2019 Adobe Summit Follow-Up: Skinning Schrodinger's Cat Q&A

56119847_10100142598678572_4581330048572719104_o.jpgAnother Summit is behind us and with it came a bunch of FAQs. As with any conference, information overload is a very real thing and 45-minute sessions can’t cover everything we want to say or answer every question asked (hence why we preface each year’s session with the words “rapid-fire”).

Julz James and I were lucky enough to get a bunch of really great questions after our session, “Skinning Schrodinger’s Cat II: Fearless MOPs in the Face of the Unknown” and we promised you we’d answer them all in a follow-up blog post. Without further ado, here's what you wanted to know:

What are the two "not mailable" segments you had?

(Jenn) This question is in relation to creating mailability segments so when you filter by mailable records, you’re left with a clean list of who you can actually target when sending an email.

The order of segments that I use is usually something like:

  1. Employees
  2. Not Mailable - System (these are your records who can’t be reached for technical reasons like invalid emails)
  3. Not Mailable - Policy (these are your records who can’t be reached for compliance reasons, namely things like GDPR and CASL)
  4. Mailable (this is what you’ll pull into your email blasts to filter only marketable records)
  5. Default (there should be 0 people in this segment because everyone should have been caught by the above filters)

(Julz) The mailable segments I use are very similar to Jenn’s but I would set it up slightly differently and split Leads and Contacts up into their own segments.  The reason behind this is because in some instances there are requests to send campaigns just to leads or contacts, and it saves having another filter in the send campaign.

When setting up segmentations always remember that you can decide on the order that you want your records to be filtered out, so for mailable segments you’ll want to filter out the bad records first.

Do you have any tips for dealing with spam bot clicks AKA Voldemort McSauron

(Jenn) It depends on your instance so you should really go through your records’ activity history to see if there’s anything you can filter out when setting up your tracking campaigns. For example, I discovered that one of my clients was getting a lot of spammy clicks on their emails but none of those emails were ever recorded as opened. Although you can’t suppress that information from Marketo’s default reporting, you can set up your campaigns to prevent them from getting added as “clicked” to your programs.

(Julz) One of the ways I try to avoid spammy clicks in programs is by adding filters to the smart campaign that is looking for the clicks.  So I’ll set up a Smart Campaign that has a trigger of ‘Visited Web Page’ with the filter of ‘Clicked link in email’ that has the matching link in it.  I also put a time filter on the trigger sometimes as sometimes the timestamp between the page being visited and the link being clicked sometimes has a delay.  One of the caveats to this is that when people have disabled Cookie tracking we can’t track the page visits. But it is at this point where you have to weigh up the options and look at what Jenn suggested and see how many spammy clicks you are getting - run a couple of smart lists on your biggest email sends and see how bad the problem is.

Where can I find the resource for bounce management program covered in the presentation?

(Jenn) You can find the Community blog post detailing where more information about bounces and how to track them here (https://nation.marketo.com/community/champion/blog/2015/07/14/making-sense-of-marketo-email-bounce-c...) and download the presentation slides going into detail about how to set up the program and campaigns here (https://resources.digitalpi.com/presentations-2/jenn-dimaria-skinning-schrodinger-s-cat-ii-fearless-...). I also filmed a PiPointer giving an overview of how it works here (https://resources.digitalpi.com/blog/pipointer-why-managing-bounced-emails-is-like-herding-cats).

Why do you keep email addresses that you can't email in your segmentation / Marketo instance?

(Jenn) Mostly for compliance purposes, although a case can be made that you can delete these records out of Marketo if you have other ways to keep track of if/when they opted out or couldn’t be contacted. For technical purposes, it can help to keep a binary sync between systems (ie: have all the records in Marketo that are in your CRM) to prevent duplicates.

For example, if you delete a record out of Marketo only because they unsubscribed but that person comes back to your website, fills out a form, and downloads content, they’ll re-sync to your CRM and create a duplicate because Marketo can’t tell that record already exists.

If you have a duplicate management system in place within your CRM, that may not be a problem for you but it’s definitely worth noting because I see it happen all the time in my clients’ instances and it can become a huge headache to manage.

