Statistics for campaigns are not accurate

Ikhan85

New Member
Hey Folks,
I have mailwizz connected to a sendgrid server to send outgoing mail, but it doesn't appear as though the stats displayed in mailwizz about the campaign are accurate. Send grid provides real time numbers for clicks, opens, etc, but they don't line up with what mailwizz is displaying.

Have I failed to do something in the backend that would allow for mailwizz to obtain more accurate campaign stats?
 
MailWizz stats are 100% accurate.
We count all opens and clicks as they hit your server. SendGrid's counting system might be different, we don't know, but as far as we're concerned, we count everything.
 
Hi @twisted1919, may I suggest reviewing the open/click tracking in line with other providers? I.e. Sendinblue are only showing "trackable" opens and clicks and seem to have blocked proxy opens/clicks from Google and Apple. Sendinblue use this message in their stats:

Trackable opens
We know for sure how many people opened your email in the trackable population (people who did not enable a Privacy Protection option on their email client).


Yes, this will make the open and click through rate appear lower. But the current situation is not sustainable. I have just sent a message to a tiny segment (23 people) and had 203 click-throughs on a single link within 2 minutes, plus a near 100% open rate immediately. It lacks any valuable insights. So I'd rather exclude a bunch of proxys incl. the people who have opted out of Email tracking via Apple, than to have those ludicrous stats which my clients keep questioning me about.

If you have an alternative better solution waiting in the wings, fantastic!
 
So I'd rather exclude a bunch of proxys incl
Sure, but then, if for certain subscribers, the open only happens via proxy, you'll lose any tracking from them.
And since all big providers do this, GMAIL/Yahoo/Etc when you use their web interface, then that will affect things considerably, so it is not an option.
You can block IPs and IP Classes in the app now, there's a settings section that allows you to do this, but all I can tell is that is a bad idea.

We know for sure how many people opened your email in the trackable population
This is literally non-sense, they mean people that generally use local email clients(except apple mail + privacy on), because for sure they have no control over what happens when people use the web client for gmail for example.

If you have an alternative better solution waiting in the wings, fantastic!
I have none, because there isn't anything that could fix this, as long as the requests to open the email and to click links are proxied, the email application will only view the proxy info, that's just how this works, it's not like you can look at the request and extract the info you need, because that info is not there to begin with.

Maybe a solution would be to delay recording of clicks/opens a few minutes after the email has been sent, this way you'd be skipping all the spam filters/checks/etc. But then if your email would be queued on the smtp server for hours before delivery, this becomes useless.
We could not track first opens/clicks, but how many of them should we skip? one, two, 10? Wouldn't we lose clicks/opens like that?
So yeah, this is not simple and I am open to discussion here.
 
Sure, but then, if for certain subscribers, the open only happens via proxy, you'll lose any tracking from them.
But what's the point of tracking proxies? You're not tracking anybody anyway, so you might as well skip them. It's not reporting on any user behaviour.

And since all big providers do this, GMAIL/Yahoo/Etc when you use their web interface, then that will affect things considerably, so it is not an option.
You can block IPs and IP Classes in the app now, there's a settings section that allows you to do this, but all I can tell is that is a bad idea.
All of them? I'm not aware of that. I appreciate your advice here though and will try the manually blocking for a period of time. I see how the other ESP platforms I occasionally deal with don't have such strangely inflated open/click-through stats.

This is literally non-sense, they mean people that generally use local email clients(except apple mail + privacy on), because for sure they have no control over what happens when people use the web client for gmail for example.
It's just communicating that they do not include proxy IPs.

I have none, because there isn't anything that could fix this, as long as the requests to open the email and to click links are proxied, the email application will only view the proxy info, that's just how this works, it's not like you can look at the request and extract the info you need, because that info is not there to begin with.
I know. Sorry, I really didn't mean to pick a fight. I sense you're a bit annoyed by questions on this subject. We're all trying to make MailWizz even better. So please don't take any of these comments as complaints.

Maybe a solution would be to delay recording of clicks/opens a few minutes after the email has been sent, this way you'd be skipping all the spam filters/checks/etc. But then if your email would be queued on the smtp server for hours before delivery, this becomes useless.
Nope. Not a good idea.

We could not track first opens/clicks, but how many of them should we skip? one, two, 10? Wouldn't we lose clicks/opens like that?
No. Because Apple's and Google's methodology is different.

So yeah, this is not simple and I am open to discussion here.
That's good to know. I know that the Email Marketing and Email Technology community are already discussing this vividly. It's a matter of joining that discussion for all of us here and figuring out the best way forward based on what other players are doing.
 
