Predictive Audiences provides marketers with highly valuable audience groups based on consumers’ predicted actions, preferences, and lifetime value. With data integrity and acquisition at our core, Bluecore predictions use terabytes of unique data, making them better, faster, and more accurate. Predictive Audiences group customers together based on their various scores for each prediction. Use Predictive Audiences to find next-best customers to grow value, retain and win back at-risk customers, and optimize revenue based on discount analysis.

The predictions we are initially launching with include:

Though calculated in different ways with different logic, each predictive filter utilizes predictive analytics and data science algorithms to measure each score per customer. These scores are refreshed and recalculated on a daily basis as new data is made available.

PREDICTED CUSTOMER LIFETIME VALUE 

Predicted Customer Lifetime Value predicts the amount a customer will spend over the course of their relationship with your brand. Customers are filtered into percentiles according to their predicted lifetime value.

Many other Customer Lifetime Value (CLV) calculations only take a customer’s past behavior into account; saying somebody has a high CLV only means that they have purchased a lot in the past, leaving non-buyers completely out of the equation until they make a purchase. However, Bluecore’s calculation is Predicted CLV, or what we expect a customer to purchase in the future.

Our calculations are forward looking rather than based purely on historical data, meaning that we can identify brand new customers that have never bought from you as having a high predicted lifetime value. Those customers with little or no purchase data that resemble your highest spending customers can be identified as such before they ever spend with you, thereby being included in your high predicted CLV audience. Because we take other online signals into account (e.g., frequency of visits, when the customer signed up, how many products were browsed each visit), we do not miss potentially high value customers. We can also remove people that have spent a lot in the past and are no longer active or don’t look like they’ll spend much in the future.

Predicted CLV is an easy way to boost ROI on CRM-targeted campaigns in paid media channels (FB, Google Adwords, etc.) by cutting out low value customers.

USE PREDICTED CLV IN BLUECORE

To filter customers based on Predicted Customer Lifetime Value, select Top, Between, or Bottom and enter the percentile or range you want to target.

When you enter a percentile or range, the graph will show the average predicted CLV and the approximate number of customers available for the filter.

Select the checkbox to include customers with an unknown Customer Lifetime Value. A customer has an unknown Customer Lifetime Value if Bluecore does not have enough information about the customer.

EXAMPLE

We want to filter for the top 5% of customers with the highest predicted lifetime value so we can send them a loyalty discount. We would select Top from the dropdown and enter 5 in the % of customers field.

LIFECYCLE STAGE

Lifecycle Stage predictions find customers based on which of four buying cycle stages they fall into:

  • Non-Buyers

  • Active Buyers

  • At-Risk Buyers

  • Lost Buyers

Many other lifecycle calculations use a static definition of time to define each step of the lifecycle. For example, somebody is at-risk if they haven’t purchased in 90 days or lost if they haven’t purchased in 120 days. Targeting these customers based on a static amount of time will force you to target your most valuable customers too late and your more infrequent and lower value customers too early.

Bluecore is able to predict which customers are at-risk on a individual basis by looking at each customer’s purchase cadence, which is based on their past purchases, as well as the past purchase cadence of similar customers. Once a customer begins to fall out of their personal purchase cadence, they become at-risk. For example, if you have one customer that buys every month and another that buys every three months, after 60 days we would notice that the monthly buyer is At-Risk while the quarterly buyer is still Active and within their usual purchase cadence.

Use Lifecycle Stages to easily automate lifecycle based campaigns:

  • Turn non-buyers into buyers

  • Turn one-time buyers into repeat buyers

  • Target customers as they become at-risk

NON-BUYERS

Non-Buyers are customers who have not made a purchase on your site yet. You can choose to filter for all Non-Buyers by selecting the main Non-Buyers checkbox. To filter for Non-Buyers who joined the audience segment within a specific timeframe, select the Became a Non-Buyer in the last checkbox and enter a number for the number of days for your timeframe.

EXAMPLE

We want to filter for customers who have signed up on our site in the past 7 days, haven’t purchased yet, and have a high Predicted Customer Lifetime Value so we can send them a special discount through Facebook. After selecting our Predicted Customer Lifetime Value filter, we would select the Non-Buyers and Became a Non-Buyer in the last checkboxes and enter 7 in the text field.

ACTIVE BUYERS

Active Buyers are customers who have made a purchase within their typical buying cycle. Each customer’s typical buying cycle is unique – one customer’s buying cycle can be one purchase every two weeks, while another customer’s buying cycle is one purchase every six months.

You can choose to filter for all Active Buyers by selecting the main Active Buyers checkbox. To further filter Active Buyers, select a type of Buying Behavior: All buyers, One-time buyers, or Repeat buyers. To filter for Active Buyers who joined the audience segment within a specific timeframe, select the Became an Active Buyer in the last checkbox and enter a number for the number of days for your timeframe.

EXAMPLE

We want to filter for customers who became Active Buyers within the past month, but only for those who are One-time buyers so we can utilize display advertising to surface a post-purchase campaign that encourages them to buy again. After selecting the Active Buyers filter, we would select the One-time buyers and Became an Active Buyer in the last checkboxes and enter 30 in the text field.

AT-RISK BUYERS

At-Risk Buyers are customers who have deviated from their typical buying cycle and are at risk of never purchasing from your site again. Each customer’s typical buying cycle is unique – one customer’s buying cycle can be one purchase every two weeks, while another customer’s buying cycle is one purchase every six months. Therefore, customers can become At-Risk Buyers after varying time periods.

You can choose to filter for all At-Risk buyers by selecting the main At-Risk Buyers checkbox. To further filter At-Risk buyers, select a type of Buying Behavior: All buyers, One-time buyers, or Repeat buyers. To filter for At-Risk buyers who joined the audience segment within a specific timeframe, select the Became At-Risk in the last checkbox and enter a number for the number of days for your timeframe.

Example

We want to filter for all customers who became At-Risk within the past 24 hours to send them a “We miss you!” email with a discount code for free shipping. After selecting the At-Risk Buyers filter, we would select the All buyers and Became an At-Risk Buyer in the last checkboxes and enter 1 in the text field.

LOST BUYERS

Lost Buyers are customers who have a high probability of never purchasing from your site again; they have exceeded the threshold of their typical buying cycle. Each customer’s typical buying cycle is unique – one customer’s buying cycle can be one purchase every two weeks, while another customer’s buying cycle is one purchase every six months. Therefore, customers can become Lost Buyers after varying time periods.

You can choose to filter for all Lost Customers by selecting the main Lost Buyers checkbox. To further filter Lost Buyers, select a type of Buying Behavior: All buyers, One-time buyers, or Repeat buyers. To filter for Lost Customers who joined the audience segment within a specific timeframe, select the Became Lost in the last checkbox and enter a number for the number of days for your timeframe.

EXAMPLE

We want to filter for all customers who became Lost Customers in the past week. We’ll send them a “win-back” email that’s probably more aggressive than our other likelihood emails, like a sweet deal that’s too good to pass up. After selecting the Lost Buyers filter, we would select the All buyers and Became a Lost Buyer in the last checkboxes and enter 7 in the text field.

LIKELIHOOD TO TAKE ACTIONS 

Likelihood to Take Actions predicts the likelihood that the customer will Unsubscribe from Emails, Open an Email, Click an Email, or Convert (i.e., make a purchase in the next 14 days). These filters can be configured using the Audience Settings modal. 

To offer predictions around email engagement, a company needs to have access to email engagement data– which is difficult to obtain unless the company is an ESP (e.g., ExactTarget) or uses feeds to extract this data (e.g., AgilOne). Bluecore is an easy solution: we already have access to all of this data and do not need to hook into a feed to acquire it.

Likelihood to Take Actions is especially effective for unsubscribes. An easy way to reduce your unsubscribe rate on batch emails is to exclude customers with a high likelihood to unsubscribe. As these customers with a high likelihood to unsubscribe get fewer emails, their likelihood to unsubscribe will decrease over time, and we can start emailing them again. Bluecore can also an target those customers that are likely to unsubscribe through other channels, like Facebook or display.

Likelihood to Convert predicts the likelihood that the customer will make a purchase within the next 14 days. Targeting customers that are likely to convert soon has been limited to ad retargeting (e.g., Criteo), and triggered emails. Until now, marketers have been unable to create batched email sends targeting people deemed likely to convert based on their on-site behaviors.

USE LIKELIHOOD TO TAKE ACTIONS IN BLUECORE 

For each action, you can filter for customers who have a Low, Medium, or High likelihood to take the corresponding action.

When you add likelihood filters, the pie chart in the top right corner of each section will display the number of customers available for that filter.

EXAMPLE

We want to give some extra encouragement to Lost Customers (see Life-Cycle Stage) that have not opened our win-back email. We will filter for high likelihood to unsubscribe and send an offer through Facebook.

PRODUCT PREFERENCES

Product Preferences predict the products customers will be interested in buying, based on their preferences for specific product categories (Category Preference) or the probability of requiring a discount to make a purchase (Discount Preference).

CATEGORY PREFERENCE

Category Preference predicts the customer’s preference for specific product categories.

When companies use only past purchase and browse behavior to assign category preferences, they miss out on important signals. Bluecore associates behavioral data– beyond just purchase and view data– to email addresses instead of anonymous pixels, so we can target customers consistently across marketing channels.

