Predictive segments and data models

Smart solution directly from the data experts

Customers who like to buy

Up to 30 Affinity Models

Customers who like to buy

Up to 30 Affinity Models

The modelling of the propensity of customers for certain actions, e.g. purchases, clicks or a registration is called affinity in practice, or e.g. product affinity.

There are many scenarios for affinity models, e.g. purchases for certain product groups for different periods of time or predictions about when a customer will uninstall their app.

The purchase history and key figures derived from it, such as the purchase frequency, sales in different product groups, but also customer master data, such as the age, gender or place of residence of the customer, play a role as initial data.

Affinity models are particularly interesting for up-selling and cross-selling campaigns because sales can be increased in a targeted manner. Unused sales potential often lies in the range of 5-15%.

Best target groups

Integrated segmentation

Best target groups

Integrated segmentation

“If we have 4.5 million customers, we shouldn’t have one store . We should have 4.5 million stores.”
(Jeff Bezos)

Buyers’ preferences and behaviour are subject to constant change. In order for you to act accordingly, a complete and up-to-the-minute picture of your target consumers is crucial.

Example:
One of the most common methods for calculating customer value is the RFM model, a scoring method that classifies customers into different segments and target groups based on three metrics: Recency, Frequency, Monetary Value.
Recency: When did the customer last buy? How active is a customer, how high is the probability of repurchase.
Frequency: How often has the customer bought?
Monetary Value: What monetary value has the customer generated ( revenue or margin)?

Get the segmentation confidence and speed you need with our guided methodology, which is customised for segmentation and can be tailored to you.

Selections

Fast & easy queries

Selections - Machine Learning for all

Fast & easy queries

Enable your entire organisation to easily access your machine learning models and integrate them into your daily workflow.

It often doesn’t make sense to mix and match heterogeneous data from different user scenarios and systems.

Acceleraid enables transparent access to all data regardless of where and how it is stored.

Data selections and analyses that previously required the help of the data processing department can now be done by users themselves with the graphically oriented tools of the Prediction Engine.

This usually increases user satisfaction and productivity significantly.

Analyse and segment your customers with a quick and easy self-service interface at the touch of a button:

Build target groups quickly and easily
No IT resources required
Individualise sales for each customer group and thus increase revenue.

AI as an add on - what?

AI for existing Marketing Cloud

AI as an add on - what?

AI for existing Marketing Cloud

This is true: The Acceleraid Prediction Engine is a marketing cloud that integrates seamlessly with your organisation’s data collection software. So we can easily help you gain valuable insights from the data you collect every day.

Compatible with any IT infrastructure

Whether you store your data in Azure, already use various data science tools or have marketing automation tools in place, the Acceleraid platform is the perfect complement to your technology stack. Your data is cleansed and integrated so that it is optimally stored for your industry’s use cases. Already trained models can be used directly and converted into automation. Matching actions and target groups make your marketing automation tools even “smarter”.

As an AI-powered marketing cloud, the Prediction Engine enables you to reach your customers with an interesting, personalised message at the perfect time when they are ready to make a purchase.

Full feature list

ADTELLIGENCE AI ACCELERATOR

USE A SET OF INTEGRATED MACHINE LEARNING ALGORITHMS WITH ONE CLICK

Activity Level based on transactions

Activity Level based on transactions

HOW

Based on number of transactions per month/quarter the machine learning calculates the activity of each user. Updated daily

WHY

Not all the users are similar. One of the best ways to measure the willingness of a customer is through estimating their activity levels. Transaction activity is a prime indicator of a user's activity.

BENEFIT/ROI

Predict future behaviour, reactivate customer before churn risk increase

Activity Change Score (absolute)

Activity Change Score (absolute)

HOW

Calculates the absolute change in the activity level of a user in the given time interval.

WHY

For users with high or low activity levels, percentage change in the activity level might not be a good enough of an indicator. Best used in tandem with the relative activity change score.

