This page provides definitions for key terms used by Evergage.
This article includes the definition of selected terms used throughout the Evergage Platform and within the Evergage Knowledge Base.
A form of statistical hypothesis testing with two variants: test and control. Within Evergage, the term A/B testing is used to more broadly define any campaign that uses a randomized split in visitor traffic to test a hypothesis. With A/B Testing you can split traffic randomly into different groups and show each group different variations of a message. This is done by creating multiple experiences, then assigning each a set percentage of visitor traffic.
In Evergage, accounts are considered to be a company or organization that has individuals who engage with your company. For example, you could have individual users who visit your site that all work for ABC Company. If you have accounts enabled on your dataset, you can see analytics by account, activate (i.e. personalize) based on account, and view account profile data.
These are page loads, clicks, or tracked events that have been assigned a name in Evergage with a process called action mapping. You can use actions to create rules for easier campaign targeting, segment a group, and gather aggregate statistics.
This means using the data you have collected and stored within your CDP to affect a customer experience in some way. This is also known as personalization in the Evergage world because you are using information you know about a person to deliver a personally relevant experience.
This reporting term gives an overview of campaign and goal statistics including trends for impressions, goal completions, and clickthroughs. Within the activity reporting section in Evergage you can analyze information about your recent visitors, see actions taken by your visitors and customers, group users by cohorts, and see other data.
This includes information on pages visited, number and frequency of visits, and page-level of engagement based on mouse movement, scrolling, inactivity and time spent.
ADMC offers options for additional configurability so you can address very specific or complex requirements by creating a query into Evergage to return a specific set of products. ADMC should only be used by experienced developers.
In the recommendations system, this is the reference point an ingredient or exclusion uses as the basis for determining which items to return. For example, if you want to use a Co-Browse ingredient, you need to set the anchor item so Evergage can return items based on what is being viewed on the page, placed in the visitor's cart, last purchased, or purchased by the visitor within a specific number days.
This is data that is unique to a user or account. This data is is collected from contact forms, survey responses on your site or from external databases such as a CRM, email marketing solution, marketing automation tools or a data warehouse.
The period between the visit impression and the goal completion of a campaign. Evergage has four attribution window options available: 30 minutes (default), 1 hour, 24 hours, and 1 week. Since some products may have a longer consideration cycle, you should review how long the timeline to conversion is and use that time frame when reviewing data. Your customer success representative will work with you on selecting the best attribution window for your business model.
There are two attribution types:
Audiences are the specific individuals, groups of individuals (i.e., segments), or accounts (i.e., companies) targeted by or who view personalization campaigns.
The total revenue collected divided by number of orders. This gives the average basket size of the visitors who saw the campaign and made a purchase during the attribution window.Beacon
This report shows information about visits to your site. You can graph out the number of visits based on new versus returning visitors, engagement level, logged in status, and purchase behaviors. You can apply one filter to adjust the results and make comparisons.
You can use boosters to tailor your recommendation recipe to the 1:1 level. When you use a booster, the visitor's affinity score is taken into account and items that match that affinity are boosted in the recommendations query results. In other words, if you boosted the brand affinity in your recipe, visitors who show a preference for a particular brand will be shown items with that brand first.
This message type helps draw attention to a particular area or feature on the page by being on the top, bottom, right or left of the item you wish to draw attention to and having an arrow point to it.
Systems that companies use to engage with customers over the phone (e.g., a 1-800 number). Using an integration, Evergage can pass data including personalized messages and recommendations to these systems to help representatives best serve customers.
Contain experiences and messages you design to personalize the interaction a visitor has with your website or application.
Are created for campaigns only and are usually tied to very a specific set of narrow actions. The purpose for applying campaign-level goals is to avoid exposing visitors to the campaign if they’ve already reached the goal. Once any visitor has reached the goal, they will not see the campaign again. One example of a good campaign goal is moving people through different audience segments. For example, once a visitor watches a getting started video, he would have completed the campaign goal of “Watched Intro Video” and would not see the message again.
Rules set at the campaign-level holistically control how the campaign will display and can be based on things like visitor segments, location, device type, time of day, visitor actions, page, company, and industry.
