I’m sure by now you’ve heard of Interaction Studio — or Salesforce Marketing Cloud Personalization (SFMCP) as it is now called. But if you’re similar to me, none of the buzzwords that are thrown around help to paint the picture of what SFMCP actually does.
In this blog, I’ll be demystifying this intriguing product and detailing the features and functionality, and where possible I’ll be calling out any “gotchas” that have personally caught me out, so you can avoid them.
What is SFMC Personalization (Interaction Studio)?
So, let me start with the easy part… What is it? Simply put, SFMCP is a real-time personalization engine. By integrating it with your website(s) and mobile app(s), you can track visitors as they interact with your digital content.
Every page, article and product they view, click on, rate or add to cart is providing SFMCP with the data it requires to build a unique customer profile for every visitor — whether they are known or unknown.
In other words, it allows you to dig deeper and truly understand what keeps bringing each visitor back, and when coupled with data provided by a CRM or data warehouse it is capable of delivering a hyper-personalized experience at scale and in real-time.
What does SFMCP do?
Although accessed via Salesforce Marketing Cloud (SFMC), that’s where the similarities between SFMC and SFMCP come to an end. Having been procured rather than built by Salesforce, it has a very different look and feel.
SFMCP is broken down into the following:
- Channels & Campaigns
- Machine Learning
By default, all users land on the Dashboard (as shown below) which provides a high-level view of revenue, visits, orders, average order value per visit and conversion rate. It also provides details on how your campaigns are performing through metrics such as impressions, new vs. returning visitors and item actions (purchased, viewed, added to cart, etc.), all of which are filterable by date ranges.
However, to dig a little deeper, SFMCP also provides a range of Activity, Results and Visit reports that can help understand how your personalization efforts drive success.
- Recent Visits: Stream view of all the visits occurring over the defined period
- Event Stream: Stream view of all the events occurring over the defined period
- Cohorts: Groupings of users by predefined criteria such as first time visiting, first time performing an action or specific statuses
- Actions: Linear graph showing key Sitemap actions specified that can be filtered by segments or keywords
- Modules: Volume of visitor actions, views/clicks that have been grouped together as Modules
- Paths: A page-by-page view of the paths visitors take when interacting with your site
- Funnels: Using Modules, create an optimal funnel for your visitors and monitor the completion of the desired sequential actions
- Day / Time: View visits, visit length, revenue, revenue per visit and bounces by day of the week and time of day
- Revenue: User, order and revenue statistics in a grid matrix and filterable by platform, engagement, visit status, and more
- Purchase Funnel: Using segments, compare how your audiences perform against all visitors when viewing products, adding to a cart and purchasing
- Goal Completions: Side-by-side comparison of your configured goals and filterable by device type, browser type, user state and more
- Goal Comparisons: A view of the goal completions and completion rate over time for created goals
- Behavior: Shows visits, engagement, logins and purchases over a given time frame
- Technology: Includes information such as browser, device, and operating system commonly used by visitors
- Referring Sources: Useful report to show how visitors access your site as well as the referring sources driving the most revenue
Channels & Campaigns
In short, it’s the marketer’s space to build campaigns and whilst it is often the most talked about, web is not the only channel SFMCP can support, it also offers mobile and email, as well as triggered and JS server-side campaigns.
Using web campaigns, SFMCP users can create real-time personalization experiences aimed at targeting visitors who meet certain criteria. Server-side campaigns give users the ability to get more creative but will require a greater skill set.
Triggered campaigns and email campaigns tie into the wider SFMC platform if you’re lucky enough to have it, by triggering Journeys or by providing Open-Time Content to include in your SFMC sends.
Through mobile campaigns, SFMCP can incorporate mobile as an additional channel by including a personalised experience on both Android and iOS.
And finally, a Third-Party section allows SFMCP to bring in data from systems such as Marketo, Act-On, LinkedIn Ads, etc. and even create custom products when and where required.
