Stop tracking mailto: links (if you want them to work)

Sanford Whiteman -

There are great reasons to use mailto: links in marketing emails — and just as many reasons to make them actually work when they're clicked.

Unfortunately, most Marketo users are doing it wrong. You're adding handy-looking CTAs that don't work for a huge numbers of leads (maybe even the majority of your leads, depending on industry). "It works for me!" you say. Well, this is one area where testing on your own computer just won't cut it, where Incognito/Private Browsing isn't enough. You need to truly simulate real-world mail clients to see why and how it breaks.

What's broken

It's pretty easy to explain why mailto: links break when you track 'em.

When you track a link (tracking is on by default when you click Insert Link) Marketo will redirect the link via your tracking domain. If your tracking domain is click.example.com, all links will be "wrapped" with that domain: http://click.example.com/z1c574fe14, http://click.example.com/e3ca486b82, etc. When your leads click such a link (the much-desired Clicked Link in Email activity), their computer/device does the same thing it does for any http(s): link: it opens a new window or tab in the user's default browser. It doesn't have any way to know that click.example.com is going to try to send them to a mailto: link next. All it knows is that http://click.example.com is an http: link.

That, right there, is the problem. The moment you exit the user's mail app (be it a webmail app, a standalone mail app like Outlook, or a mobile app) to open a browser window, you leave that app's control. It's then up to the browser to decide what to do with subsequent pages, docs, or redirects. 99%+ of the time, the next location is another http(s): page or download, so the browser can handle it. But when the next link is a mailto: link (the same applies to other non-http: links, too, like itunes:) the browser throws up its hands and asks the OS for help.

In this case, it asks the OS, "What's your default app for new email messages?"

Now, if the lead's go-to mail app is Outlook (not Outlook Web Access), chances are they (or their IT admin) have set up their OS to use that app as the default for mailto: links. That's the best-case scenario for mailto:, since links do open a new message in Outlook (even though they take a circuitous path from Outlook, to the browser, and back to Outlook, they still work).

Problem is, the best case may also be your case, and it's the critical reason why testing mailto: links on your company workstation is not sufficient.

There are other scenarios that have a far worse outcome:

  • A lead using Gmail, Yahoo, Outlook.com or Outlook Web Access in their browser, with no special OS-level setting for mailto: (this is the typical setup)
  • A lead using webmail on a machine that has an outdated mailto: setting
  • A lead using niche or legacy webmail where no mailto: handling is even possible

As bad as it gets

As a marketer, you must assume that your leads will not have made any special changes to accommodate your emails. I recorded some videos so you can see what a lot (quite possibly a majority) of your leads are seeing when they click mailto: links. To record these, I used a completely fresh machine, with a default mailto: setup, so as not to obfuscate the problem.

Click to experience a mailto: link as:

The videos should be disturbing. The first two leads get a seemingly dead link (their browser opens to an empty click.example.com page and stays there). The third gets prompted to set up what their OS thinks is their default mail app — Apple Mail — for the first time, which is not just a hugely confusing UX but impossible in most cases (the lead would need to know things like their SMTP server hostname and credentials, which may not exist in their environment).

And the recorded experiences above aren't the only bad outcomes:

  • Some Windows users will get a "Set up your mail app for the first time" prompt like the Mac example.
  • Some users will get a New Message window, but will get an error when they click Send.
  • Some users (this is the case you should dread the most) will get a New Message window, and they'll successfully click Send... but unbeknownst to them, the message will never be sent. This is because they had a default mail app that was seemingly usable, but had an obsolete configuration. So the mail just stayed in the Outbox. Since they don't really use that app day-to-day, they'll never know there was a problem. (If you hear a lead say, "I emailed you" and you never got it, this may be why.)

The fix

When you insert a mailto: link, always disable link tracking by unchecking Track Link:

This will in turn disable Include mkt_tok, so the dialog will look like this:

When you disable tracking, the mailto: link will remain just that, a mailto: link. It will not be wrapped in an http: link. Thus all your leads will get their best shot at a usable mailto:, and their experiences will be more like these:

But doesn't this mean —

Yes, it means you cannot both track mailto: links and have mailto: links function widely + reliably. Either:

  1. You turn on tracking, but only those leads who are lucky enough to see something other than a blank page or "first time setup" dialog can send an email, and all others are alienated and confused (and you possibly even lose inbound mail); or
  2. You turn off tracking, and just about all your leads can click the mailto: and successfully send you a message

I hope the choice is clear.

A broken mailto: will still log an activity

An important detail: in the videos showing the broken behavior, the lead does not get a New Message window, so obviously the mailto: did not work. Yet simply by opening click.example.com, the lead will generate a click in their Activity Log either way. This is because click.example.com first logs the click, then tries to redirect to the mailto:. It never knows that the mailto: didn't work.

You can use this detail as a way to audit how many people attempted to send you an email: every registered click on the mailto: should result in an email delivered to the mailto: mailbox. Of course, if the mailto: recipient is in Sales, this is practically impossible. If it's a shared Marketing mailbox, though, you can run the comparison. Needless to say, any disparity is bad. And even if you don't have a disparity today (after all, the only leads who notice the breakage are people who chose to click) the examples above demonstrate that you eventually will.