Glossary

Term	Definition
Ad Stock	The prolonged impact of advertising on customer decisions, extending beyond the immediate spend and encompassing both short- and long-term effects.
Admin	The highest level member of an organization. Admins have full control over account settings and can create new connectors, add new members, and view all tabs.
Attribution Lag	The delay between ad exposure and conversion, affecting revenue attribution.
Backtesting	A statistical validation process where Prescient’s models are tested against historical data to ensure accuracy and improve forecasting reliability.
Base Revenue	When it comes to MMM Revenue, this is also known as first-order revenue, which is most analogous to last click attribution, where the customer purchased as a direct result of seeing or interacting with this campaign.
Base ROAS	This is the ROAS that you would see as a direct result of the campaign. Simply put, it’s Base Revenue / Spend.
Channel Interactions	The relationship between marketing channels, such as how Facebook prospecting increases branded search conversions on Google.
Channel Revenue	Also called “Reported Revenue”, this refers to the revenue attribution model used by the associated platform to report revenue for a given campaign or channel, such as Facebook Ads, Amazon Ads, or Pinterest Ads.
Channel ROAS	Also called “Reported ROAS”, this is Channel Revenue / Spend for a given campaign or channel, such as Facebook Ads, Amazon Ads, or Pinterest Ads.
Clicks	Platform reported ad clicks.
Confidence	Confidence in Prescient is a 0–100 measure of how reliably our forecasts reflect real outcomes, calculated by comparing back-tested model predictions against actual results. Confidence bands outline the possible range (upper and lower bounds) for each forecast, while confidence scoring (in Spend Optimization) indicates how reliably budget recommendations can be made, based on the volume, variance, and consistency of historical data.
Connector	A data source connection that transfers data from the data provider to Prescient. Built using Fivetran - a tool for replicating data from one place to another - or internally by our engineers.
Counterfactual	A “what if” scenario imagines alternative marketing actions—such as reducing ad spend on a particular channel to zero—without altering live campaigns. By modeling these hypothetical changes, teams can gauge potential outcomes and refine their strategies with minimal risk.
CPM	Cost per 1 thousand impressions.
Customer Acquisition Cost (CAC)	The cost to acquire a new customer, calculated as Spend / New Customers. Prescient tracks modeled CAC to help manage acquisition costs efficiently.
Customer Success Manager (CSM)	A dedicated point of contact who assists with onboarding, troubleshooting, and optimizing Prescient’s platform for your organization.
Decay	The rate at which marketing impact fades over time, used in Prescient’s models to predict long-term campaign effects.
Diminishing Returns	The decreasing marginal revenue generated from additional ad spend.
Ecommerce Revenue	The total amount of money brought in by your ecommerce stores, such as Shopify and Amazon. We calculate this by looking at the summed subtotal of orders, meaning this number is equivalent to the total amount of revenue a business generates from sales after accounting for discounts, customer returns, and other deductions.
Ensemble Modeling	A modeling technique unique to Prescient that layers multiple additive models together to improve prediction accuracy by capturing different aspects of marketing performance.
Flat Spend Effect	A scenario where consistent daily spend prevents the model from identifying incremental revenue impact. Some variability in spend improves model accuracy.
Frequentist Statistical Methods	An alternative to Bayesian modeling that relies purely on historical data without incorporating prior probabilities.
Halo Effects	Revenue influenced indirectly by a marketing campaign, such as increases in organic search, paid search, or direct traffic conversions.
Halo Revenue	As part of MMM Revenue, also known as halo effects or second-order revenue, this tells us how a campaign helped to drive across paid search, organic search, and direct traffic.
Impressions	Non-unique ad views.
Incrementality	The additional revenue driven by an ad campaign that would not have occurred otherwise.
Interaction Effects	The impact of one marketing channel on another, such as how prospecting spend influences retargeting performance.
KPI Report	A visualization on the Performance Page that allows users to analyze key performance indicators (KPIs) by channel.
Last-Click Attribution	A traditional attribution method that assigns 100% of the credit for a conversion to the last marketing touchpoint before purchase. Prescient’s MMM provides a more comprehensive view of marketing influence.
Marketing Mix Modeling (MMM)	A statistical method used to measure the impact of marketing spend on revenue by analyzing variations across channels and campaigns. Unlike Multi-Touch Attribution (MTA), MMM accounts for factors like seasonality, delayed effects, and market trends.
Member	The lowest level member of an organization. Members are read only accounts that cannot edit organizations in any way.
MMM CAC	Algorithmically calculated cost of new customer acquistion attributed to your ecommerce Shopify store. Put simply, this is New Customers / Spend.
MMM New Customers	Algorithmically calculated New Customers attributed to your ecommerce Shopify store.
MMM Revenue	Algorithmically calculated revenue attribution. Millions of simulations are run to determine how each campaign impacts total revenue. The results of those simulations are then used to determine the actual amount of money generated by your campaigns.
MMM ROAS	Algorithmically calculated ROAS. This metric is calculated by dividing True Revenue by ad spend for a channel or campaign.
Modeled Attribution	A methodology used by Prescient to allocate revenue impact based on MMM rather than last-touch attribution. It accounts for delayed conversions, channel interactions, base vs. halo effects, saturation curves, and ad stock effects.
Multi-Touch Attribution (MTA)	An attribution method that assigns credit to multiple touchpoints in the customer journey, typically focusing on short-term impact rather than long-term trends.
No Change Spend	Sometimes referred to as Current Spend, this is forecasted amount of money we expect you to spend over the next n days within the prediction. Note that this is different from 'Spend', which is what was actually spent in a historic timeframe.
Optimizer	Prescient’s algorithmic engine that uses our attribution model to generate AI-driven budget recommendations—at the channel, tactic, and campaign levels—and simulate different spend scenarios. By forecasting the impact on revenue and ROAS, it guides marketers in making data-driven decisions to maximize ad spend efficiency.
Platform Bias	The tendency of advertising platforms (e.g., Google Ads, Meta) to over-attribute revenue to their own channels due to self-reporting. Prescient’s MMM helps brands counteract this bias with independent, modeled insights.
Predicted Revenue	Forecasted revenue for over the next N days configured in the prediction.
Profit	The amount of money spent on ads subtracted from the amount of money brought in by your ecommerce sales.
Recommended Spend	An algorithmically identified spend range that will provide optimal profit for a given campaign.
Saturation	The point where additional ad spend stops delivering proportional returns, reflecting diminishing gains as a campaign nears its maximum effectiveness.
Seasonality	Predictable fluctuations in marketing performance that occur at specific times of the year due to external factors such as holidays, sales events, or industry trends. Prescient's models account for seasonality by isolating its impact from other marketing variables, ensuring that optimizations focus on sustained performance rather than temporary spikes.
Spend	Actual spend on a campaign or channel over a given period of time.
Spillover effects	Spillover effects (or “halo effects”) are the indirect impacts of a marketing campaign beyond its direct, last-click conversions, such as one ad’s influence on organic search traffic or Shopify sales even if there’s no immediate click.
Tactic Labeling	The process of categorizing campaigns into mutually exclusive Tactics for streamlined reporting and optimization.
Trends	Long-term patterns in marketing performance that indicate business growth or shifts in effectiveness. Prescient's models isolate these trends from short-term fluctuations, providing insights into trajectory and performance over time.