Universal scenarios: The key to comparing personalization technologies

In part 1 of this 3-part series, I explored different options seating personalization capabilities in a larger marketing technology stack context. In part 2 of the series, I looked at different platform components required for building a holistic personalization technology strategy. In this concluding part 3, let’s explore some canonical scenarios that you can employ to support architecture and vendor-selection decisions.

Scenario analysis

While you should always consider product functionality and vendor predilections, the most important key to comparing technologies lies in how well they fit your particular business use cases, what Real Story Group calls “scenarios.” In our experience, scenario analysis provides the most efficient shortcut for finding the best-fitting solutions and architectures.

Explicitly or not, different personalization technology platforms target different use cases. This is usually because they were often created for addressing a specific need. Over time, they may have broadened their scope but initial roots remain visible — and typically decisive.

In this case, several personalization platforms started as simpler services for providing A/B testing. Some others started their journey offering website personalization and have broadened from there, though typically the newer services remain less rich.

Understanding the business scenarios that fit better or worse for the different platforms enables you to see deeper into their relative strengths, weaknesses, and architectural compatibility for your particular circumstances. Therefore, RSG has identified four common scenarios against which personalization platform vendors can be judged.

Before we get into details of each, some important considerations to keep in mind:

• These scenarios are abstractions. In practice, your own efforts here are likely to represent variants or a hybrid combination of scenarios. The cases overlap somewhat, but they are useful for understanding what types of products tend to work better for different types of projects;
• RSG uses these as model scenarios for evaluating vendors, including personalization platforms. However, in your own tech selection efforts, you should specify your own unique use cases against which to test vendors;
• The scenarios in the figure roughly form a maturity spectrum from left to right. As you move across that spectrum, you will need more preparedness in terms of capabilities required as well as a deeper understanding of how to deploy personalization services strategically. But as you mature, you can use these scenarios to reinforce each other; e.g., using Testing & Optimization to inform Ecommerce Recommendations.

Now let’s dive in on each.

Scenario 1: Experimentation

Experimenting with logic, content, design, and other elements is an almost universal requirement, and increasingly begs omnichannel capabilities. This is a scenario that most personalization vendors support. In fact, several personalization tools, including Optimizely and Adobe Target, started life with this scenario.

Common capabilities are:

1. A/B Testing, or more advanced A/B/..N testing: This compares two or more versions of content to see which works better for specific goals;
2. Multivariate testing: As the name suggests, it compares multiple variables, so you can compare combinations of different elements that can include not just variations of content elements (e.g., headline) but also variations of design elements (e.g., image or call to action); and
3. Optimizations based on test results.

Most tools now support machine language-based mechanisms to carry out splits, tests and optimizations. Where they differ — substantially — is in their omnichannel capabilities, e.g., the ability to stripe a single test across multiple different customer touchpoints. 

Scenario 2: Web Personalization

As the name suggests, this scenario targets web sites and applications; i.e., deploying personalized content or services on one of your own digital properties. Like testing, it can be based on behavioral and contextual signals, but increasingly enterprises are trying to leverage first-party profile data.

This sort of inbound personalization isn’t new, and some of you have been fiddling with rules engines for as long as two decades. Today, rules-based techniques are slowly giving way to Machine Learning-based algorithms, often based on session behavior rather than customer profile. RSG has found that enterprise experience with these techniques remains mixed, however.

Scenario 3: Outbound Personalization

This scenario caters to personalizing messages, mostly via email but also via text and in-app messaging.

Personalization here allows you to tailor message content to segments or individuals, and perhaps trigger messages based on behavior or events, as well as testing/optimizing outbound communications as you would inbound web experiences.

Some personalization platforms integrate with email marketing platforms. However, the sophistication of integration varies. A few platforms provide advanced capabilities for email templates and email content on their own. In other cases, the email marketing vendor itself may provide channel-specific personalization services.

Scenario 4: Ecommerce Recommendations

Online retail and ecommerce more generally is a special use case for personalization. Since it promises a direct impact — increased sales — vendors have focused on advanced capabilities here.

Key functionality might include product recommendations, cross- and up-sells, cart-related triggers and more. Several personalization solutions provide specific point solutions for ecommerce, whereas others integrate with ecommerce platforms. 

Machine Learning-based recommendations can also play an important role here in selecting the right audiences, or an optimal set of products, bundles, offers, and so forth.

Dig deeper: The first two parts of this three part series

What you should do

Scenarios offer the most useful initial approach for contrasting key strengths and weaknesses of different personalization platforms. There are at least two ways you can use these scenarios for your benefit.

First, scenarios can help you clarify architectures. In part 1 of this series, we addressed an important question: Where should personalization reside in your omnichannel martech stack? Business use cases should weigh heavily here. For example, if you’re really keen on just website personalization (second scenario) and nothing else, then channel-based personalization embedded into your WCM may not be a bad option to consider. But if you wanted to support all these scenarios, you probably need a dedicated Personalization Engine.

Secondly, once you’ve decided where it should reside, you can use scenarios to select the right products for your needs. At RSG, we use scenarios customized to your specific needs to generate custom vendor quadrants suitable for your requirements. You should follow a similar approach. Let me know if we can help.