[onion2k]

You're talking about optimization. In your opinion emails with fancy layout and "on-brand" design will get a lower level of engagement than plain text with well-written copy. You may well be right (in a lot of cases.)

But you don't know that.

It's a guess based on your biases and assumptions. This is why it's critically important to measure things when you optimize them. Believing that you should engage with your customers in a way that you like because "they're just like me, so they'll like what I like" is an easy way to kill your messaging dead. Optimization of something like email is a matter of constantly measuring, refining, and measuring again.

[buro9]

What I do know from measuring my marketing and transactional emails is that the least possible formatting results in the lowest probability that I end up in the spam folder.

The hypothesis I had here (unverified) is that using a tool to produce email makes your email look like spam produced by that tool (by other people). My hypothesis more specifically is that spam detection software weighs similarity in the structure equally to similarity in content.

My HTML emails have since been modified to be absolutely minimal HTML as hints to the layout and nothing else. They are more like plain text emails that have been polished with just the lightest sprinkle of HTML and CSS but nothing more.

The result of keeping things simple is that my deliverability is over 99% and the open rate also phenomenally high.

[aarondia]

When I receive an html email, I default to thinking there is nothing of value in it, and I don't read it. I use email as a more formal text messaging, or often group text messaging, system. Marketing newsletters, etc are just a clutter. But I do get a lot of them, so maybe they work for others.

[andrei_says_]

Are you aware of any studies verifying this hypothesis? Would be very useful in my team’s roadmap.

[citrin_ru]

Spammers sometimes (often enough to bother) are trying to fool spam filters by adding invisible or hard to see text copied from ham messages (e. g. it allows to bypass Bayes). To counter this spamfilters are trying detect and penalize invisible text in HTML emails e. g. <span style="font-size: 0px;">....</span>, display:none, text with the same color as background, e. t. c.

Running something like headless browser (Chrome) to parse HTML emails would require too much resources so spam filters use either simple ad-hock HTML parsers or just regexps. If you use complex HTML markup, e. g. lots of nested div/span tags, different CSS styles on different levels and global <style> block on top of all, they can be confused and will detect invisible text in a message where all text is visible (when CSS/HTML is rendered in Google Chrome; who knows how it would look in different mail clients).

If you use clean and simple HTML, there will be no leftovers like <span style="font-size: 0px;"> and you will not confuse spam filters.

[andrei_says_]

Ironically <span style="font-size: 0px;"> is a must for spacing tables if you want outlook compatibility.

[citrin_ru]

Ironically this technique was used in phishing for MS Office 365 credentials (users of which often use MS Outlook): https://www.avanan.com/blog/zerofont-phishing-attack

Rendering in Outlooks changes from time to time - it is worth to check if this workaround is still required.

[contravariant]

Normally you'd think the burden of proof would be on the side requesting considerable resources and infrastructure to send something that doesn't render properly under the most basic privacy settings and cannot be displayed text-only (like in notifications).

[passivate]

Normally I would agree with this sentiment, but what kind of proof are you looking for here? We're not discussing logical absolutes or claims of absolute fact! We're discussing peoples preferences - what they like and/or expect an engaging email to look like. This area is highly subjective and contextual (culture, region, age, demo, etc). Neither your nor OPs experiences invalidate each other. Both of you can be correct at the same time.

Personally, I've seen engagement metrics increase when marketing added more "fluff" to their emails. Its no different than those obnoxious thumbnails people use on YT - it works - maybe accidentally, but it does, for now. That may change at any time in the future, just like what people like to see in emails/magazines/brochures is bound to change in the future.

[contravariant]

Sure if you just prefer if your emails look a certain way then there's no point in discussing proof, just do the thing you like.

However if the discussion is whether you should just send plaintext emails or use a complicated and/or expensive tool to send out fancy emails then the more time/money consuming side has more to prove.

Given that the first few replies I got were essentially "Well I assume somebody checked this", I'm not sure the assumption that fancier emails are necessarily better has been questioned enough.

I'm also reminded of several articles where online marketing itself turned out to be nowhere near as effective as people think. [1,2]

[1]: https://thecorrespondent.com/100/the-new-dot-com-bubble-is-h... [2]: https://news.ycombinator.com/item?id=21465873

[passivate]

>However if the discussion is whether you should just send plaintext emails or use a complicated and/or expensive tool to send out fancy emails then the more time/money consuming side has more to prove.

I am curious, how many plaintext emails do you receive as a percentage? I don't receive any so maybe my opinion is colored by that fact.

>Given that the first few replies I got were essentially "Well I assume somebody checked this", I'm not sure the assumption that fancier emails are necessarily better has been questioned enough.

I'm 100% with you that conventional wisdom should attract scrutiny from time to time, and I don't mean to discourage that. My only point was that there is nothing to "prove" here. Marketing styles/trends/methods are not linked so neatly and linearly that we can test individual elements by scientific method. There are far too many variables. I think it is natural that each company wants to be able to choose the font, add graphics to make the email attractive, add tracking pixels to measure performance, etc. Its sort of become the default. So maybe the question really is what do you stand to gain by removing all that?

>I'm also reminded of several articles where online marketing itself turned out to be nowhere near as effective as people think. [1,2]

Heh, maybe marketing was never as effective as people claimed, its that now we have the tools & metrics to detect just how bad it is. :)

[anamexis]

I would assume that side has, in fact, established that proof.

[TrueDuality]

I'm also going to make a pretty big assumption, which is that a lot of people using mailing tools that give them stats such as MailChimp do not think about how privacy settings affect their metrics. If I have images disabled by default I will almost certainly never show up in an "opened" metric.

Images off by default is becoming more common.

[contravariant]

It's a common assumption, but also a dangerous assumption when the people who you assume have established the proof are the same people who stand to profit a from it being proven.

[anamexis]

I'm not sure I follow - if marketers proved that plain text emails are more profitable, they would use plain text emails.

[hedora]

They certainly haven’t proven its efficacy to the people whose resources are being consumed.

[3np]

I hope you’re not suggesting inserting and relying on tracking pixels and other unreliable junk into outgoing emails in order to measure.

Besides, guess what, many times, “engagement” is precisely what I don’t want to be coerced or manipulated into doing.

Cursing while clicking around your funneling website to find a place to do a GDPR data delete request and eventually giving up while filing your sender as spam also counts as engagement. Whereas plaintext email is way less likely to trigger that response.

[linsomniac]

As someone who spent a very long time in a text-only mail client, on purpose, and then switched to Thunderbird and set it to the "show the text version if at all possible" mode, I'll say that I've become ambivalent about it.

Now actually appreciate being able to scan emails based on the 10,000 ft view. Sometimes, images can convey a lot in a glance, the overall graphical layout can convey a lot too.