Confusing what works for YOU with what gets results
I originally started this post about a week and a half ago. The plan was to write about an interview I read with a web standards advocate - she was asked about the Email Standards Project and replied that SHE preferred text emails but those damn marketers are just insistent on using HTML emails so they were clearly here to stay. As if the evil marketers created HTML email! This was going to lead into a rant about how it was a marketer’s job to adapt to the tools that would achieve the goals they have been charged with.
Then yesterday on Twitter I made the comment that graphic designers typically make terrible email marketers, which seemed to play right into this topic. Some would (and did!) say the comment was disrespectful to graphic designers. I disagree - what graphic designers do is amazing. I sure as hell can’t do it. But making a pretty picture is not the same thing as building an effective template. Wouldn’t you know, but some people actually agreed with me!
In this case, I don’t even need to make the argument - @mStonerblog and @doctorious did it for me!
- A design can be the prettiest thing in the world, but if your users have their images turned off it’s all for not. Worse yet, image heavy emails tend to have their main calls to action in the form of an image - a MAJOR no-no. You’ve cut your own feet right out from under you. It’s OK to use images in email, but only to compliment the content - not to be the focal point of it.
- “Ugly” mail usually has higher response - it’s not necessarily that it’s ugly but it’s a whole lot easier to make something look slick when its a graphic rather than utilizing 1990s-style HTML, with a little bit of CSS thrown in. A simple template that focuses on quality, timely content and overt calls-to-action is statistically what works. It’s about finding a happy medium between image-heavy templates and text-only email.
I love Urban Outfitter’s emails. They’re slick, different, and visually striking. But will that style work for my audience? No. So I put it aside in favor of templates that I consider to be boring at best, but that get results. A good marketer does that. It’s not about you. It’s about achieving your goal. Put your personal tastes, and dare I say ego, aside and acknowledge that what works for you may not work best for achieving your goals.
~~~~~~~~~~~~~
Like what you read here? Subscribe to this blog through RSS or sign up to receive email updates when new content is posted.
Hire Karlyn: Karlyn is the President & Principal Consultant of DoJo Web Strategy, a consultancy dedicated to helping colleges and universities to utilize the web in their marketing efforts. Find out how she can help your institution by requesting a proposal today »
February 27th, 2009
I think the issue is that a lot of managers (and some designers) don’t fully comprehend the coding limitations of HTML e-mail. They think that if it will look good in their program (usually Outlook with images turned on) it will look the same for everyone else. What they don’t realize is that ‘images off’ is the default for many programs and that if the call to action is in a photo many readers won’t see it at all.
Back in December I rec’d one too many such e-mails. A local publisher had sent out a ‘breaking news item’ but the news was in the graphic. I had images turned off in my email program and thus the only information I could see immediately was the name and address of the company. That’s the only part of the email they had coded in plain text. For me it was annoying, for those with vision impairments–who rely on screenreaders–it was just spam, because they had no way to read the message at all.
This prompted me to write 5 Tips to ensure your readers can read your HTML e-mail messages in which I made some of the above points and give examples on how to code for e-mail to avoid these problems.
As you indicate this doesn’t mean we have to throw design out the window. It just means we have to code carefully so our e-mails will degrade cleanly and not lose the message in the majority of e-mail programs.
February 27th, 2009
It would be a mistake to blame “designers” for HTML email failures. A quality designer puts function and message above “pretty.” Anyone who doesn’t do that is a decorator.
That said, I don’t like HTML email. IMO, the web is for web pages, and email is for plain text. Email clients don’t follow any sort of web standards, so coding HTML email is hell. IMO, extraneous graphics just clog up the tube. With the recent explosion in reading mail on mobile devices, does HTML email serve or defeat message delivery?
In an ideal world, we could have well-rendered HTML email on all devices, and we’d all have quality designers to make that experience pleasurable. But that’s not the current world.
My company is considering offering HTML email for our newsletters. A handful of customers have asked for it, and we’d sure like the tracking you get from serving images. But we’ve been resisting for all the above reasons.
So where does this leave us?
~brian
@briandigital
February 27th, 2009
@Brian - Designers tend not to know the “rules” of doing HTML emails (like Heidi points out in her post) - that’s not a failure of the medium. That’s their failure in working with a medium that they don’t understand.