BENEFIT/ROI

Gain insight to recent trends in user activity levels: target specific user groups based on the recent changes in their activity levels. Complements the relative activity change score.

Activity Change Score (relative)

Activity Change Score (relative)

HOW

Calculates the percentage change in the activity level of a user in the given time interval.

WHY

While being a universally accepted metric, percentage change values might not be representative enough to show the entire picture. Best used in tandem with the absolute activity change score.

BENEFIT/ROI

Gain insight to recent trends in user activity levels: target specific user groups based on the recent changes in their activity levels. Complements the absolute activity change score.

Optimized delivery of Content variations per campaign

Optimized delivery of Content variations per campaign

HOW

Considering individual differences and track record of your customers, our automated optimization algorithms can always send the most conforming content to each recipient in a target group.

WHY

The devil is in the details. The subtle differences in the characteristics and preferences of customers can lead to them being more responsive to different campaign content.

BENEFIT/ROI

Reach sub-groups of users without much effort. Catch the nuances among seemingly similar users and act on them. Increase customer engagement.

Revenue Prediction

Revenue Prediction

HOW

Accurately predicts how much revenue a particular product/customer group is slated to generate using various Machine Learning and AI models, scores and features we have built for you.

WHY

Anticipating how much revenue a particular configuration is set to generate can also help pre-emptively tackle some problems that may occur in the future. Also helps managers make better budgeting decisions.

BENEFIT/ROI

Accurate predictions on how much revenue you are set to cash in.

One-Time Clustering

One-Time Clustering

HOW

Using customized feature engineering and selection, we can build clustering models that are tailored to a specific need. You then can learn what kind of groupings can be found in your data.

WHY

There's almost always more to learn from data. Getting to know elusive patterns about your customers/products can help you target your audiences more effectively. Finding more about lookalike users makes it much easier to find efficient ways to reach them.

BENEFIT/ROI

Pin-point targeting of certain easy-to-miss user groups or products (or the unit of your own interest). Dives deeper and much accurately than the average customer segmentation method.

Variety Score

Variety Score

HOW

Calculates whether a customer focuses on a few product categories or rather displays an even spending behaviour among different product categories.

WHY

If a customer's spending portfolio consists of only a few categories, the customer can easily be encouraged to spend in other categories as well. If there is a strong preference for a handful categories, then the user's tendency to spend can be leveraged.

BENEFIT/ROI

Detect users with high cross-sell and up-sell opportunities. Identify those that are not really utilizing their potential. Target the users who are likely to spend and engage more.

Incentive Optimization

Incentive Optimization

HOW

Our algorithms identify users who are more likely to respond strongly to certain incentive-based campaigns. Learning from past experiences, our models can tell how much incentive is the correct choice for a given customer.

WHY

How incentive should be allocated to a target group to maximize the revenue is not a straight-forward optimization problem. Allocating sub-par incentives to users that require a little more stimulation mostly causes the incentive to go waste.

BENEFIT/ROI

Simply, squeeze out more revenue from your campaign budget.

Revolving Likelihood

Revolving Likelihood

HOW

Uses machine learning models to tell how likely a user to go "revolving", not being able to pay their entire credit card debt. Our models utilize a wide array of user and transaction features to tell you who to pay attention to.

WHY

Not being able to tell if a customer is only able to partially pay up their bill, can have severe ramifications on how accurate the revenue and cost predictions are. Finding users who are likely to "revolve" gives you the opportunity to address any possible shortcomings caused by it.

BENEFIT/ROI

Find out who is more likely to not pay their entire bill, help determine credit card limits more effectively.

Next best Category

Next best Category

HOW

Our models can tell if a customer has a high potential to be a frequent buyer in a certain category considering their previous activity records and customer characteristics.

WHY

Sometimes the users themselves don't know if they have a potential to spend in a particular category! Such users can easily go under the radar and result in missed opportunities. Encouraging these users to make purchases in categories that they are likely to be champions in, has immense benefits in the long run.

BENEFIT/ROI

Take advantage of up-sell and cross-sell opportunities created by your customers. Increase customer loyalty and your overall revenue.