Campaigns in Evergage can be in one of three states:
This is the collection of your products and content as well as related categories, and tags - such as brand, gender, style, keyword, and author. Evergage dynamically populates your catalog in real-time with product and content data automatically extracted from your site. You can provide this information through a data feed, but gathering your data directly from your site means that your product and content information is always up to date.
Process for integrating your site with your Evergage Catalog. Once your Catalog is mapped and Evergage Recommend is implemented, you can recommend any item from your Catalog to your visitors.
In the most basic sense, a customer data platform (CDP) is a software solution that creates a persistent, unified customer database that can be accessed by other systems. Evergage further defines a CDP to have three levels.
Channels are the different touch points you can use to interact and communicate with visitors and customers. Depending on your Evergage license, you may have access to channels for Web, Email, Mobile, and Third Party (i.e., integrations with other systems).
These happen when a visitor clicks a button or a link being tracked by Evergage. This might be a specific action you want to track, or it might be a button or a link driving visitors to another domain not tracked by Evergage. For example, a login button that takes visitors to your application from your marketing site.
This is the number of clicks that happened in the campaign. Be aware that you will only see clickthrough data if there is something to click in the campaign, like a button or a link.
This is the number of clicks divided by the number of impressions. Keep in mind that if clickthroughs are zero, the clickthrough rate will also be zero.
A business who uses Evergage services.
Probably the most often used Evergage Recommend algorithm. It is a machine-learning algorithm, based on the alternating least squares method of matrix factorization, but is more commonly referred to as a “people like me” algorithm. This ingredient learns about the visitor’s intent, browsing, and purchasing behaviors to determine which affinity categories the visitor falls into, and then clusters the visitor with other visitors who exhibit similar behaviors. The results returned are based on the visitor’s cluster.
What you select in campaign statistics so you can compare campaign experiences against each other or a control. If experience A is the baseline, you could see if the revenue or conversion rate was higher for experience B as compared to experience A.
In Campaign Statistics for an A/B tested campaign, this is the probability that the click-through rate for the updated campaign experience will be greater than that of the original version (control). For example, if confidence is 95%, the chance of making an error—deciding that the updated experience gets a higher click-through rate when it does not—is 5%.
A group of people who qualified to see a campaign but were not shown it. A control group helps you create a benchmark to measure the success or impact of your campaign. Since visitors in this group won’t see the campaign you created, you can compare the results against the group that does see the campaign.
This feature, part of Evergage Decisions, gives you control over which factors will be trained in the machine learning models responsible for serving up recommendations. Once configured, the Contextual Bandit algorithm analyzes these factors for each visitor to determine which promotions to show based on which will generate the most lift.
Data that is passed to Evergage when a visitor lands on your site based on the visitor’s IP address and referring source. This also includes things like geographical data and device type.
This is the number of orders or goal completions collected and then divided by the number of unique visitors who saw the campaign.
This algorithm looks at items or content commonly viewed by previous visitors along with the same item or content the visitor is currently viewing or has placed in his cart. In other words, if a visitor is viewing an article, this ingredient will return articles that other visitors commonly viewed during the same browsing session as the one in which they viewed that particular article.
This algorithm looks at items or content also purchased or downloaded by previous visitors along with the same item the visitor is currently viewing or has placed in his cart during the lookback period. For example, if a visitor placed a jacket in her cart, this ingredient will return items that other visitors have commonly bought with that jacket in the past 30 days.
Using the supplied Apache Zeppelin notebook to combine Python, R, Scala, and SQL code, the Data Science Workbench gives data scientists access to a web-based portal to access, analyze, and build models from the raw data collected and stored by Evergage and other systems on your visitors and customers.
Dates and times are displayed in local time, but the underlying data may be based on UTC. There are several places within Evergage (segments, unified customer profile, campaigns, etc.) where you can adjust the date and time range to reflect the time period you are interested in. Some common times include:
This defines the length of time a product or piece of content should remain marked as new.
Gives a clear picture of each visitor’s true interests and intent based on different types of information including:
The number of times the message was dismissed using the close icon.
The number of dismissals divided by the number of impressions.
A message level rule that limits how many times the message should be displayed - either by a specific interval, all-time, or during the visitor's current visit.
You can use dynamic message content to dynamically render product or visitor variables like location, name, and company. Dynamic message content is controlled on the Promoted Content tab in Message Settings.