It's not reporting on any user behaviour.
I dunno, I think it does though because the user itself opens the email, the exception is the open is done via the proxy. So you lose a valid open in that case. It is the user that did the open, but behind the proxy, so how do you handle that? because it looks like a proxy open, but it's the user at the other end. So do you keep or discard the open?

All of them? I'm not aware of that.
Yahoo and Gmail for sure, and, while I don't have the numbers, I think it's safe to say that's like almost the whole internet.

It's just communicating that they do not include proxy IPs.
True, which we can do just fine, but then again, you risk losing real opens done via the proxy, which for they might not matter since they do transactional emails, but it does matter for you doing marketing emails.

Nope. Not a good idea.
No. Because Apple's and Google's methodology is different.
Yeah, I had a hunch that's the case, but throwing ideas out there at least help us rule them out.

I know. Sorry, I really didn't mean to pick a fight. I sense you're a bit annoyed by questions on this subject. We're all trying to make MailWizz even better. So please don't take any of these comments as complaints.
I am sorry if I came out like that, that was not the intention, I went through these answers really quick on a late afternoon, trying to wrap the day, knowing these are subjects we don't have a fix for yet.
I am also keeping tabs on a few different discussions related to this watching how they evolve and looking to see if there's anything that will come out out of them.

For anyone not knowing the problem, and doubting this affects all email marketing platforms, this video show that even the largest email marketing platforms suffers the same problem:
 
So we did a few changes, in the campaign setup step, you will now see a new section, dedicated to tracking:
Screenshot 2022-07-05 at 11.19.57.png

The first two options, you know already, they just got some additional help text:
Screenshot 2022-07-05 at 11.32.47.png

The third option, "Open tracking from url tracking ", which is new, is explained like:
Screenshot 2022-07-05 at 11.20.59.png
Which means that you can disable open tracking completely, so that no tracking pixel is added to your emails BUT you can still track opens when a link in your email is clicked. Of course, you can still enable open tracking and it will work just like it does now.

The other two additions, "Open tracking - Exclude crawlers *" and "Url tracking - Exclude crawlers *" are explained like:
Screenshot 2022-07-05 at 11.25.36.png
Which means we will try to detect if the open and/or click is made by the subscriber itself or by a proxy / bot / crawler / spam filter, etc.
And if we detect the action is not made by the subscriber, then we will not track the open/click at all.

You might ask why do we have the "Open tracking from url tracking" option if we can exclude bots/crawlers/proxies using the "Open tracking - Exclude crawlers *" and "Url tracking - Exclude crawlers *" options.
The answer is simple, sometimes, when open tracking is enabled, the tracking pixel for opens might be blocked, so in this case, we will track the open when the first click in the campaign is done by a real subscriber.

I think these are the right tools to allow you to decide how to do the tracking per campaign basis, based on your own logic.
 
So we did a few changes, in the campaign setup step, you will now see a new section, dedicated to tracking:
Will this be available with the next update? I'm aware that it obviously wouldn't magically appear. Or does it need to be enabled in the backend? I'm on version 2.1.14.
 
yeah, which was scheduled for today, but we found a bug in the tracking links and we're now fixing and testing it, so tomorrow most likely.
I'm having an issue with my tracking of opens on some campaigns (but not all) could this bug be causing it. I just updated to 2.1.16, should that fix it?
 
I'm having an issue with my tracking of opens on some campaigns (but not all) could this bug be causing it. I just updated to 2.1.16, should that fix it?
Check again with the latest version because has been fixed a bug in tracking urls that would prevent transforming successive links on same line.
 
Check again with the latest version because has been fixed a bug in tracking urls that would prevent transforming successive links on same line.
I tried with the latest version (although they were a continuation of recurring campaign before upgrade) and not much improved).
 
When you say they are accurate, what do you mean? Can you please be more specific?

Cosmin
 
@twisted1919 Hi, recently updated to version 2.3.9. Really happy to see the option to exclude crawlers/bots etc. Can you give any information on how a bot/crawler is identified? I know this issue is industry wide. Thanks, J.
 
Thanks @twisted1919 I know there's lots of schools of thought on this. And i've read lots of other threads here on Mailwizz, and within the larger email community. Does that library exclude Amazon, Google and Apple IP addresses used for privacy opt outs? I sent a test campaign within my company that would have been opened on 4 iOS devices, and also by my test mail-tester.com address. I was surprised to see these emails shown in the Opens list:
1704476028615.png

Particularly as the mail-tester.com opened all of the links within 1 minute of receiving.

Is there an option to discount behaviour that opens/clicks all links either immediately or multiple times? Moving forward I'll be adding a 'hidden' url in all my templates, and discounting any address that clicks that link from the reports i give to my clients. But I'd be interested to know if i can add a filter somewhere to do that within Mailwizz.
 
Back
Top