It seems as though you’re resisting based on your personal preferences rather than what your customers want. You can have HTML email that works well on all devices - it’s really not that hard if you follow the rules. If your users want to receive text message only, they can set up their clients to do that. Tracking from images isn’t the only benefit from HTML (actually it doesn’t really do you much good because “open” doesn’t mean much - it doesn’t tell you if they read it) - enhanced conversion tracking is really what you’re looking for. With text emails, you can’t integrate Google Analytics into your track (well, you can, but it looks god awful). With HTML, you can. That’s the real info you’re looking for.
February 27th, 2009
I’m going to have to disagree with you here, Karlyn. A well-educated web designer knows the rules, but mail clients don’t interpret them properly. I’ve been following the issue at The email standards project> where you can see the results of their research. Especially interesting is their recent post on mobile devices, but also see their “state of the clients” side bar, where they grade the common email clients… even some web-based clients can’t handle HTML mail.
It’s a mindfield, and right now, doing what’s best for our customers (getting our message through unscathed to their choice of email client) is winning out over what may be best for us (tracking their attention). I hope the two merge some time in the near future.
February 27th, 2009
You can disagree all you want Brian but you haven’t said anything to convince me. I’m aware of the email standards projects and, as someone who’s been building HTML emails for close to 6 years now, I’m well aware of the problems. If you follow some pretty simple rules, it’s not as much of a minefield. But most designers won’t follow the rules because they don’t allow them to do the things they want to do.
February 27th, 2009
I’m curious if your “simple rules” include things like using tables for layout. Convince me that designing HTML email that works everywhere is not a hack. I want to be convinced, but nothing I’ve seen yet has done so. I’ve been designing with web standards for a long time and the thought of shipping emails filled with hacks makes me feel nauseous. Especially with our clientele.
Opt-in to HTML email solves my personal qualms with non-text email, because it’s serving the recipient what they ask for, and some people genuinely prefer HTML email.
Can you offer something to convince me? I’m all ears. (PS-this is not trolling, I’m honestly interested in what you may say here, or at least, link to.)
February 27th, 2009
Yes it absolutely does involve using tables for layouts because that is what works and I won’t convince you that it’s not a hack because it is. Designing for email and designing for web are two completely different things - again you’re confusing what YOU WANT with what WORKS.
If the recipients are using clients that read HTML email, wouldn’t you assume that’s what they want? Otherwise, why would they use it?
February 27th, 2009
“you’re confusing what YOU WANT with what WORKS.”
I want standards-based code because it’s proper and sustainable. There’s a reason it’s called a “hack.”
“If the recipients are using clients that read HTML email, wouldn’t you assume that’s what they want? Otherwise, why would they use it?”
Making assumptions about your users is *the most dangerous thing you can do.* Really, do you think the vast majority of your audience chose their email client? And how many email clients out there don’t support some, bastardized form of HTML? Only a handful.
How many people ever change the defaults on their computers? Probably less than 10% How about on mobile phones? The better question is, how many even know they can get different email clients on their phones?
Tables for layout does not degrade well for the mobile experience, on any platform. And how do people who have screen readers access content in tables? Answer: they’re deleting your emails, never getting to the opt-in text-only option link at the bottom. I love getting Timbuk2’s HTML emails, because they’re visually stunning. But they are completely unusable on my iPhone.
Now, all that said, you have to talk to your customers… if you’re marketing internally, and you know IT has forced Outlook on everyone, and no one’s going to get this on their Blackberry, and you don’t have users with accessibility needs, I’m fine for a hacked layout in that case.
But I’m less convinced than ever that HTML “works” in email.
February 27th, 2009
“I want standards-based code because it’s proper and sustainable.”
Why on earth does an email need to be sustainable?
“The better question is, how many even know they can get different email clients on their phones?”
Sounds to me like you have a problem with the way they’re using technology because they’re not doing what YOU want them to do. Again, you’re taking it personally when you shouldn’t be.
“Tables for layout does not degrade well for the mobile experience, on any platform.”
Emails don’t need to degrade.
“And how do people who have screen readers access content in tables?”
If you use a simple layout, this isn’t a problem. Simple layouts, BTW, statistically work best with email when it comes to conversion.
“But I’m less convinced than ever that HTML “works” in email.”
Well, I’ve raised hundreds of thousands of dollars in the last year using my hacked HTML email so I’m going to assume that my users like it just fine. On the other hand, you can’t even track what’s going on with your text emails so you have no clue if they’re working or not.
I’m going to say this again because I really don’t think you’re getting it - this is not about what YOU like. What YOU like is irrelevant. This is about what is going to get RESULTS. If you need to use something that you don’t like to get results, then suck it up and get it done because that’s what marketing is.
