I'm wondering if I'm using the right way to implement a default program with personalised emails, let me explain :
I created two versions of the same email, but with different greetings and use a smart campaign to differentiate recipients :
"if member of list..", send Email with Token 1 ("Chère Madame")
"If member of list ...", send Email with Token 2 ("Chère + FirstName")
My issue is the following : in the next phase of my scenario, I send a different email depending of the behaviour of the recipients : if click link in Email ... send ... or not
But the link in both emails is the same... is it an issue? I suspect it is, because when I test my scenario, I don't get the coherent greeting... (e.g. : Chère Madame in both emails but I get : Chère Madame and then Chère + FirstName)
Here is an example of the first flow : (hope you can read that! sorry it's in French and Marketo keeps changing the default language...: step 2 = Send an email)
Can you help please? Any better/simpler way to do this?
Make the list a field or segmentation instead and use a Velocity token. Using lists and diff assets for this kind of basic personalization leads to major confusion.
Why the two lists for "Madame" vs "First Name"? If the "Madame" list is for people whose first name you don't know, there is a much simpler way to implement that doesn't require lists at all. Simply use "Madame" as the default value in the token. If First Name is empty, the token will automatically use "Madame" instead:
(I assumed this was for A/B purposes as opposed to not knowing the First Name.)
You may be right. If it's for A/B split purposes why not just use random sample rather than split the audience into lists?
Dunno. I would've started with Velocity anyway. Not necessarily half-and-half?
That makes sense for you. For someone (like me) who doesn't know how to use Velocity, random sample is an easy solution. It doesn't have to be half-and-half. You can use Random Sample to split into multiple segments.
Sort of... but it's not really random and can leave orphans when there's a small # of members.
In any case, we don't know that the goal could've been met with sampling. Let's assume the members of each cohort could not be defined reliably in any other way but Static Lists. It would make sense, then, to either: