​
Overview

// Image Innkeepr generates incrementality forecasts based on recent user engagement data. We track interactions across all digital properties—including websites, mobile apps, and backend systems—and convert them into audience and conversion signals. This provides marketing and growth teams with an analytics infrastructure for optimizing user interactions based on incremental impact. To activate Innkeepr, two components must be configured:
  1. Event collection: Send user interaction data via Innkeepr’s client-side tracking libraries or server-side APIs. This includes events such as checkout_completed, signup, or product_viewed, along with relevant traits (e.g., order_value, is_new_customer, source).
  2. Objective definition: Specify a measurable target variable that Innkeepr should optimize for—such as customer lifetime value (CLV), purchase probability, average order value (AOV), or retention likelihood.
Once both components are in place, Innkeepr begins analyzing daily traffic to identify cause-and-effect relationships between marketing touchpoints and defined objectives. These relationships form the basis for forecasting incremental impact on an individual user level. The resulting predictions are transformed into activation-ready signals that can be synced directly into connected media platforms.

​
Common use cases

Growth and marketing teams use Innkeepr’s signals to:
  • Build lookalike audiences focused on new-customer acquisition
  • Optimize conversion values based on incremental impact
  • Target high-CLV customer segments
  • Exclude low-AOV purchasers from retargeting
  • Suppress spend on low-impact or non-incremental traffic

​
How you can collect data

You can collect data by implementing Innkeepr’s tracking libraries as your Sources:
  • Innkeepr.js – Our JavaScript source is the most powerful way to collect customer data from your website. Innkeepr recommends it as the default installation for web tracking.
  • Server-side sources – Ideal for sending analytics data directly from your backend when client-side tracking isn’t feasible, or when sending mission-critical data like revenue or transaction events.

​
The Innkeepr methods

Innkeepr’s tracking libraries generate structured messages about user activity across your digital properties. These messages are standardized and fed into our causal analytics engine to generate activation-ready signals. The core tracking API methods are:
  • Identify – Who is the user?
  • Page – What page or screen are they viewing?
  • Track – What action are they taking?
All tracking methods share a common structure and fields. Using these methods consistently ensures that Innkeepr correctly detects the context and transforms the data into audience and conversion signals for downstream activation.

​
How Innkeepr builds signals of incrementality

Signals are the mechanism through which Innkeepr communicates incremental impact to connected media platforms. The system is built around two core concepts: objectives, which define the business outcomes you are optimizing for, and treatments, which represent the marketing touchpoints users were exposed to. Together, these inputs power Innkeepr’s causal analytics engine and enable the generation of accurate, activation-ready signals.

​
Objectives

Objectives represent measurable business outcomes that serve as guiding metrics for signal generation. They typically include metrics such as customer lifetime value, purchase probability, order value, or retention likelihood. An objective can be derived either directly from tracked events (e.g., a checkout_completed event with a revenue value) or from calculated traits associated with a user profile (e.g., is_new_customer, revenue_30d). Multiple objectives can be defined and versioned independently, allowing teams to run parallel analyses or optimize different growth surfaces. Objectives can also be filtered or segmented by additional traits—such as “CLV for new customers in the EU market” or “7-day retention for users from paid social”.

​
Signals

​
Audience signals

Audience signals are custom audience segments built on predicted incremental value. Instead of targeting users solely based on historical engagement or demographic traits, Innkeepr generates audiences from segments that are forecasted to deliver the highest lift against a defined objective. These segments can be directly synced into connected platforms (e.g., Meta Custom Audiences) to focus ad delivery toward users with the greatest expected incremental impact. Audience signals enable growth teams to expand into new, high-value segments while reducing spend on non-incremental traffic.

​
Conversion signals

Conversion signals are adjusted conversion values that reflect the incremental effect of previous marketing treatments. Rather than passing raw transaction values or unweighted conversions, Innkeepr assigns a value multiplier to each conversion based on its predicted lift over the control scenario. This means platform bidding algorithms receive optimization signals that are aligned with actual causal impact, ensuring budget is directed toward actions and touchpoints that contribute the most to long-term business outcomes.

​
How signal quality improves over time

Signals in Innkeepr are powered by the continuous analysis of marketing treatments and their measured impact. Treatments represent entities of engagement. In Innkeepr, treatments correspond to ad sets, ad groups, or asset groups across platforms like Google, Meta, and other media accounts. Each incoming user session is matched to the treatment that first drove the visit using our internal attribution system, which applies a 7-day first-touch logic. This correlation allows us to determine which treatments caused each session and to track how those sessions later behaved across your website or app. At the core of Innkeepr’s learning process is the estimation of treatment effects. A treatment effect quantifies the incremental impact of a specific treatment - such as a Meta ad set - on a defined objective, for example customer lifetime value (CLV) or purchase probability. Innkeepr learns to predict treatment effects from historical engagement and conversion data. These predictions are continuously refined as new data flows in, ensuring that the underlying models stay aligned with current user behavior and market dynamics. Over time, this feedback loop increases the accuracy of audience and conversion signals, so connected media platforms can optimize bidding, targeting, and spend allocation based on the most up-to-date causal insights available.