February 27th, 2009
(Well this has been fun but I’ll have to let this go after this reply…)
“Sounds to me like you have a problem with the way they’re using technology” “Emails don’t need to degrade.”
These statements don’t make any sense at all. Your customers are choosing to access their email on a mobile device. Text email gets to this increasingly growing segment, HTML FAILS here. We need to allow them to read our message on their terms, not your terms.
“If you use a simple layout, this isn’t a problem.”
Have you ever seen a screen reader (or properly, listened to a screen reader) read a table-based layout? Since they have to follow rows and columns, they don’t follow the visual flow, and they announce aloud each table row, header, cell, column.
“I’m going to assume that my users like it just fine.”
This is a big problem when you give someone analytics, they start to think they know what’s going on in the real world. After all, analytics that lead you to believe the wrong thing can be worse than no analytics at all. In fact, you know very little unless you’re visiting your users and observing their behaviors. Secondly, have you ever measured the difference in $s made off of people who use HTML emails and those who chose text-only? (I’m assuming you have a non-HTML version; dangerous, I acknowledge).
We won’t know the answers to our conundrum fully until we test an HTML option against our text-based campaign. But at least right now, we know that a) we’re making money and b) we’re not excluding anyone and c) we have a lot of people who tell us they really love our newsletter. To do this testing, we first have to find an email package that we like and that can send to our enormous list without breaking the bank.
Sadly, I think your “results” are based on less reliable data than you believe. Maybe your email is working just fine for your users. But I’m not sure you have the full story to back that up. Maybe you’re leaving money on the table. After all, you could assume that those people actively getting their email on the go are more affluent, right? But you won’t know until you do real life user research. Without that, it’s all guess work, analytics, or not.
And why leave your business up to guess work?
(Any how good luck to you. It’s been a fun debate.)
February 27th, 2009
“These statements don’t make any sense at all. Your customers are choosing to access their email on a mobile device. Text email gets to this increasingly growing segment, HTML FAILS here. We need to allow them to read our message on their terms, not your terms.”
1) Do my users access emails on mobile devices? Not really. Maybe yours do, but mine don’t
2) Even so, I test all my emails on mobile to make sure they work. Haven’t had a problem with it.
“Have you ever seen a screen reader (or properly, listened to a screen reader) read a table-based layout? Since they have to follow rows and columns, they don’t follow the visual flow, and they announce aloud each table row, header, cell, column.”
Yeah, I actually have. This is why I use SIMPLE LAYOUTS with one column and one row, and the table wrapped around it just to keep it at 600px.
“This is a big problem when you give someone analytics, they start to think they know what’s going on in the real world.”
You’re not even reading what I’m saying here. I’ll spell it out:
users clicks on link in HTML email + gives my institution money = my crappy hacked HTML emails get results.
What other way is there to interpret that? What are you results from your text emails again?
“Secondly, have you ever measured the difference in $s made off of people who use HTML emails and those who chose text-only? (I’m assuming you have a non-HTML version; dangerous, I acknowledge).”
Wow condescending on my blog batman. Yes, I use multi-part emails. No I don’t track text emails because I can’t without making them look god awful with links that span like 3 lines. The point is mute, however, because only 2% of my users ever see those emails.
“We won’t know the answers to our conundrum fully until we test an HTML option against our text-based campaign.”
Well, you can’t accurately track text emails so this is basically an impossibility…
“But at least right now, we know that a) we’re making money and b) we’re not excluding anyone and c) we have a lot of people who tell us they really love our newsletter.”
I think you’re a prime example of why I say web people need to understand marketing - you don’t crate materials for the minority. You create materials for the majority so you can get as much from them as you can. When you create materials for the minority, you end up not getting the best results you can from the majority. Anecdotal evidence really doesn’t mean a whole heck of a lot to me - you’ll always have some people who love what you do and some who hate it. Unless you can show me what you’re doing is producing results through data, it’s meaningless. Warm fuzzy feelings don’t pay rent.
“Sadly, I think your “results” are based on less reliable data than you believe.”
Really…because I can track a user from the moment they click on a link in my hacked HTML emails through the moment they give a gift…and I know how much that gift is for…and my conversion rates from click-to-gift (not to mention delivered-to-click and open-to-click). How exactly is that unreliable? How awesome of you to make assumptions about what I’m doing when you haven’t taken the time to remotely understand how I’m doing it.
“After all, you could assume that those people actively getting their email on the go are more affluent, right?”
Now who’s the one making assumptions? Doesn’t everyone and their mother have a cell phone now-a-days? Why on earth would you assume affluence because of that?
