Selecting an Nth of a file - Please help ASAP

[Anonymous]

Hi,

Need some help on this.  I'm creating a smart campaign to select an NTH of a static list to mail over the course of several days.  I want to be able to mail new names from that list over consecutive days, without repeating people that received a message the day before.  Eventually all should be mailed from that list at least once.  Here's what I have so far:

WHO:  Member of List:  List Name

What:  If Random Sample is 5% then email:  EMAIL Design Concept X

Default Choice: Do Nothing

When:

Campaign Status:  Not Run

Smart List Mode:  Batch (shouldn't it be Trigger??)

Smart List Status:  Affecting XXXX people

Smart List Reoccurrence: 

Repeats every weeks on Monday, Tuesday, Wednesday, Thursday and Friday

End date:Oct 13, 2017

Next three runs:

Oct 10, 2017 12:35 PM EDT

Oct 11, 2017 12:35 PM EDT

Oct 12, 2017 12:35 PM EDT

In any response can you please include screenshots of how I should adjust the format/setup?  What might I be doing wrong?  This would be so much easier if Marketo had a Throttling option for email sends.  I'm hoping this approach will work.

Any help is greatly appreciated.

Thanks.

Dave

[SanfordWhiteman]

I agree with Drew that that's one way to do it, but I have a particular dislike for arbitrary/approximate wait steps and believe waits should only be used for positive purposes, i.e. waiting 7 days to send a follow-up email. The other problem is as you winnow down the list you have to constantly recalibrate the percentages, it won't be 20% followed by 20% because your population has grown smaller so the percentage needs to be higher to send to the same count.

My way of approaching this would be a single Add to List step with choices for each random sample (i.e. 20, 20, 20, 20, Default where N = 5) that adds to N lists. Then those lists are immediately ready for sending independently.

[Anonymous]

Sanford,

Since I'm more of a visual learner would you be able to share a screen capture in your reply as to what that would look like?

Thanks

Dave

[SanfordWhiteman]

[Lynn_Ray_Pardo]

Sanford, I've said it before, I'll say it again! You are a freaking Ninja!

Thank you so much!!!

Big fan, not a stalker, ha, ha, ha, ha!!

Lynn

[SanfordWhiteman]

Ha, if I were a real ninja I wouldn't enjoy the compliments.

[Anonymous]

Thanks Sanford.

For List Name:  That is the same list over and over again in these 25% subsections right?

Also when running these campaigns you're focusing on just the volume of the send bucket not any sort of subject line testing etc?

[SanfordWhiteman]

For List Name: That is the same list over and over again in these 25% subsections right?

No, they're different lists (how would it work with the same list?).

Also when running these campaigns you're focusing on just the volume of the send bucket not any sort of subject line testing etc?

You can do whatever else you want. The point is that to divide a source list into N buckets you add to 100/N lists. Technically speaking here the randomization is irrelevant, it might as well be "Rounded Slice" instead of "Random Sample" but there's only one relevant feature.

[Anonymous]

Hi Sanford,

Thanks for the update.  I'm not sure the list add would work then in this scenario.  What I'm trying to accomplish is to mail smaller volumes of a specific list across several days.  I already have the static list in Marketo but now looking for a way to take a % of the list to mail.  For example, let's say the list is at 1,000 records, I want to mail 200 records on Monday, another 200 on Tuesday, another 200 on Wednesday, Thurs and Friday.  By the time I reach Friday that list is done and I can create a new campaign or start with a new list.

Make sense?

Thanks for your patience with this, much appreciated.

[SanfordWhiteman]

Thanks for the update. I'm not sure the list add would work then in this scenario.

Thought I explained exactly how it would work for this scenario.

There's no such feature as "Send to % of an existing list" without doing anything else.

There's only:

• Send to a % and then remove from flow and then send to some higher % of remainder and then remove and then send to some even higher % of remainder and then remove, etc.
• Split the list into exact N lists of % percent each and send to each list.

[Anonymous]

Starting to get it. 

So the flow under bullet # 1 would need to have the random sample, followed by a remove, followed by another random sample argument to be valid and work as we're detailing here, correct?

For bullet # 2:  I suppose what I'm missing here is how do you divide the lists automatically so you list each segment separately?

