Email Clicks on Delivery But No Matching Web Page Visits Logged

Great insight, Sandy.  For the past couple of months, we've been using this approach to eliminate false positives - where we include a "visited web page" as a choice within the "change program status" flow step (after a 15 minute wait step, since it can take time for Marketo to log that activity after the click):

But like you said, if anyone has cookies disabled, do-not-track enabled, or opted out of specific tracking cookies (per upcoming GDPR legislation), this won't work.

So it looks like we're back to square one as the approach mentioned in this thread doesn't even appear to be a reliable solution to weed out link scanners.

Hi Lynn

We are having the same issue. Could you please share more details of the filter used? Is it working as expected?

Thanks

Axel

And thanks!

May I ask if you have a document showing the firing order for activity logging, i.e., order of operations? (I just searched the MKTO docs page to no avail.)

No doc, just testing.

Thank you so much, Sanford!

Your info wrt the order of CE activity firing before VWP is most helpful.  (Guess, I'll have to go back and correct quite a few SLists, created since speaking with MKTO Support ha, ha)

May I ask if you have a document showing the firing order for activity logging, i.e., order of operations?  (I just searched the MKTO docs page to no avail.) 

I did find the excellent community post for controlling the order of operations.  But I am helping train new MKTO users next week and hoping to provide a cheat sheet.

In the meantime, thank you for this and all of your MKTO community posts, Sanford! You are a gift!

Lynn

@Sanford Whiteman​ we have seen some patterns that are helpful in terms of user agent filtering. But to your point, anything that can successfully distinguish the scanner publicly will be actively developed against by the mail scanner company and likely won't last long.

@Roxann McGlumphy unfortunately we have not found this statement to be accurate:

"While this may not catch all test clicks from the security software, it catches most of them."

Our testing showed most of the mail scanners were sophisticated enough to not blindly click every link in every email. Only a small subset would click a hidden link and be identified that way.

Quite so. A mail scanner worth its salt will not be distinguishable from a human click behind the same public IP.

Can you give us some clarity on how Marketo determines the 'source' of these clicks? It seems likely that a legitimate click could be ignored if it happens after a bot-click for the same email.

You could expect these real clicks to happen just as often in your 'bot-click group' of recipients as among your 'non-bot-click group' of recipients.

It's important that a solution doesn't throw out some valid data with all the bad data.

One of your brilliant co-workers, Chris, suggested that until we are sure that the MKTO feature is implemented for us, that we include a filter for visited web page.

That won't work if the click wouldn't result in a Visit Web Page, however (i.e. a direct download link has no VWP associated with it, even though you shouldn't send such links anyway, many people do).

It'll also be a false positive on anybody who's using an privacy plug-in (which is kinda exactly where you do want to trust the CIicked Email activity, since it's all you've got).

More important, the VWP fires after the CE. So you can't filter the CE on VWP.  You can try to filter the VWP on CE but all our attempts to do this have failed because the look-back in the trigger doesn't have the necessary granularity.

I don't have confirmation on this yet, but so far every account that I personally have checked has had it implemented, and the examples of bot-clicks that I have seen were emails sent before the September 29 release.