Sitecore Personalize: A Guide to Decisioning

Leveraging Decisioning in Sitecore Personalize

According to the 2017 Gartner Customer Experience in Marketing Survey, over two-thirds of marketers say they compete primarily on customer experience. To gain a competitive advantage in terms of customer experience, businesses must focus on meeting customer needs through personalized experiences fuelled by data-driven decisions. Decisioning is one of three main features in Sitecore Personalize (formerly known as Boxever). Personalize is a SaaS-based marketing tool within the Sitecore Digital Experience Platform (DXP) that enables businesses to deliver personalized, relevant, and engaging content to their customers and recommend the next-best-action. It combines customer data, behavioural information, and machine learning to create tailored experiences such as personalized offers, messages, or actions for fast, accurate, and precise customer engagement. It can also integrate with platforms like Sitecore Customer Data Platform (CDP) to integrate more customer data aggregated across multiple sources.

Decisioning is the key to making personalization work. It's like the brain that helps marketers figure out the best way to interact with their customers and keep them engaged. This process takes valuable customer data, about what their interests and preferences and uses decision models and business rules to decide what to show them next and when. This helps create experiences that are unique and engaging for each individual customer.

This blog post will provide you with an overview of Sitecore Personalize's decisioning features and guide you through what it is, how it works, and why you need it. Here are all of the topics we plan to cover:

• What is decisioning?
• What is decisioning in Personalize?
• Key features of decisioning
• How decisioning improves business outcomes?
• Benefits of decisioning
• Creating decision models
• Connecting decisioning to Sitecore CDP
• Exploring the Decision Canvas
• Decisioning use cases

What is Decisioning?

Decisioning, also known as decision management, is a process in digital marketing that combines big data, business rules, and predictive analytics to make intelligent decisions about customer engagement. It enables businesses to automatically take action on customer needs based on this data, allowing them to create personalized offers, messages, and actions. By understanding customer behaviour and preferences, decisioning helps deliver relevant, contextual, and personalized experiences.

In today's digital marketing landscape, decisioning plays a crucial role in creating exceptional and relevant customer experiences. As customer expectations soar, having the right decisioning technology is vital for delivering personalized experiences that make customers feel valued and understood, while still driving business growth.

Implementing decisioning effectively requires vision, collaboration, and persistence, as it varies for each organization. Here are six important steps to consider for a successful approach to decisioning:

1. Think Big, Start Small: Personalization strategies often face internal challenges, particularly in organizations with legacy systems. Begin with one or two use cases aligned with strategic objectives to demonstrate value, gain stakeholder support, and initiate progress. Progress will be more important than perfection.
2. Embrace Decision-First Approach: Coordinate a company-wide effort focused on understanding customer experience objectives and target customers. Abandon traditional methods and standard offers in favour of a decision-first mindset.
3. Structure Your Business Around Customer Experience: Rethink traditional business structures to prioritize customer focus and facilitate collaborative thinking, enabling decisioning to reach its full potential. Build your organization around customer experience, and not the other way around.
4. Adopt a System of Operations for Decisioning: Establish teams, structures, and processes to monitor results, measure outcomes against defined KPIs, and continuously refine your approach to decisioning.
5. Recruit the Right Skills: Decisioning demands a diverse skill set. Assemble forward-thinking, adaptable teams capable of evolving alongside the ever-changing technological landscape.
6. Invest in the Right Tools: Technology is the foundation for decisioning. Equip your organization with a flexible decisioning engine that fosters collaboration and evolves with your business over time. Prioritize tools that empower your marketing and business users to accelerate their time to market and make delivering personalized experiences quick and easy.

What is Decisioning in Sitecore Personalize?

Sitecore Personalize features a powerful decisioning engine with a user-friendly, flexible drag-and-drop canvas for creating decision strategies. Utilizing various data types, AI, and programmable logic, it enables brands to achieve advanced decisioning and hyper-personalization at scale. A/B testing and experimentation helps optimize decision models to better align with business goals, while the intuitive interface allows for easy rule editing. Sitecore Personalize requires minimal technical knowledge, so marketers can easily deliver impactful, revenue-generating customer experiences that scale.

Decisioning in Sitecore Personalize starts when a customer interacts with an "experience" on your website or mobile app. In Sitecore Personalize, an "experience" is what your customers see and do on your app or website. This experience is linked to a decision model that you create using the Decision Canvas in Sitecore Personalize. Decision models help show different things to customers, like pop-ups, banners, or emails they get after triggering an event on your site.

For example, if a customer visits your online store, they might see a pop-up message with a special discount. The decision model can decide what discount percentage to show individual customers, making the experience more personal for them. This way, you can offer better experiences to your customers based on what they like or what they interact with on your website or app.

