Recurring email -- incomplete stats showing in email performance report?

[Dory_Viscoglio]

We have a recurring (daily, M-F) email for our subscribers that has custom recommendations that are populated by a company called Certona and presented as images in the email. The email itself works great, but I'm having some issues with email performance reports.

So that we would be able to look at a week's reporting as a whole I have set this up as a Monday email, a Tuesday email, etc, each with their own respective smart campaign for sending. Then I set up an email performance report to report on the past week Sunday-Saturday since we have some Monday sends that start on Sunday due to time zones. Week one our reporting worked fine and showed all of the emails that were sent. Week two, however, the actual send numbers are similar (confirmed in the email send campaigns) but the email performance report is only displaying about 2% of the total recipients -- which I'm thinking are our new people who are receiving the report this week.

I've looked through all of the settings of the report, but can't find any reason why it wouldn't be displaying all sends for the timeframe specified. Has anyone seen this before? Or have any suggestions on how to get all of the send information to appear in this weekly report? Technically I can get someone to pull the daily information from our daily send smart campaigns, but that's the point of report subscriptions! 

Thanks in advance for the help!

[Justin_Cooperm2]

Dory,

The best way to do this is setup an Email Program (or regular program with SC and Email Asset) and then clone it for each send, using good naming conventions. Most customers do this and then Archive past programs that they aren't interseted in anymore.

There are some nuances for why certain reporting metrics won't always be accurate unless the send went out via a new campaign and email asset.

[Dory_Viscoglio]

Hey Justin, wow. To be honest, that really puts a kink in our workflow --- who has time to set up a daily send on a regular basis?! Isn't that the point of automation?

Is resolving this anywhere on the roadmap? Thanks

[Josh_Hill13]

Are you sure the email report is selecting the right emails in the right time frame?

If so, I recall some other threads where people had isues with the performance report not displaying data.

[Dory_Viscoglio]

Yup, Josh... right emails, right timeframe. Based on Justin's response, this is going to be a huge uphill battle, and a very manual email. Agh. 

[Justin_Cooperm2]

Yes, we are enhancing our entire analytics platform and fixing this is a top priority. This is why the email program only lets you send a blast once, to make sure analytics work as expected.

In the meantime, what you need to do is either use a different campaign and/or different email asset for every send. Using the email program is a great way to force you into doing this. The reason you need to do this is that if you send a mailing out with the exact same campaign and email asset, the opens won't be differentiated in the second send from the first send, so it will look like the open rate is way lower than you'd expect.

 

[Denise_Valles]

Are you using the same email more than once? If so, try checking the metrics under the email tab of the smart campaign and see if this shows more accurate data.

The email reporting will only report on the first send, open, and click of an email. To have accurate reporting on all sends, you must clone the email and send the clone each time. That way, there is a brand new email that will report on everything.

[Dory_Viscoglio]

Thanks for the input, everyone. Justin, aside from having bad experiences with email programs in the past, we have 7 different time zones that we send to and 5 emails - one for each day of the week... So I think I'm better off cloning the 5 emails weekly and then updating my smart campaigns. Creating 35 email programs and assets/week just for this one thing doesn't sound like the most efficient method here. 

I would love to have more insight on when advanced analytics for email reporting will be available -- maybe there will be some discussion of this at Summit?


 

[Dan_Stevens_]

Fixing this can't happen soon enough - thanks for sharing a perfect use-case, Dory, that illustrates how this is broke in Marketo.  We've been dealing with this for over a year and continue to build what would seem like unnecessary assets and smart campaigns just to accomodate the limitations here.  This happens a lot with event programs.  We'll have the program setup, run the first wave of email invites, and a couple days later, receive a request from the marketer that "we've received 5 more names from our sales guy - can you send them the invite?".  And a couple days later, the same request.  Rather than just add these names to the single "targeted list" in the program and re-run the "send invite" smart campaign, we now need to clone the smart campaign for every new request like this.  Before you know it, we have 5+ additional smart campaigns that do the same thing.

Like Dory, we never use Email Programs - we use Marketo programs for everything - even within Engagement Programs.  Speaking of Engagement Programs, we're starting to use these more often for multi-wave email campaigns.  Like the above scenario, when a marketer sends us another list of leads to send to, we simply add them as members of the EP and Marketo takes over from there.  No need to set up additional smart campaigns just so we can have realistic reporting.

[Dory_Viscoglio · Accepted solution]

Hey Dan, interesting use case for engagement programs! I'll have to think of that for next time 🙂 

Our final solution was to create a template program that we clone once a week (we're doing this a couple of weeks in advance so it isn't exactly a weekly process) that contains local assets for the emails, smart lists for reporting and also email performance reports, and they all reference to local smart campaigns. So the only thing that we need to do is update a unique daily identifier token, and then schedule the sends. 

It's still slightly more work than I would have wanted, but it's about as "automated" as we could get it given the constraints.