(Julz) I’m the opposite to Jenn.  I tried to purge email addresses as much as possible as long as the CRM system is keeping track of the invalid emails, unsubscribes, no longer with account, do not contact, etc etc. This is down the Database size being one of the pricing factors in Marketo. But I’m also in total agreement with Jenn that this is only possible due to the CRM keeping all of the records and also a good duplication system to prevent too many duplicates happening.  I always think there will be 2 differing opinions of what to remove and what not to remove.

How much does using segmentation affect processing? I've heard each segmentation transaction is a single batch campaign run per lead. Does that slow down larger instances?

(Jenn) I’ve heard this is an issue, as well, but have never experienced it. By optimizing your segments (ie: not including filters like “contains”), you can ensure they’re functioning quickly and efficiently. Segmentations themselves are also much faster to use in smart campaigns (partially because a lead can only be in one segment) and can often take preference over smart lists if you’re targeting a broader group.

That said, the first time you run a segmentation could take quite some time - I’ve seen my mailability segments take up to 24 hours to run on larger databases.

Here’s a great resource in community about common segmentation issues and how to correct them (https://nation.marketo.com/community/product_and_support/blog/2018/12/18/segmentation-health-check-u...).

Have you ever had issues with the history smart campaign not always running to record the field updates?

(Julz) I’ve never seen an issue with the UTM history fields not updating.  This is probably due to the fact that I’ve got all UTM updates happening in one campaign, so that the smart campaign is only for one change per field.  Its set up in a way that the trigger is just looking for just the UTM ‘most recent’ field to change. This change then triggers the ‘original’ and ‘history’ field updates.  I also put a filter on the smart campaign settings so that people can only run through the campaign once an hour, so that you don’t get someone who is clicking all over your website and landing page with UTM parameters.  That might be one of the causes of the history field not being updated.

One more thing to note - always make sure that your forms have the correct UTM parameter fields, and that your URL’s are built correctly too.

Have you noticed a processing slowdown with the bounce campaigns due to the volume?

(Jenn) I haven’t but it’s important to always monitor these type of campaigns in relation to your other initiatives to diagnose any processing delays or race conditions.

That’s why it’s also important to keep tabs on the sources of your invalid emails. For example, if you determine that a specific source is sending over loads of bad emails, you can consider dropping that source and fewer bad emails will flow through your bounce management program.

For your first few bulky sends after you set up a bounce management campaign, you can expect to see a lot of traffic and data value changes. But if you’re acting on the results and not batch-and-blasting willy nilly to your non-engaged database, that should slow down.

Note: this shouldn’t be a replacement for a deliverability monitor but it can help give you an idea of how your marketable database is performing and allow you to quarantine potentially bad records for further investigation.

Can you track the link click inside the velocity content token?

(Julz) Yes, you can track link clicks inside Velocity tokens.  The Velocity Scripting was used for the Real Time Content Delivery program that I spoke about in our session has a different token per piece of content, and different variables based on different fields.  It’s a bit more work, but after running reports after each event that has a content delivery email set up for it it's worked without any issues so far.

If I wanted to use 1 Velocity Script Token for all of the content then it would be a different story.  For that you can check out Sandy Whiteman’s blog here for a more in-depth break down of how to get work arounds for tracking links in Velocity: https://blog.teknkl.com/multiple-marketo-tracked-links-in-velocity/https://blog.teknkl.com/multiple-marketo-tracked-links-in-velocity/

Where can we find out more about velocity scripting and where those live?

(Julz) I’m still learning Velocity but Sandy will always be my go to for this: https://blog.teknkl.com/tag/velocity. There is also any work on the developer site for Marketo: https://developers.marketo.com/email-scripting/ and a full site dedicated to Velocity: http://velocity.apache.org/engine/devel/user-guide.htmlhttp://velocity.apache.org/engine/devel/user-guide.html

I’m excited to spend more time playing with velocity and also javascript to see what fun things I can set up in Marketo.


Have more questions for us? Let us know in the comments and we’ll answer as soon as we can.

If you weren’t able to get into our session or if you did and want to review any of the info, you can play the session recording here: Adobe Summit 2019—The Digital Experience Conference | March 24–28, 2019 .

PS: Julz and I will be presenting "Schrodinger's Remix" at the Marketing Nation Roadshow: Boston on June 18th where we'll focus on our Summit highlights and your Q&A. If you want to see us in person and are in the area, register here: http://fieldevents.marketo.com/marketingnationboston