Any area on a webpage that can be selected on your site by Evergage — whether that’s an image, a button, a textbox, an Evergage message, or web copy — is considered an element. Evergage automatically places a container around each element on the page, which helps you to identify and select each one. Evergage assigns different colors and patterns to the element containers.
An option you can select in the Path report. This report shows the number of actions recorded for the page you selected in the drop down, the bounce rate for this page, and the path for that page.
This acronym stands for extract, transform, load and is the process used to pull data out of other systems and into Evergage and push data out of Evergage and into other systems.
Includes any interaction a visitor has with your site. An event is everything that a visitor does - a click, a page load, and a message view which is also called an “impression.”
This feature allows you to leverage each individual’s unique intent and preferences to deliver true 1:1 personalized content within your email campaigns regardless of the email marketing or marketing automation system. Personalization is delivered each time an email is opened, this means that messages can account for a customer’s latest actions on your site, offer expiration dates, or even show product inventory levels - all in real-time.
This is a simple and scalable way to provide your website visitors an easy way to enhance online experiences, improve product and content discover, and gain valuable insights into visitor behavior. As visitors explore your website, the products and content assets they view and interact with are automatically added to a “smart bar” dashboard that’s easily accessed during that session or in a future visit. No action is required by visitors – such as selecting a favorite or adding products to a shopping cart – to build their SmartHistory. When visitors access their SmartHistory dashboard, items are sorted by engagement level. This “smart sort” capability considers the active time spent reviewing the item to determine a visitor’s true interest. Visitors can also turn “smart sort” off to view their browsing history chronologically.
The core Evergage application. Gives you access to all parts of Evergage including segments, reporting, configurations, integrations, and campaign management organized by channel. It is best used for managing campaigns or segments, analyzing reports or statistics, adjusting configurations, and making minor edits to existing campaigns.
Combines deep behavioral tracking with machine learning algorithms to suggest individually relevant search terms and results – the instant someone begins typing. Either on it’s own or layered within your existing search technology, It leverages each visitor’s current and past session behavior, implied intent, purchase history, and even offline sales data, to dramatically increase search result relevance, product discovery, and conversion rates. SmartSearch automatically detects out-of-stock and discontinued products and removes them from search results, without needing a integration with your catalog.
Combines behavioral tracking with machine-learning algorithms to present visitors with personalized search results and category views. It can be woven easily into your existing website or mobile app and can also be combined with Evergage SmartSearch without a catalog integration. SmartSort automatically reorders search results and category page views based on each visitor’s unique interests so the most relevant items appear first.
Can be used to change the way visitors navigate your website by displaying personalized navigation to every visitor based on their unique interests and preferences. While SmartNav can be used to personalize a site’s main navigation, it is most often used to promote brands and categories on the homepage. You could also use SmartNav to promote a visitor’s preferred brands or categories or the one he’s most interested in within the main content area of any page on the site.
This feature allows you to use the power of social validation to increase conversions. Use SmartTrends to leverage the “wisdom of the crowd” to influence visitors on your site in real time and build confidence in products viewed by using view and purchase counters. These view and purchase counters can be used to drive both validation and urgency to ultimately influence conversions on your site. Additionally, you can apply badges to product and content images to help draw attention to an item. Badges can be used to promote hot deals, newly added items, closeout products or must-read articles.
Google Chrome browser extension that overlays Evergage editing and campaign creation on your website or application. The Visual Editor makes it easy to “see” your edits since they appear on your site as you create your personalization messages. It is best for creating new campaigns, making web page changes, seeing how your personalization messages will look with your web site, and creating advanced personalization campaigns.
A Web application clients use to create campaigns, view reports and configure Evergage, among other things.
A user of the Evergage Web App. Typically an employee or a contractor of a client.
Allow you to specify what items you do or don’t want to show in your recommendation. For example, you can use Exclusions to tailor your recipe to avoid recommending products that are in the cart, and also use them to include items in the same category as what each visitor is currently viewing.
Data that is provided intentionally by the visitor like survey responses, purchase history, content downloads, or clicks.
You can create different personalization results within the same campaign using Experiences. For example, suppose you would like to create a campaign for first time visitors to your site, but you want to show different message to visitors from Boston and San Francisco because you have events coming up in those two cities. You would create 3 experiences in the same campaign: one for Boston visitors, one for visitors from San Francisco, and one for visitors from everywhere else. Then, you’d use rules to control the visibility of these experiences.