Key Features of Decisioning in Sitecore Personalize

Sitecore Decisioning empowers marketers to define, strategize, document, and execute decision management strategies using a best-of-breed visual experience. Sitecore’s approach ensures efficient collaboration and flexibility through a seamless separation of requirements, logic, and execution.

• Drag-and-drop decision models — Empower all users to use AI by easily assembling decisioning features on a user-friendly canvas.
• Decision tables — Construct and deploy straightforward or complex rule-based decisions using a table structure.
• Offer management — Craft dynamic offers and deliver any message to individual customers that connect with other systems, ensuring customers receive the best content on every channel.
• Programmable decisions — Access greater flexibility and decision-making power by coding custom decisions with JavaScript and other programming languages.
• Recommenders — Utilize powerful product, offer, and content recommendations, or configure your own. Sitecore Personalize supports everything from rules-based to collaborative filtering and machine learning.
• Artificial Intelligence — Connect to models hosted and executed in third-party systems, ensuring all your AI workstreams collaborate seamlessly through Sitecore Personalize.
• Data connections — Link to any web service to import data that supports digital decisions. Examples include dynamic product catalogs, live inventory, pricing, risk, and credit models. Use the retrieved data in decision tables, programmables, and recommenders.
• Decision model testing — Sitecore Personalize enables safe decision model testing before going live. Unit testing allows for model outcome assessment against specific segments and individual profiles, while silent testing simulates model performance against live traffic.

How Decisioning Improves Business Outcomes?

Decisioning uses important information about your customers to suggest the next-best-action for them based on what they need. This helps marketers create personal, 1:1 messages that show customers new products, give them useful updates, and let them share their thoughts about their experiences. All of this makes the customer's journey more personal and enjoyable, all while considering what's important for the business.

For customers, this means they get more personalize experiences, like:

• Product recommendations for customers chosen based on things like their age, where they live, or other information about them
• Different ways to communicate with customers, such as special offers for loyal customers
• The best times to connect with customers, reminders for things they left in their cart, and emails to keep them engaged and up-to-date

Here are more examples of other way you can use decisioning in Sitecore Personalize to suggest the best next steps for your customers:

Benefits of Decisioning in Sitecore Personalize

Sitecore Personalize's decisioning capabilities are built on the Decision Model and Notation (DMN) standard, which allows marketers to model and run decisions using business rules. The same decision models can be run using different experiments and across different channels.

Key features of decisioning in Sitecore Personalize include:

• Managing complex decisions through business rules to deliver the next-best-action or offer.
• Delivering superior customer experiences by ensuring decisions are accurate, precise, visible, and managed.
• Utilizing analytical models and data systems for maximum impact through intelligent connections.
• Testing decision model variants in real-time from the Decision Canvas and preview offers and images returned, just as they display on customer devices.
• Accessing full revision history and reverting decision model variants.

Fast, Accurate, and Data-driven

Sitecore Personalize’s decisioning is typically calculated in about 200 milliseconds or less. This speed can vary based on how quickly external data connections respond, as they play a crucial role in the overall decision-making process. By connecting to platforms such as Sitecore CDP, the decisioning is data-driven and accurate at delivering hyper-personalized experiences and product recommendations to customers.

User-friendly Canvas

Sitecore Personalize is designed to be user-friendly, with an emphasis on providing a simple and intuitive interface. Its drag-and-drop functionality and visual Decision Canvas make it easier for marketers and other non-technical users to create personalized experiences without needing advanced technical skills. You can easily move around rules-based ideas for your decision models and include sources of data, events that start actions, and create flexible decisions.

Creating Decision Models in Sitecore Personalize

The lifecycle of a decision model includes creating decision model variants, testing, and moving them into production. Consider following best practices when managing your decision model lifecycle:

1. Decision discovery process: Before building a decision model, think about what it should do, what information it needs, and how it should work, specifically identifying its outcome, required inputs, structure, and logic. Consult different stakeholders in the organization such as analysts, architects, data scientists, marketers, IT members, and people from different business units. This collaborative approach ensures a comprehensive understanding of the decision model's purpose and requirements within a business context, which is very important.
2. Create a decision model: Establish a decision model that serves as a container for storing all related decision model variants. This allows for better organization and easier management of the variants.
3. Create a decision model variant: After creating the decision model, develop at least one decision model variant to represent different approaches or scenarios. This helps in testing various strategies and selecting the most effective one.
4. Test the decision model variant: Choose a guest or a sample user to test the decision model variant. Evaluate whether the expected offers or actions are returned, and adjust the variant as needed to improve its performance.
5. Move the decision model variant to a test state: Transition the decision model variant to a test state, allowing for further evaluation and refinement before implementing it in a live environment.
6. Configure an experiment using the decision model: Set up an experiment that employs the decision model, initiating a silent test to gather data and insights without affecting the actual customer experience.
7. Move the decision model variant from test to production: Once satisfied with the silent test results, transition the decision model variant from the test state to production. This ensures that the variant is ready for real-world application.
8. Change the experiment status to live: After finalizing the decision model variant and moving it to production, update the experiment status to live, allowing the model to impact and enhance the actual customer experience.
9. Add comments throughout the process: Always add a comment to a decision model variant each time you save. This allows you to view a list of all revisions and comments, helping other team members and yourself keep track of changes, especially as the list of revisions can grow significantly.

