The Most Overlooked Benefit Of Einstein Recommendations

27 Sep 2022 in

Einstein Recommendation Builder is Salesforce Marketing Cloud’s easy-to-use recommendation engine for products, content or banners. The way it works is quite simple: it requires an installation of snippets on the website so a cookie can track the user behaviour and return recommendations of content for the user. This way, the user will find the content he/she is looking for faster; dropping the bounce rate and thus increasing the conversion rate.

Great technology for web and email but what if this is not the only thing we can use Einstein Recommendation Builder for? What if we use it as a data collection tool for our subscribers and use that data to build highly personalised 1:1 communication!

What’s Powering Einstein Recommendations?

At heart, Einstein Recommendations in Salesforce Marketing Cloud is powered by tracking actions that a user performs on a website or email. These actions are tracked similarly to Google Analytics and then stored in Salesforce. The algorithms behind Einstein will bring this data together and create sets of data that (1) have a correlation with one another and (2) build a profile of each user. It’s this latter point that is being overlooked by Salesforce’s customers!

Below are 2 examples of an “Affinity Profile” of 2 different subscribers within one of our hospitality clients.

Affinity Profile

Affinity Profile

Based on these 2 profiles, you can clearly see that they are 2 different subscribers. Not only are they searching for hotel stays in different geographies, but one is more interested in adventurous outdoor activities while the second is more interested in fine dining.

All of this data is based on the tracking data collected by Einstein. So next to Einstein Recommendations, how can we leverage this? Below you can find a few ideas.

One Step Closer to 1-1 Email Personalisation

Emails have a limited amount of real estate to show products you’d like to promote so why not personalise this based on the recipient's affinity profile. A lot of retail brands send a single weekly newsletter containing the latest promotions or products they would like to push but what’s the point in pushing a product to the wrong gender?

We reckon that the best approach is to list the favourite categories based on the affinity profile and push products based on preferences. A good example would be that if a man purchases products for adults and boys (vs. women and girls) then a better personalised email should contain products from only those 2 categories. If a woman only looks at products for women and nothing more, then the choice is obvious as well!

Advanced Audience Segmentation

Stepping up your segmentation game is another possible benefit of using Salesforce Einstein’s Affinity Profiles. If you need to push a certain product or service, you can base yourself on existing data from Service Cloud by checking preferences and or historic purchases but you can expand this segmentation by checking the Affinity Profiles. You’ll get a wider audience based on the most recent tracking data.

Particularly useful if product stock levels remain high, to generate a boost in sales, you can target those with a high affinity rating for that product or the category it resides in.

The reverse is also an option, products that just returned in stock might have missed sales recently, what better way than to segment your most likely customers within your data set.

Bonus Tip for Einstein Email Recommendations

There are also ways to leverage data generated from Einstein to reduce costs with Salesforce Marketing Cloud. Collecting data for Salesforce Einstein is free and super messages are only charged when Einstein Email Recommendations are generated & shown. If a profile is unknown, Einstein will revert back to the wisdom of the crowd to provide recommendations; which is not necessarily a bad thing, but a simple piece of code (AMPscript) can be added to check if the subscriber is known or not in Einstein. If known, we can then display the recommendations as per normal whilst not known, you can add your own default set of products.

From Good to Great: A Framework to Create Salesforce Marketing Journeys

14 Sep 2022 in

How well are you designing your journeys? What are you considering when you are designing your Salesforce Marketing Cloud Journey? How many journeys are you planning to build and how do they all relate? Which segments do they cover? All these questions are very important for any marketer as well as any Salesforce Marketing Cloud implementation and below Transformative will share a little peak at our framework in helping create Salesforce Marketing Cloud Journeys.

Think about you want to achieve

When you drill down from your journey map, you’ll end up having to think about the specific tactics needed to achieve those specific goals. If I want to increase retention through loyalty, what do I as a business have to offer to my customers that increases loyalty. How are these offerings currently being communicated and are they being underutilised? Let’s drill down one more level; how many touch points do you need to communicate your message and which channels would communicate this message best?

Which triggers are available

Choosing an entry point is also incredibly important for engagement. Behavioural triggers such as sending an abandoned cart email to a customer generates more engagement because there is buying intent, relevant content and excellent timing. These 3 elements combined make communication effective; regardless of the channel used.

Think about the content

As mentioned, relevant content is incredibly important. People open an email but stay for the content. If the recipient has a dedicated account manager, why not show the account manager’s name and profile in the email to more easily converse? If you are sending a weekly newsletter for your online store, why send male products if the recipient has a purchase history for female and children products only? It would be a waste of valuable real-estate. And what value does the message bring to the recipient? Which channel is best to inform my customer? SMS works incredibly well but is expensive, it should be used to directly communicate with customers with time-sensitive information and you are sure people will read it! Emails tend to stack up so important information gets lost or is read too late! The younger generation also doesn’t read emails, however, because they are tied to their phones, SMS has also made a comeback! But the content you can communicate in an email is vastly larger than in an SMS, so think about the content and how you can bring value to your customer.

Think about the KPI’s

The final step is to assign KPI’s to each communication and or journey! After all, you want to know how effective they all are. If your abandoned cart email converts both online and offline, you need to measure both metrics. (ie, your brand provides tailored suits, someone goes online to check which suits he or she is interested in and then goes to the shop to have professional advice and some tailoring.) Another example, is your reminder email working? If the KPI’s indicate it doesn’t, then why doesn’t it? Maybe it will work better if you add a sense of urgency or convenience. Does this increase performance with regards to conversion, click & open rate?