Churn Prediction

Churn Prediction

HOW

Using the various foot prints the users leave behind and their characteristics, our machine learning models can tell if a user is about to become inactive, deactivate their accounts or become nonresponsive for a prolonged time period.

WHY

In most cases, users have some valid reasons that drive them away from you. Addressing these would encourage most of them to change their minds. But how to reach these users before they become inactive, rather than after they are already lost? That is the million-dollar (literally) question.

BENEFIT/ROI

Pre-emptively identify the users that are about to leave you. Target them and help them continue making profits for you!

Activity Level based on revenue

Activity Level based on revenue

HOW

Based on the amount of revenue generated by the user per month/quarter, the score approximates the activity level of that user. Updated daily

WHY

Not all the users are similar. One of the best ways to measure the willingness of a customer is through estimating their activity levels. Like transaction activity, how much revenue a user generates over a specific period of time is a prime indicator of a user's activity.

BENEFIT/ROI

Predict future behaviour, reactivate customers before they churn.

Automated Clustering for ”new” customers

Automated Clustering for ”new” customers

HOW

Our cluster models are built to find similar users with respect to their behavior and their characteristics. With the help of these we can tell how likely it is that a new user fits the user clusters we previously have discovered.

WHY

New customers are rarely blank slates, they have reasons to be here, and affinities to drive their behaviour. We can only do good by helping them reach their goal and more. One of the ways of achieving this is by identifying what a new customer might look like in near future. So that we can better address their needs.

BENEFIT/ROI

Make sure users are making the most of your products after their honeymoon stage by introducing them to opportunities without losing much time. Increasing their loyalty, activity and revenue.

Delivery Timing Optimization

Delivery Timing Optimization

HOW

In addition to the "how" of reaching your customers, our machine models can also find the "when" of reach them by finding the best time to send out your campaign content.

WHY

Timing matters. Even if you reach a customer, if it is too late or too early chances are that your correspondence goes unnoticed. These are the little things that make or break your campaign. Thanks to our algorithms, you can always know that your customers are contacted at the right time.

BENEFIT/ROI

Increases customer engagement, increases revenue, boosts campaign performance, prevents customer churn.

Lookalike Audiences

Lookalike Audiences

HOW

By analyzing the characteristics and behavior of your existing customers or users, our machine learning algorithms can identify patterns and similarities among them. Using this information, we can create "Lookalike Audiences" that consist of people who share similar attributes and behaviors with your existing customer base.

WHY

Instead of relying solely on your existing customer base, Lookalike Audiences allow you to expand your reach and target new potential customers who are likely to be interested in your products or services. This enables you to tap into untapped markets and increase your customer acquisition.

BENEFIT/ROI

With Lookalike Audiences, you can proactively reach out to new customers who are highly likely to convert and engage with your business. By targeting these audiences, you can optimize your marketing efforts, increase conversions, and ultimately drive more profits for your business.

Recommendation Likelihood

Recommendation Likelihood

HOW

Word-of-mouth is no longer a thing of mystery thanks to our models that calculate how likely a customer is to recommend your product to a peer. Again, this model utilizes the wide array of features at the disposal of our models.

WHY

Sometimes, for a customer, a recommendation might just be a little incentive away. Monitoring the likelihood of the customers to recommend your products can have big boost

BENEFIT/ROI

Acquisition of new users. Increase in user engagement.

Purchase Likelihood

Purchase Likelihood

HOW

Our machine learning models can also tell how likely a customer is to make purchase on a given product, using their characteristics and track record.

WHY

Sometimes it is just the first step that starts a long-lasting user-product relationship. Would you pass up on the opportunity if you knew that you could make a touch at the right place at the right time? Our models help you do just that.

BENEFIT/ROI

Recommend users what they are likely to purchase, increasing your revenue and customer loyalty and activity in return.