Connecting Decisioning to Sitecore CDP

Decisioning in Personalize easily connects to Sitecore CDP to ensure decisions are fast, accurate, visible, and data-driven. Customer data is aggregated in CDP to create a profile for each customer. Once these profiles are created, decisions and next-best-actions are activated in Personalize based on the customer profile, as detailed in the table below:

Decisioning adds personalization to your customer journey that is fast and responsive. By using customer data, these responses can provide your customers with stories tailored to their interests and behaviour. Decisioning turns browsing into a value-added activity, giving your customers more relevant information in their shopping experiences.

Behind the scenes, Sitecore CDP makes an identity resolution to find who the customer is. Suppose the customer has selected a particular product. In that case, decisioning pulls in external price data and returns a personalized offer to the experience, which is then displayed back in the browser for the customer. The numbers illustrated in the image below shows the flow of data between a customer’s web browser experience to the decision model in Sitecore Personalize and Sitecore CDP. Here’s exactly what’s happening behind the scenes in more detail:

1. Browser to Experience — The customer interacts with a Sitecore Personalize experience on their desktop web browser.
2. Experience to Decisioning — The decisioning engine in Sitecore Personalize reviews the data from the experience through a Stream API.
3. CDP to Decisioning — As the decisioning engine reviews the experience data, it also pulls in data from customer profiles in Sitecore CDP.
4. Price API to Decisioning — The decisioning engine in Sitecore Personalize also pulls in external price data through an API. This data would be stored in a separate database of your choosing.
5. Decisioning to Experience — After pulling in data from Sitecore CDP and a Price API, decisioning returns the next-best offer or action to the experience.
6. Experience to Browser — The customer is shown a new customized offer or discounted price.

An applied example of this, would be an airline using a price API to return the real-time price of a flight. An external third-party system, accessible though a REST API, is called to retrieve the real-time price required for their search and displays it. Recommended products are also show based on their customer profile from Sitecore CDP which can include their point of sale, recent searches, order history.

If you’re looking to learn more about Sitecore CDP and Personalize as standalone, SaaS-based products, read our blog What is Sitecore CDP and Personalize?

Exploring the Decision Canvas in Sitecore Personalize

Now that we’ve covered the basics on how decisioning works in Sitecore Personalize, we can learn how to use the basic functions of the Decision Canvas. Here you can connect the different elements of your Decision Model and create Decision Tables to denote which segments of your audience will get each offer.

The following screenshot shows the Decision Canvas and its basic terminology. Everything in the Decision Canvas is customizable from a business user perspective. Instead of adopting a black-box AI model, Sitecore Personalize allows you to show other non-technical people in your organization exactly how the decisions are made using a systematic approach.

Using the Decision Canvas in Sitecore Personalize, you can:

• Customize the discount percentage displayed on website banners or in-app mobile banners for loyal customers according to their order data. By analyzing customers' previous purchases and spending patterns, you can offer personalized discounts that are more likely to resonate with them and encourage further engagement.
• Adapt the content shown to customers based on their session data. By tracking and analyzing customers' browsing history and in-session behaviours, you can present them with relevant content, products, or services, increasing the likelihood of conversions and enhancing their overall experience.
• Utilize data from third-party systems like credit score databases or price-lookup APIs to further refine your personalization strategy. Incorporating external data sources allows you to gain a more comprehensive understanding of your customers and tailor your offers and messaging more effectively. For example, you can offer personalized financing options based on a customer's credit score or display real-time price comparisons using data from price-lookup APIs.

The Decision Canvas in Sitecore Personalize

Let’s explore all of the terminology in the Sitecore Personalize Decision Canvas. Here’s a screenshot of the menu on the right side of the Decision Canvas and what each numbered element means:

