Litmus recently published a free report (see link at the end of this post) that contains predictions by experts about email marketing in 2020. The very first one they highlighted was this:
How hyper personalization will be crucial in delivering the right content at the right time.
I wholly agree: personalization is a major trend in digital communications, email included. I would like to propose a distinction between two types of personalized emails: Superficial and Deep. I prefer ‘deep’ over ‘hyper’ because there’s too much hype in ‘hyper’.
Superficial personalization refers to what we used to call ‘mail merge’ in the good ol’ days, whereby you utilize a token to display, within the email subject line or body, the recipient’s first name, company name, salutation, etc. This type of personalization is expected by your audience and failing to use it means you’re not doing a very basic part of your job as a database marketer.
Noticed I wrote ‘expected’ there? What I really mean is, it’s non differentiating. When you use superficial personalization, you’re simply leveling yourself with the market’s standards. You’re doing what you should be doing because everybody else is, and not doing it will actually differentiate you as worse.
To differentiate, you need to go a level or two deeper into your data.
What do you mean, ‘go deeper’?
Simply put, I’m referring to the non-trivial manipulation of data, in order to significantly increase the degree of relevance that your email message has with regards to a specific recipient. Here’s an example: let’s say that you want to tell someone that an X number of their peers from the same region have already signed up to a certain activity or otherwise responded positively to an offer that was made to them. Clearly this isn’t available as a ‘flat’ database field, and requires data manipulation. The result of such manipulation is what I call Deep Personalization.
OK, and how is it done?
There are a couple of approaches for manipulating data in order to achieve deep personalization in marketing emails.
Manipulating Data via Email scripts
This requires your email marketing automation software to support server-side scripting. Marketo includes, as part of its email capabilities, a scripting language called Velocity, a language based on Java that allows the scripting of HTML elements, and can access the Opportunity and Custom Objects of a Marketo-Salesforce instance. Scripts are executed upon the email’s release from Marketo’s servers for distribution.
I’m not familiar with other marketing automation vendors that offer email scripting similar to Velocity, but if you do fee free to share in the comments below.
Manipulating Data in the CRM
The other, probably more conventional approach, would be to create custom fields in your marketing automation-connected CRM software (typically Salesforce), and populate them using CRM formulas or scripts.
Regardless of the approach you choose, you’re likely to require assistance from a developer proficient in the relevant scripting language & methods, so do take that under consideration when assessing such projects.
Alright, now that we’ve figured out how what deep personalization entails, let’s consider a few other use cases:
Fetching data from custom objects
Probably the most common use case for deep personalization. The fact is, depending on your CRM data structure, there are other objects in your CRM besides the Lead or Contact that hold interesting data about the persons or accounts you want to communicate with. It could be anything, really: products, support tickets, campaign associations, account related opportunities, partner information, and the list could go on and on.
Here’s a simulation I did of a non-personalized vs. personalized email, based on a real one I received via one of my (too many) subscriptions:
Which version do you think would have resulted in more responses? Well, this wasn’t really tested of course, but whenever we’ve tested personalized messages vs. non personalized for our clients, personalization proved the better. Let’s explore a few more examples.
One of our clients, a SaaS provider, manages a custom object that holds information about their users’ SaaS instance, and we reference this information in specific campaigns using email scripting.
You may want to send an email about a product upgrade, and you want the email to reference the contact’s actual product, information that is stored in the Product custom object nested inside the Contact’s parent Account object.
Or, you may want the email to mention all of last year’s marketing events that the recipient actually attended, information that is typically stored in the Campaign object to which the lead or contact is associated.
Perhaps you want to send an email to customers who were referred by your partners, providing contact instructions that change based on the partner, information that is stored, you guessed it, in the Partner custom object…
Now, most email marketing and marketing automation platforms have solid integrations with CRMs like Salesforce, but it only goes as far as being able to exchange data about the Lead or Contact object. Try and access anything deeper than that in your CRM data model, and things get more complex and vary greatly between the marketing system vendors. Hence the need for data manipulation via email or CRM scripting.
Displaying data that results from a mathematical calculation
Let’s say that you want to send out an email campaign to your clients offering a discount if they upgrade before the end of the month. You could use a generic phrase like ‘get 10% off’, but what if you could add ‘and save $XX’, where ‘XX’ is exactly the discount value that each recipient will get based on their individual rate? This could have a positive effect. At least, it’s worth A/B testing, no?
You could also use a calculation showing the number of days remaining until a certain date, where the date will vary per recipient.
Displaying data that is based on a logical calculation
If-Then-Else decision trees, simple True/False options, complex AND/OR operator combinations – logical calculations can seriously boost the personalization effect of an email by making it a notch or two more relevant to the recipient. Examples? Aplenty, here are a few:
- Show the exact hour per the recipient’s US state-denoted timezone.
- Use the right job level denomination (Director, Executive, etc.) based on whether the Job Title field contains certain keywords or not.
- Display the email footer in the recipient’s language based on whether the fields Preferred Language and/or Country have values. For example, if Country = FR and Language Preference = German show the German text, but if Language Preference is empty, show the French text.
- If Country=US/Liberia/Myanmar then use imperial system for distance units, otherwise use the metric system. (you did know that only three countries in th.e world don’t use the metric system, right?)
When should you use deep personalization?
Simple: use it whenever the effect, or potential value, of showing your recipients you truly know them and can be relevant to them is greater than the effort of implementing this data manipulation 🙂
On a more serious note, even if you end up not implementing deep personalization, I hope this post at least got you thinking about the value that’s locked inside your data, and that next time you add the First Name variable to your subject name, you’ll at least spend a moment thinking about other ways you can make your audience connect to your message.