Thanks.

Dave

[SanfordWhiteman]

For bullet # 2: I suppose what I'm missing here is how do you divide the lists automatically so you list each segment separately?

Sorry, I don't understand the question.

What does this have to do with segments?

The process (as in my earlier automatically divides the original list into random cohorts, each a fixed % of the original list's size.

[Anonymous]

Hi,

So something like this?  So in part 1 I'm actually defining the email template/creative to use and then in # 2 I'm defining the random samples to be sent in 25% segments?  Under this layout how does Marketo know that there should be a break between 25% increments and exactly how long to wait before it sends the next 25%?  Apologies for all of these questions but nothing is detailed out on this subject in any clarity and want to get to the bottom of it.

Thanks.
Dave

[SanfordWhiteman]

Hmm, not at all. That would send to the whole original list and then split the list 4 ways and add each of the 4 cohorts back to a single list!

You want Add to List with 3 x Choice/Random Sample and 1x Default Choice for the last list, exactly as in my screenshot above. Then you follow this with one of two things:

1. Send Email with a 1 x Choice/Member of List  that matches one of the lists and 1x default Choice that does nothing, then a one-day Wait, then the next list, etc.

or

2. Separate scheduled batch campaigns that filter by Member of List, since you'll now have everybody pre-divided.

[Anonymous]

Yeah, this just isn't making sense to me.  I realize you're doing your best to explain it but it seems like a lot of manual effort for something that should be more automated.  I'm going to see their office hours tomorrow and see if there's any other resources.  My last ESP had a slider bar on how to adjust the volume releasing per hour/over time.  Marketo seems to complicate this issue significantly.

Appreciate your time.

Best.

D

[SanfordWhiteman]

It's never going to get more automated than Drew's and my takes on it, that's just the way it is w/the platform.

I wish I knew what you weren't getting above, it's ultimately just 2 steps -- add to N lists evenly and send to each list after a delay.

[Anonymous]

Here would be my latest stab at this before I start banging my head on my desk

Thoughts?

[SanfordWhiteman]

Alas, still not it.

Here, you're trying to do Drew's version, but you aren't placing any Was Sent Email conditions (Choices) on the Remove From Flow step. You must remove people who have already been sent the mail (and only those people!) like Drew said. Otherwise you're removing everybody after waiting 2 hours!

Also in Drew's version, you have to increase the sample % as you decrease the count of qualified leads, because the Random Sample is not guaranteed to produce the same exact slices of the original list if the count of people still in the flow changes from the original qualified count. You're changing the count (downward) as you do Remove From Flow.

Thus the second 25% is, in theory, 25% of the remaining people in the flow, and therefore less than 25% of the original qualified count (that isn't technically how it's done under the hood, but you can think of it that way).  In contrast, with my approach, you can be absolutely sure of a 25-25-25-25 distribution since you're performing the Random Sample within a single flow step when all the original leads are still in the flow.

[Anonymous]

Wow, lots of steps to finally equal the 100% of the sample.  Hopefully I have it now.

a. added the Was Sent Email Choices on the Remove From Flow Step

b. increased the sample % from 20, 30, 35 and 15 as the last leg, (100%)

c. Not sure if I need step 10 on this flow?

Let me know if this works in your eyes.  Again appreciate the insight.

[SanfordWhiteman]

Yes -- just catching your update now -- this is Drew's recommendation.

[Anonymous]

Hi Dave,

As with many things Marketo, there are a few different ways to accomplish this. Based on the setup you have described below, the path of least resistance is to add two flow steps to your campaign. The first flow step you're looking for is "Wait"... give it a few minutes. The second flow step you're looking for is "Remove from Flow" with a constraint of "Was Sent Email" or "Was Delivered Email".

Basically, what you need to do is to wait a few minutes for the email to be sent/delivered (wait time should be determined by the number of people you're sending to) and then removing anyone that received the email from the flow so that they can't be sent the email a second time.

The results of this are, the first send will pull a random sample of your starting list. The next send will pull a random sample from the 95% remaining, and so on. It essentially creates a waterfall type of effect where the list you're pulling the random sample from slowly decreases after each send.

Make sense? Let me know if you have any other questions/concerns. Thanks!

Drew