Control the qualification criteria of each experience based on things like location, page, and viewer specifics like segments, visit referrer, and browser.
Allows you to display content from a system outside of Evergage. External message content is controlled on the Promoted Content tab in Message Settings.
Filters are created from segments or global goals. They can be applied to the data you see in campaign statistics and some reports. When you apply a filter, only the people who meet the filter condition will be included in the metrics. For example, you can compare how first time and repeat visitors performed for a campaign, so you can cater experiences to each group. Or, you can see if your campaign is performing well on all devices, or only mobile, or only desktop. In addition to any filters you configure for your dataset, there are a number of pre-existing – or default – filters that can used in reporting and campaign statistics.
This is an extension platform you can use to add custom business functionality to the Evergage platform, including new features, reusable templates and custom reports, and connectors. An Evergage administrator can install Gears from the Gears Marketplace interface and components can be created by developers with Administrator access to Evergage.
These are primary actions you want visitors or users to complete on your website, SaaS app, or mobile app. For example, global goals could include things like: completed a purchase; added to cart; downloaded or viewed content; or completed a form. Global goals don’t need to be restricted to online actions. Since global goals are created from segments, which also consider any offline data you have sent to Evergage, you can also include offline goal metrics. Once a Global goal has been created, you can view how any campaign is performing against that goal. Applying global goals to a campaign gives you a broader view of the effects of a campaign beyond the immediate metrics.
Before you create your campaign, you should consider your business goals and develop a test, or hypothesis, so you can determine how to craft a personalization campaign to support your goals. Your hypothesis should answer the questions:
The aggregation of all of the known information about a customer or an account gathered through on-line and off-line implicit and explicit data, and stored within a Unified Customer Profile in Evergage. Since the same person or account could be identified in a variety of ways by a variety of different systems, Evergage is able to hold multiple identities in the same profile, matching each against a hierarchy of identities, then merges these different profiles together according to customizable prioritization logic.
Information that is tracked by Evergage as a visitor moves around on your site. These are things like pages viewed and time spent on a page.
Data that is passed to Evergage when a visitor lands on your site based on the visitor’s IP address and referring source. These are things like geographical data and device type.
How many times the campaign was seen. This may be greater than the number of unique visitors since they may see a campaign more than once.
Available on the Reports Dashboard, you can see the influence of web recommendations and email campaigns on revenue, orders, visits, AOV, and conversions.
Type of message that appears in the header or footer of the page.
The core algorithms driving a recommendation recipe. You can think of Ingredients as the engine that determines which items are returned for the recipe. You can use multiple ingredients in a recipe, or use alone, to return the most effective recommendations for your visitors.
Way to seamlessly add content or messages within the existing structure of the page.
Refers to kiosks, POS systems, clienteling and mobile apps that help unify an in-store experience with a digital experience. Evergage can pass data including messages, offers, and recommendations to these systems to help personalize a customer's in-store or in-branch experience.
An element type that can be used to easily deploy product or content recommendations to any web page. Rather than designing your own dynamic content and styling the message yourself, you can insert an Item Block element to easily add a preformatted layout to your message. It is a container for content - similar to how a picture frame is a container for a photograph or piece of art.
Provides an overview of content views and product purchases for top performing products, blogs, or articles, as well as product brands, styles, and categories.
Item Templates are configured in the Catalog tab using simple HTML. Item Templates work in tandem with Item Blocks. The Item Template defines how a product or piece of content, and its related details, should be displayed. You customize your Item Templates to match your existing brand guidelines. For ease of use, you can create more than one template. Item templates are the content displayed inside an item block - similar to how a photograph or piece of art is displayed inside a picture frame.
A snippet of code provided to you by Evergage that you add to your site. The beacon places a first party cookie on each visitor’s browser and sends data for that visitor back to Evergage. The data sent back includes pages visited, links clicked, time on the site, number of visits, geolocation, referral source, as well as other custom data you want collected.
The number of days you want the recipe to look back at the visitor’s history. When deciding how far you’d like to look back, think about what will support your product and content catalog and business goals.