As it sounds, Audiences are where users create visitor segments. Segments are a large part of SFMCP and go beyond simple segmentation. In SFMCP, segments can be used to display AI recommendations to a selection of visitors based on your pre-defined criteria. But they can also be used to create goals by setting the actions you wish visitors to perform (i.e., order a minimum of $xxx).
The number of visitors who meet the criteria directly relates to your goal’s success. Furthermore, any segment created can also be added as a filter to ensure certain visitors are excluded from recommendations and personalised content (i.e., exclude visitors from the US for UK-specific campaigns).
The audience section is also where you can see all users in your SFMCP account, whether imported via feeds or tracked through the SFMCP Sitemap. However, it is important to note, that visitors captured via the sitemap will be anonymous (created using a randomly generated profile ID) until they perform an action (i.e., create account, checkout, etc.) that provides SFMCP with a chosen identifier such as an email address or username.
The Catalog section is where you begin to add business context and it will require plenty of planning. SFMCP offers standard catalog objects such as Products, Categories, Blogs and Articles but also provides the opportunity to create custom objects to support your particular business needs. The objects can then, where required, be related to each other with varying cardinalities that will provide SFMCP with the information needed to understand visitor interactions.
For example, when a visitor clicks on a specific pair of trainers it will increase that visitor’s affinity towards that particular product, but it will also create an affinity to the categories of running and shoes. In turn, products in the same categories can then be suggested through recommendations and promotions.
Note: An important callout here is to distinguish which method will be used for consuming data: ETL or Sitemap. Passing the same attributes, such as stock, through both methods can cause issues in SFMCP. I’d recommend leveraging the ETLs where possible and using SFMCP’s Sitemap for passing IDs and activity types to avoid data discrepancy issues.
If you’re like me, this is where the fun begins. The machine learning section is all about recipes and decisions. The Machine Learning piece can be broken down into two main capabilities, Recipes and Decisions,
As the name suggests, by mixing ingredients SFMCP can create unique content or product recommendations for each customer.
Ingredients make up the basis of Einstein’s machine learning-powered algorithms. Elements such as co-browse, trending or smart bundle (to name a few) will show recommendations based on what other visitors have browsed, what’s currently trending or what products other visitors have bought together respectively. It’s important to note that these recipes can even include offline activity that is imported via ETL feeds.
Rather than manually creating and testing the recipes yourself, why not let AI do the hard work for you? This is where decisions come into play. Decisions use visitor data to predict and display the right promotion based on the chance of completion and the highest business value.
Einstein automates the process of deciding who should see what content by evaluating each promotion viewed, whether or not the session is a returning visitor, device type and much more.
Gears and Feeds
Gears are extensions for the platform that can be added based on your specific needs, although most will be enabled by default. The most important gear, in my opinion, is the Flicker Defender which, when implemented correctly, stops SFMCP sitemap from flickering on page load.
Feeds, on the other hand, are SFMCP’s Secure File Transfer Protocol (SFTP). SFMCP has a specific feed and format for each standard object, custom object and related catalog object which you can use to ingest data to support your catalog setup. Sharing data with SFMCP provides the information it needs to turn anonymous clicks into insightful data that can be used to make informed decisions that ultimately impact the bottom line and revenue.
Further Considerations Before Jumping Into SFMC Personalization
Before jumping straight in and procuring SFMCP, it’s worth evaluating the following considerations:
- Personalization Requirements – As its name suggests, SFMCP is well suited to real-time personalization. But if your requirements are undemanding and single channel, it might be worth considering alternatives such as SFMC Web and Email Recommendations. However, if your requirements are omnichannel in nature and supported by a complex catalogue of data, SFMCP might be the right tool for you.
Once you’ve made the decision to procure SFMCP, the final consideration you should take into account is the implementation approach itself — you should begin reviewing your short-term gains vs long-term strategic objectives.