some of the modern mail clients flag mails that use a link inside the link element but href to another site as scam. This is an issue with click tracking. Of course this is not an issue of Discourse but I don’t see any down sight about changing the two links in the mail template to avoid getting flagged as scam with link tracking.
This would mean, instead of
To respond, visit http://example.com/t/our-next-big-goal/119/2 in your browser.
we use
To respond, visit |Our next big Goal - Example Community| in your browser.
I think what @Slind is talking about is changing it so the URL does not appear in full in the HTML version of the email, just the linked title. Not doing away with the link altogether.
It does seem puzzling and suspicious to me to see URLs in emails from the forum that look different when you mouseover the URL.
So yes I would turn off link tracking which stops that issue for right now. That does come with a cost however which is that we can’t track in mandrill who’s opened links. We all have to weigh the pros and cons and decide what’s more important… participation or tracking of participation in mandrill. Another trust building effort might be to tell your forum participants to add your forum from address to their contacts to prevent this from happening.
The alternative @slind is suggesting is actually a good one methinks which is to show the title of the topic as a clickable link rather than the URL which is converted by mandrill.
If you don’t want to disable click tracking for all mail in your Mandrill account you can also add the following option to your discourse setting for email custom headers X-MC-Track:opens
The detail on the header options can be found here:
AFAIK, this won’t fix the “possible scam” flag in Thunderbird until they get their software patched (which, keep in mind, is less than 1% of email client market share).
I just realized that we have two threads going on this topic… one or the other should be closed methinks.
Honestly, I personally am also affected by this issue - don’t know about anyone else and haven’t asked - because I tend to mouse over links I get by email before clicking on them and look to see where it goes. I happen to know mandrillapp.com is trusted but for someone else new to my forum who doesn’t know mandrillapp but has my habits would be suspicious to be asked to click a link that says it is one thing but leads me somewhere else.
It’s not a big enough issue for me to want to make any big changes to discourse or to turn off click tracking on mandrill for my sites, but it’s definitely something I am aware of and think about. So having a way to improve this in the emails discourse sends out would be great.
If possible I would like to focus on the change of the one phrase (a text content instead of link content).
If we leave the tracking/scam thing by side, would you prefer it to be the thread title or the link? While the link is optimized for human readability I still prefer pure text.
Edit:
I see a big possible issue with these stats: http://emailclientmarketshare.com/
The open rate can only be tracked when pictures get loaded or a link clicked. Some of the clients do show images always others ask the user to confirm it. Further more I would evaluate the open rate of advertisements higher on mobile devices than on workstations, which might also have a huge impact that should be considered. One more thing to be added is the user group of the discourse forum users, I don’t think that the average stats shown do only very slightly. I do see discourse used a lot for open source projects, administration topics, indie games… I do think that open source software (e.g. thunderbird) is used more frequently in these areas …I’m drifting far away from the topic, sorry. Just wanted to add my concern on these stats.
We all have to weigh the pros and cons and decide what’s more important… participation or tracking of participation in mandrill. Another trust building effort might be to tell your forum participants to add your forum from address to their contacts to prevent this from happening.
