effects, and determining the impact of your ad. Ultimately, an RCT can help determine how much to spend on each marketing channel in order to maximize results.
	Reach	The number of people exposed to a medium at least once during a given period.
	Reach and frequency	A method of buying ads on Facebook that allows advertisers to predict the audience size reached and control the average frequency that the audience is exposed to their ads.
	Regression adjustment	In marketing research, an observational method whereby people build a model based on the observed data and use the model to predict the outcome of the campaign based on a set of observed variables, including whether or not people were exposed to the advertising campaign.
ROAS	Return on ad spend	An economic indicator used to evaluate the effectiveness of ad spend. It is calculated as the ratio of the amount gained or lost relative to the amount invested.
ROI	Return on investment	An economic indicator used to evaluate the efficacy of total media investment, including trade. It is calculated as the ratio of the amount gained or lost relative to the amount invested.
	Rule-based attribution model	Different attribution models distribute different amounts of credit for conversions across ads. They can be based on a rule or a set of rules, or a statistical model. Rule-based models let you select the rule that will determine how conversions should be attributed to different touchpoints. Single-touch models give credit to only one touchpoint, while multi-touch models give credit to multiple touchpoints in the consumer's conversion path.
	Sales lift	The additional revenue caused by ads in a test, when compared to the test and control groups.
SEM	Search engine marketing	The practice of marketing a business using paid advertisements that appear on search engine results pages.
	Single-cell test	A test with one test group, where group A is shown the ad and group B is not. cf. Multi-cell test.
SDK	Software development kit	A piece of code installed on an app that captures app events.
	Statistical attribution model	An attribution model that uses algorithms to determine credit for each touchpoint.
TRP	Targeting rate point	The number of impressions bought per 100 people in an ad's target audience.
	Test duration	How long a test runs for, from beginning to end.
	Third-party data	Data collected by a party that does not have a direct relationship with the user (such as DMP-Data Management Platform).
	Total value calculation	(Advertiser Bid x Estimated Action Rates) + User value
	User value	A prediction of how engaging or relevant a user may find an ad based on available signals.
	Variable	A quantity or other piece of data that can assume different values.
	Visit	In Facebook Attribution, the number of times a person loaded your website from any channel (including paid, direct, organic, and untracked channels), as recorded by the Facebook pixel of SDK.

 

Marketing Science Study Guide