Our category affinity models include co-view, co-purchase, and co-cart data to discover relationships between products and categories that may not be immediately obvious. Co-view, co-purchase, and co-cart are collaborative filtering models (i.e., “customers who viewed this also tend to view that). For example, if someone buys a TV we’ll know they have a certain level of affinity to a sound system.

The more complex a client’s product catalog is, the more interesting this predictive filter will be. Compare a site that sells fashionable items to millennial women (essentially one customer type) to a site that sells a wide range of electronics, like computer components to build your own computer, TVs, entertainment systems, laptops, software, hobbyist stuff, etc. An electronics site like this would have many different types of customers, and being able to push a new product to the right people is very difficult– which is exactly where Category Preference excels.

USE CATEGORY PREFERENCE IN BLUECORE 

Use the Add category filter dropdown to select the categories by which customers should be filtered. Then use the slider to select a preference level for the selected category.

EXAMPLE

Celebrate a sports team championship by sending to customers that love the Miami Heat. Set the Team is Miami and Category is basketball. We want to send emails to super fans, so we will select Very High Preference on the slider.

DISCOUNT PREFERENCE

Discount Preference predicts the customer’s preference to buy products that are full price or discounted, and to what extent. The Discount Preference filter can be configured using the Audience Settings modal. See Predictive Audience Settings for more information.

Typical offer affinity calculations may account for past purchase behavior, completely ignoring past browse behavior. This approach is limited because it does not extend to non-buyers, and can misrepresent preference when a customer has limited purchase data. To overcome these limitations, Bluecore computes a preference score using both browse and purchase data as a weighted average between the browse discount preference and the purchase discount preference. The relative weighting in the average is designed so that it assigns more weight to the purchase discount preference with each additional purchase the customer makes.

Marketers can distinguish between full price and discount buyers, and can use this information to offer a discount to discount buyers, and no discount to full price buyers. This has two major benefits for a marketer:

  • Don’t give away margin when you don’t need to. Giving a discount to a full price buyer is like giving money away.

  • Don’t discount your brand to your full price buyers. The more someone sees discounts from you, the less likely they are to ever buy from you at full price.

We can segment audiences by the degree of discount they typically need to engage with products or make a purchase. For example, we can find the group of people that have a preference for a 20% discount and separate them from the people that have a preference for a 30% or 40% discount. A customer will have an unknown preference if we do not have enough data.

USE DISCOUNT PREFERENCE IN BLUECORE 

Discount Preference predicts the customer’s preference to be a Discount Buyer, a Full Price Buyer, have No Preference, or has a preference that is Unknown. Select the preference designation(s) for which to filter. A customer will have an unknown preference if we do not have enough data.

EXAMPLE 

We want to filter for Full Price Buyers so we don’t send them emails about sales or discount codes. These customers don’t care about waiting for deals and will pay top dollar, allowing us to keep some of our margins.

REPLENISHMENT PREFERENCE

Note: This feature is in beta. For access, please contact your CSM.

Replenishment predicts which of your customers are “Replenishers,” those who have purchased a specific product or category in a pattern that implies replenishment, and the cadence on which they will want to purchase those products.

There are two main steps for identifying replenishable products and categories. First, we identify customers who have purchased a specific product (or category) in a pattern that implies replenishment, and determine their personal replenishment cadence for these products. The exact details of the pattern may be customized, but generally we are looking for a minimum number of repeat purchases which are spaced out over time.  

Once we have identified these customers (Replenishers) and the products they replenish, the next step is to determine which of these products have a minimum number of Replenishers. These products are now identified as Replenishable Products.  

Now that we have identified the Replenishable Products and their Replenishers, we are able to target the Replenishers at their dynamically calculated personal replenishment cadence each time we detect a new purchase. In addition, we target all new buyers of Replenishable Products at an optimized cadence based on the average of all the personal replenishment cadences of the Replenishers. This way, we are able to message the highly valued Replenishers at their own personal cadence, while motivating new buyers to become replenishment buyers at a timing which is automatically customized for each product.

USE REPLENISHMENT PREFERENCE IN BLUECORE 

Replenishment Preference determines which of your customers are replenishers and which of your products are replenishable. Use the eligibility dropdown to select eligible customers. Use the frequency dropdown to filter customers based on how many times they have purchased the replenishable product. You can use additional filters to refine the targeted audience.

EXAMPLE 

A replenishable product for a makeup store is mascara. We want to filter for customers who are deemed eligible to replenish their mascara within the last day and have bought the product at least once. We will email them a reminder to buy a new mascara before they run out.

Did this answer your question?