The content the visitor will see if he qualifies to see the campaign, the experience, and the message itself. A message is contained within an experience and you can have more than one in an experience. For example, you could create a different message for visitors depending on whether they are using a mobile device or a desktop computer. You would use rules to control who sees which message.
Control message visibility based on viewer device type, visit frequency, and user behaviors like time on the page and page scroll behavior by applying rules at the message level of a campaign.
There are four types of messages you can create in Evergage.
You can configure the minimum amount of time a visitor must spend viewing an item to be considered interested in it. The default is 30 seconds.
This algorithm uses the published date captured in the Catalog. Use this algorithm to prioritize the most recently published items or content on your site.
You can add a rule to define what must occur in order for the message to be displayed. For example, you could configure the message to be delayed on the page and have it display only after the visitor has spent a specific amount of time on the page, or only display the message after the visitor has scrolled down a specific percentage of the page, or only display if the visitor has been inactive on the page for a specific amount of time, or if the visitor is showing signs of bouncing away from the page.
The number of orders placed by visitors who saw a campaign experience during the attribution window. These are not only product orders, but are considered by Evergage to be a goal completion on your site - product purchases, content download registrations, form submissions, newsletter signups, and other goal completions.
If you’ve configured outliers in Catalog Setup, you can exclude outliers in Campaign Statistics by switching Outlier Filtering to ON. For example, suppose a visitor placed an exceptionally large order - which was far above the expected deviance from the average order value - this order could skew the statistics for your campaign and you may want to exclude orders outside of the general range you’ve configured. Outlier filtering is retroactive.
These occur when a user visits a particular page on your site or in your application.
Details which pages on your site people enter on arrival and where they go next so you can better target your communications and content strategy. For example, if people are abandoning a demo request form, and you can see that they are going to any one of four other pages, messaging them about the demo on those four pages will help increase demo sign ups. Also, if no one enters your site on the homepage, posting a “Getting Started” video on that page might not be the best approach.
A report that enables marketers to gain insight into the the flow of visitors as they navigate through your site. This report also includes bounce rate information.
A type of activation, the premise of personalization is understanding your visitor’s mindset and intent so you can create the best experience for each visitor as soon as he lands on your site, or opens your email.
Type of message, which, as the name implies, pops-up on a visitor’s screen.
Actions occur in less than 20 milliseconds in Evergage. This means that as your site or app visitors or email recipients make different choices, or as their behavior changes, Evergage responds instantly and seamlessly to deliver personalized experiences to each one.
All of the visitors who have been to your site since the beacon was activated, updated in real-time. A green dot indicates a visitor who is using your site right now.
Recommendation strategy in Evergage. You build recipes by adding together one or more ingredients - or the core algorithms driving the recipe - as well as exclusions, and boosters. You can create a variety of combinations to serve up the right content or products based on an individual visitor’s behavior and affinities on your site.
Algorithmic configurations designed to present products and content to individual visitors based on their unique preferences.
These types of campaigns display information to people based on their unique preferences driven by a set of algorithms. Typically, these campaigns are used to deliver product and content recommendations.
Details how people arrived on your site so you can customize your messaging for each visitor. You can quickly see key statistics about visitors referred from top sources including referral percentage, number of visits, and revenue. Hover over the graph to see date, time, and revenue detail. Switch to the list view to see information like bounce rate, average visit duration, average number of pages viewed, as well as average and total revenue.
This allows you to configure how long a product or piece of content will be included in recommendations. This factors in the amount of time set for the Decay Age. For example, if you enter the decay age as 48 hours, and you set the Relevancy Retained to 50%, that means that after 48 hours have passed, the recommendation score Evergage uses to determine which items to return will be reduced to half of what it was at the moment the item was published. In other words, if the original score was 1, after 48 hours, the score will be .5, and it will remain .5 from that point forward.
You can apply a results filter when looking at Campaign Statistics. There are several default filters listed in the drop down in addition to any global goals you’ve set. Once you have selected the filter, the campaign statistics will recalculate to show the performance of the campaign only for the visitors from the selected filter.
Amount of revenue generated from the people who saw the campaign and then made a purchase or a download during the selected attribution window. In reports, the Revenue statistic shows daily revenue totals and a summary of the revenue generating activities happening on your site.
Total revenue generated divided by the number of visitors who saw the campaign experience and then made purchases or downloads during the selected attribution window.