Customer Lifetime Value Prediction

Customer Lifetime Value Prediction

HOW

Thanks to the ability of our models to learn about your customers, they can tell you how much value a customer can bring over their lifetime during the time they use your products. Our models make the most of the data about your customers to generate lifetime value predictions.

WHY

We stated before that customers are rarely blank slates and that they bring their own set of affinities and preferences. Having an idea of what your customer is likely to bring helps you achieve their potential by using the myriad methods available to you.

BENEFIT/ROI

Predict future customer behavior, identify strategies for better targeting, increase customer engagement, increase revenue.

Success Stories

Financial Services

65.000 APPOINTMENTS FOR MORTGAGE ADVISORS

Find the right consultant directly through intelligent websites

65.000 additional appointments

+20% more appointments

30% faster appointments

Financial Services

DOUBLE-DIGIT INCREASE IN CREDIT CARD SALES IN FOUR COUNTRIES SINCE 2013

Through AI-based automation of customer approach, efficient partner onboarding processes and highly available server infrastructure

+25% more credit cards applications

+38% more payment protection insurances

99,9% server availability

Financial Services

APPROX. 100 MILLION EUROS IN ADDITIONAL CREDIT CARD TURNOVER

Through personalized email automation to deliver the appropriate content

>€100 Mio. additional sales

øRevenue per user doubled

>10.000 users at risk of termination reactivated

Financial Services

HOW AN ONLINE BROKER ACHIEVES MORE DEALS THROUGH PERSONALIZED PRODUCT PAGES

The right content for every visitor through AI-based optimization

+40% increase in clickout rate

8.000 additional entrances in the application process

400 additional closings

Financial Services

LEAD MORE USERS INTO THE APPLICATION PROCESS

Increase product sales through personalized customer approach in the Internet branch

+120% more credit card applications

+260% more current account applications

+200% more appointment requests for construction financing

Financial Services

INCREASED CUSTOMER ACTIVATION AND APP DOWNLOADS THROUGH AUTOMATED AND PERSONALIZED CAMPAIGNS

+120% more account activations

7.750 more vouchers sold

+16% more positive app reviews

Telekommunikation

20% MORE MOBILE TARIFF AGREEMENTS & 4% MORE CONTRACT RENEWALS

Through intelligent automation of customer approach using AI

+20% more mobile phone tariff terminations

+4% more contract renewals

Better understanding of the user journey

Telekommunikation

32.000 CONTRACT RENEWALS AND 16.000 CROSS-SELLS FOR A LEADING TELECOMMUNICATIONS PROVIDER

The right website for every visitor through AI-based optimization

32.000 additional contract extensions

+8% higher contract- renewal rate

16.000additional partner card cross-sells

Highlight

Machine Learning

Automated machine learning models boost productivity

Machine Learning

Automated machine learning models boost productivity

We add insight and functionality to every application to automate your data science.

No Code User Interface

Define and select target groups without programming

No Code User Interface

Define and select target groups without programming

All algorithms and models for specific use cases are already integrated in the system and only need to be configured by you, which does not require any data science or programming skills.

Predictive Segements (RFM)

„Machine Learning is the automation of data science. Automated Machine Learning models boost productivity when it comes to personalized customer interactions along the lifecycle and they highly increase scalability. If data is the new oil, this is the biggest and still nearly untouched oil field.“

Michael Altendorf
CEO & Co-Founder, Acceleraid

What data is used by prediction engine?

Acceleraid products

Acceleraid - approachable, smart, reliable

Simply add our AI products.

Experience it in action

We can easily integrate into your IT systems
Simply start Acceleraid has developed a simple and secure concept that allows you to get started quickly and generate your first revenues
Data import as well as export via API Integration via REST-API directly into your backend system as well as integration of output channels via standard interfaces
Hosting on premise Use Acceleraid as an intelligent data layer and host it in your data center, integrated into your processes.
Hosting in private cloud Use Acceleraid’s high-performance, high-security infrastructure and everything will be taken care of, from deployment to monitoring to security checks.
Individual integration solutions Do you have special requirements due to specific data protection regulations? Feel free to contact us!