1. Input Data — This section in the menu contains the data entities required by decisions to determine an output.
2. Guest — This component enables you to provide guest data (from Sitecore CDP) that is required by a decision to determine the output. Click and drag the component to the canvas and connect it to the dependent decision.
3. Order — This component enables you to provide order data (from Sitecore CDP) that is required by a decision to determine the output. Click and drag the component to the canvas and then connect it to the dependent decision.
4. Sessions — This component enables you to provide real-time session data (from Sitecore CDP), including event data, that is required by a decision to determine the output. Click and drag the component to the canvas and then connect it to the dependent decision.
5. Decisions — This section in the menu contains decision components that can perform a selection or calculation based on inputs, then determine an output as a result, such as the next best offer or action.
6. Decision Table — A decision table enables you to write a set of business rules in a tabular format to determine the next best offer of action, which can be edited on a separate screen. Click and drag the component to the canvas and then connect it from the required input component.
7. Programmable — A programmable enables you to use JavaScript in a decision, often to use nested data that is required by a decision table to determine the next best offer or action. These Javascript nodes can be customized by a developer. Click and drag the component to the canvas and then connect it to the required decision table.
8. Decision Template — A decision table enables you to write a set of business rules in a tabular format to determine the next best offer of action, which can be edited on a separate screen. Click and drag the component to the canvas and then connect it from the required input component.
9. Knowledge Sources — This section in the menu contains the Knowledge Source components which enable you to use the internal or external authority required by a decision component, for example, the offer to return.
10. Offers — The offers knowledge source enables a decision to access an offer catalogue set up in the platform, and to select the specific offer returned by the decision model variant. Click and drag the Offers component from the menu to the decision canvas and connect to it from the required decision component.
11. External Systems — This section in the menu contains the internal or external authorities required by a decision component.
12. Data System — A data system enables you to retrieve dynamic data from connections to external and third-party data systems and other AI/ML models. A typical reason for using a data system in a decision model is to retrieve a value from the data system to use as an input into another decision component.
13. Analytical Model — An analytical model enables you to use the latest machine learning technologies in your decision models. A decision model can pass parameters to an analytical model to retrieve data or perform a calculation, then pass it back to a decision component.

Decision Model Variants in Sitecore Personalize

A decision model variant consists of components that you can easily drag and drop from the Sitecore Personalize menu onto the Decision Canvas. When structuring a decision model variant on the Decision Canvas, start from the bottom and work your way up. Place the ultimate decision at the very top. To ensure optimal performance, have only one ultimate decision. A consistent structure in a decision model variant allows users to quickly identify the main decision by looking at the top of the decision canvas.

The following screenshot is an example decision model variant. Keep in mind that the following components from our exploration of the menu earlier:

• The yellow input data buttons (guest, order, and sessions) represent data from Sitecore CDP.
• The purple decision rectangles are editable business processes
• The red knowledge source is a database containing current offers configured in the system.
• The green external data system might include a CRM like Salesforce or an AI model in AWS.
• Lastly, the blue decision table is the ultimate decision in our model that we want delivered as the next-best-offer to our customers

Decisioning in Sitecore Personalize: Use Cases

Example: Credit Score Check

A customer visits their bank’s website to check their credit score. The website sends behavioural data into Sitecore CDP through stream events using the Stream API. When a customer initiates the credit check, a full stack experience and decision model will call a credit check system using a Credit Score API. The customer’s credit is checked in real-time and a specific product is then displayed.

Example: Personalized Shopping Cart Offer

A customer is in the process of ordering a product from your website and adds it to their cart. Sitecore CDP captures their order data and decisioning calculates a discount based on their customer profile. We can calculate and deliver different discount percentages based on if they are new to the website, if they’ve made a purchase before, if there are other products in their cart, etc. The full stack experience triggers a personalized message with a specific discount offer for the customer.

Example: Product Promotion Website Pop-up

A customer visits a specific page on your website and triggers a web experience that displays a customer-specific pop-up message. These pop-ups can deliver different products, content, or offer recommendations matched to your customers’ current context and historical behaviour. Deliver a personalized message to first time visitors or a loyalty discount to repeat customers.

Example: Custom Discount Using a Mobile Banner

A customer visits your mobile app. The app sends a request into Sitecore CDP full stack interactive experience that pulls in customer order history and behavioural data and displays a personalized product banner with a custom discount on the mobile app. By connecting external third-party systems such as a price API, you can price your products in real-time. Airlines are a prime example of how dynamic pricing models are crucial to delivering great customer experiences while still prioritizing business context.

Example: Abandoned Cart Email

A customer is adding products to their cart, but abandons it and doesn’t complete their order. The experience looks up a rule indicating that their order wasn’t completed and the customer is sent an abandonment cart email through their email service provider. A custom discount code or personalized message is delivered to encourage the customer to make a purchase.

In Closing

Decisioning is a powerful tool that enables marketers to deliver highly personalized experiences to their customers. By intelligently analyzing customer data and preferences, decisioning helps determine the best way to engage each customer, resulting in more effective and personalized interactions. This not only enhances the customer journey but also leads to increased loyalty and improved business outcomes. In today's competitive market, it's crucial for marketers to leverage decisioning to stay ahead of the curve and truly connect with their audience.

If your team is ready to invest in personalization and want more information about how Sitecore CDP and Personalize can help you get there. Contact the friendly folks on our marketing team and get started today!

Until then, happy decisioning!