Control how a campaign, experience or message shows based on certain criteria.
Allows you to deliver specific experiences to specific groups or segments of people based on who they are or where they are in your site or app. The simplest way to think about rule-based testing is with IF/THEN statements. IF a person falls into segment A, THEN show him experience A. IF a person falls into segment B, THEN show him experience B, and so on.
Real-time grouping of accounts or individuals based on a set of criteria you define. You can use segments for many things including analytics, campaign goals, and even to set parameters for rule-based campaigns. Segment updates happen in real-time, so any membership changes occur immediately, even during the same visit. Segments can be user or account based.
User segments - Segments of individuals who visit your site or log into your application
Account segments - Are available when you have B2B detect configured for your site. They are segments of companies and are typically used to target visitors coming from a particular organization
You can use segments declared as filters can be utilized to filter important reports. Segments declared as goals can be applied to campaigns for measuring lift. For the sake of reporting, goal segments will also be considered filter segments.
This algorithm will return recommendation results for products or content that are similar in nature to the item the visitor is currently viewing or has placed in the cart. The Similar Items ingredient is commonly used on a product display page to help visitors discover more products. Often, you'll see it at the bottom of the page with a label like "Other Similar Products".
Includes data such as a visitor's referring site, email or ad campaign source, IP address, geo-location, device type and operating system.
This is a merchandising algorithm. It looks at products which are bought together in the same cart and returns recommendations based on explicitly defined categories. With this ingredient, Evergage will recommend paired items that are commonly bought together. For example, if the visitor is viewing men’s dress shirts, show items in the “tie” category.
This algorithm uses the expiration date captured in the Catalog. Use this ingredient to prioritize items that are about to expire.
Measure of how certain you can be with the reported lift. A metric is determined to be statistically significant based on a number of factors built into Evergage’s algorithms.
You can add a to rule to define where the campaign or message should be displayed. For example, you could set a rule so the campaign only appears on the desired pages and does not appear on any pages where it should not appear.
Provides insight into the types of browsers and devices your visitors are using when they view your site. It is important to understand this information so you are aware of how your visitors are interacting with your site so you can set-up your campaigns accordingly. For example, knowing whether visitors are accessing your site from a mobile device will help you determine if you should be developing campaigns for mobile browsers.
Determines the point at which the visitor is considered affine for the booster. The slider has five positions with the first position being the lowest weighting and fifth, the highest. In basic terms, you move the slider to specify how much a visitor needs to interact with the selected affinity to be considered affine and boost products from that Affinity. Once the visitor meets that threshold, Evergage will boost products within that Affinity.
During training, the recipe reviews all of the implicit and explicit data that Evergage is continuously collecting about each visitor, learns about the different data points, and compiles the ingredients, or algorithms, that will power the recommendation.
Informs you where your visitors originated from before they came to your site. This includes other sites like social media, search based on search terms entered, and other types of web sites.
Use this algorithm to recommend products or content based on what is trending on your site right now. You can define this ingredient to display items or content based on the highest number of purchases or downloads, the most revenue generated, the most views, or the most amount of time spent viewing.
Triggers are specific events like visitor actions, cumulative behavior, or environmental changes that Evergage can take action on to send an email or perform some other activation (like personalization).
Visual representation of all the data about a single visitor including preferences and affinities. You can roll this up to the account level so you can see the relationship between visitor behaviors at the same account.
Number of tracked individuals who have seen the campaign.
A visitor to a Client website.
Counts all visitors’ sessions on your site, regardless of how many times the same visitor may have been to your site.
Represents the individual user or visitor to your site. In reports, the visitors statistic shows you the number of unique daily visitors and the number of visits during the set timeframe. You’ll also see an indicator showing whether traffic is up or down, where visitors are coming from, and whether they are known or anonymous users
One of the key things you can configure for every ingredient is its weight. Every item returned by an ingredient has a weight, and items that have the highest weight will be displayed first by the recipe. You control how the ingredient will be weighted against the other ingredients in the same recipe by increasing or decreasing the weight and the importance of items returned by an ingredient. Evergage then multiplies the scores for the items returned for the ingredient by the assigned weight, and displays the top weighted items. Since the recipe will attempt to return items for the ingredient with the highest weight first, give your highest priority ingredient the highest weight.