Let’s visualise this in our framework

Below is a single step in your customer journey, you would attach several of these together to complete a full customer journey and if you are interested to get the full overview, please download our ebook right here.

A Framework Marketing Journeys

For each step in the journey, we would define who will receive this message and what behaviour got them there. Followed by choosing and defining a piece of content to be delivered and through which channel. Then defining when would be the best time to deliver this content and which KPI’s should we attach. Once this is defined for the first step, we could continue with this framework for each of the steps following this journey.

Keep in mind the big picture

Lastly, as the title suggests, it is important to keep in mind the big picture. Are you excluding subscribers when they enter a particular journey from all the other journeys? Or do you want them to be in 1 or more journeys simultaneously. Some clients want to make sure that anyone and everyone is in a journey at all times. But this a topic for another blog post!

How to Leverage SFMC's Einstein Recommendations

13 Sep 2022 in

What Are Einstein Recommendations

Web & Email Einstein Recommendations have multiple benefits. Let’s look at some of them: 

  • Einstein Recommendations Web – This feature improves your website by personalising your customer’s view by displaying products or content based on Salesforce’s Einstein algorithms which in turn looks at the wisdom of the crowd as well as relevancy based on attributes you associate with the catalogue.
  • Einstein Recommendations Email – This feature improves your emails by personalising your customer’s view by displaying products or content based on Salesforce’s Einstein algorithms and calculates the next best thing upon opening (not sending) the email.

How to Leverage Einstein Recommendations for Web

The greatest benefit of Einstein Web Recommendations is that they are based on First-Party cookies which means all incoming traffic on your website is being tracked. The cookies will therefore recognise returning visitors as well as provide recommendations based on their current & previous online behaviour. Below you will find the 5 best places as to where to install Einstein Recommendations for Web.

Home Page (Banner recommendation)

It’s important to personalise not only the products for sale on your website but also to tailor the look and feel of the website to your audience. If you provide insurance services, you might wish to display images that reflect your visitor’s age or family situation whilst if you sell sports gear for different sports, you might wish to display the main images concerning the most relevant sport more prominently on the home page.

Home Page (Product recommendation)

When a visitor has not returned as a customer, the visitor might still be in the consideration phase. It is important to remove as many hurdles as possible which includes allowing the visitor to find the products he or she viewed recently more easily. By presenting the recently viewed products on the homepage, the returning visitor will need less clicks to find the product and in turn reduces bounce rates and improves our chances of converting.

Category Page (Product Recommendation)

If you look at it from a UX perspective, most ecommerce websites are based on category pages. If you from the main page, you drill down to categories from which you drill down further to find the product you are looking for. So the category page is key and that’s why it is extremely important to frontload the products in each category based on recommendations tailored to each visitor! Imagine a google search but on the category page. Very important if colour or design is important to a visitor, that the visitor sees this immediately upon clicking a category.

Cart Page (Product Recommendation)

This one is especially interesting when there are a lot of accessories to a product. Think matching clothing based on a lookbook or batteries if you are selling electronics. It’s the final upsell before a sale happens and has a place in all check-outs.

Product Page (Product Recommendation)

When we first speak about adding recommendations to a client’s online store, we generally speak about this one. Not because it is the highest performing (see category page) but because it’s the classic. People who like this product also like this product and Amazon made it mainstream. 

Usually when visitors have reached the product page, they found what they are looking for so adding product recommendations on the product page is more like adding a back-up. It’s there in case the visitor is still looking for something but can’t really find it. We tend to see lower CTR and conversions with this block.

How to Leverage Einstein Recommendations in Email

The greatest benefit of Einstein Email Recommendations is that they are based on the same First-Party cookies as the Web Recommendations are which means you have a wealth of data to use in Email. The great part about it is that people build their tracking on Einstein before they sign up to your mailing list but also as soon as they start clicking in emails, they’ll be adding to their tracking data. Below you will find the 3 best ways to use Einstein Recommendations for Email.

The Welcome Email

One of the key benefits of having Einstein running on your website is that it is collecting information about your visitors based on first party cookies. Once that visitor has signed up to your mailing list, then that subscriber is considered identified and we can apply Einstein Recommendations in Emails. 

Imagine if the first email you ever send to a new subscriber is completely personalised? From the header to the content of the email, all tailored based on gender, preferences and budget? You’ll leave a strong impression. Not only does this improve conversions, but the look and feel of an email drastically improves based on each new subscriber.

Product Recommendations per Category

One of the obvious differences between an email and an online store is that emails have limited real estate when it comes to displaying products. If you wish to display 4 rows of 4 products in your email which tends to be the usual amount, you are only able to display 16 products. This while your online store probably contains over 300 products. Therefore it is incredibly important to use that real estate as efficiently as possible and you do this by making sure that the products displayed are personalised based on the recipient.

Recently Searched

The search bar is such an underutilised tool in marketing. Your customer is directly inputting what he or she is looking for and we don’t use this data enough for retargeting. Einstein isn’t a search engine but it could help your visitors in finding the products they need by providing recommendations based on the search term and matching those with what others have bought