“But you won’t know until you do real life user research.”
Honey, I work for an organization in the process of raising $1.3 billion. So far we’ve raised $1.16 billion. You think that happened without real life user research?
:::waves goodbye:::
February 27th, 2009
OK, I guess I’ll try to make this my last reply because of two things I really want to set straight:
1) “Wow condescending on my blog batman”
I was absolutely not condescending on your blog. I was afraid you were taking this all as an attack, and this comment seems to support that belief. I came for a productive and civil debate. I was acknowledging the fact that I was making an assumption (for the sake of argument, not as a replacement for research) after just railing against assumptions. Thus, I was making fun of myself.
2) Really…because I can track a user from the moment they click on a link in my hacked HTML emails through the moment they give a gift
–right. For the people who click. To grow, wouldn’t you want to reach those who don’t click? And know why they haven’t been?
3) “you can’t accurately track text emails so this is basically an impossibility”
– we have a lot of links in our emails which they currently are clicking and buying. It will be easy to A/B test this by simply giving them different links… then tracking those who click is easy. We can also look at the percent of people who click versus those who don’t in each category. But this data won’t replace observation. It will supplement it.
4) “everyone has a cellphone” - yes, but most aren’t using smartphones for email. Does your mother get her email on her Blackberry? Email users on mobiles are a minority of cellphone users, but are growing. Smartphones and data plans are expensive, thus it’s logical to note they’re sold to more affluent people. This isn’t a giant leap here. And all smartphones render text emails reliably.
5) “you don’t crate materials for the minority. ” - right, this is why we have accessibility laws… because too many won’t just do the right thing. We believe something that’s accessible to people who need a little help makes them more usable for those who don’t need that extra help. Just as it is easier to get around an accessible building than it is to get around one without, say, an elevator.
6) “Honey, I work for an organization in the process of raising $1.3 billion[…] You think that happened without real life user research?”
First off, if billions came through email, I’d be awfully impressed. Secondly, we have worked with billion dollar orgs (”names you know”) and seen them make really bad decisions based on false conclusions based on analytics and without actually visiting their customers. It happens way too much.
Again, all I’m saying is that you have to do what’s right for your users, and your users are possibly quite different from mine. So blanket statements about HTML being the best solution are dangerous.
I also don’t think it’s a leap that we can agree to disagree without being disagreeable.
February 27th, 2009
This is a fun debate you’ve been having!
In the end HTML e-mail isn’t going away. I was in the pro-text camp for years. Then I kept seeing stats showing HTML had a higher conversion rate, so I had to stop being a curmudgeon about it.
I don’t think Karlyn or anyone else enjoys having to layout e-mails with tables and inline CSS. It takes time, it goes against our desire to adhere to standards, and even then the result is less consistent across e-mail programs than a regular (well-coded) Web page is across browsers.
That said there are things we can do to alleviate the problems. Many services and bulk e-mail programs will allow you to send both an HTML and plain text version of the e-mail.
This let’s users get whichever they prefer. While we know most users don’t change defaults, we also know that we’re not the only ones sending them mail. If HTML email bothers them enough they’ll find the preference to turn it off, or they’ll discard every HTML email they get.
Tables do control the flow of information, but we as designers control the tables. Just as we have to put floats in the right order with CSS, we just have to make sure our content is well ordered in tables. If keep table embedding to a minimum and use a fairly simple layout this is quite manageable.
In the end we’re working for our clients/bosses and our readers. As such we just have to do our best to make sure our e-mails are as readable as possible, no matter what format we use.
February 27th, 2009
Heidi you’re so diplomatic
February 27th, 2009
As a designer (ducking Karlyn’s flying daggers), I’ve done a little bit of everything: print, email, web (yes, standards-compliant and proudly so).
And I’ll say that I agree with Karlyn’s general sentiment. There are a lot of designers who don’t take into account the specific medium in which they are working nor do they understand the marketing rules and reasons why a lot of things need to be. Just ask any web programmer who gets a print-minded layout that they have to manage creating in the much less flexible web world. We designers can be an unruly bunch.
The issue is that table-based layouts ARE the standards for email. I think of it the same way I would any language. Standards change based on the specific media through which you are communicating and you can’t treat email as if it were web because they are two different media. Just because they both use HTML and are read on screens doesn’t make them the same medium.
But Brian is right that specified link creation and good A/B testing is a good means of tracking the effectiveness of text-based emails. But it really isn’t realistic for people in positions that require email marketing to fight the HTML vs. Plain Text battle. The gods have brought fire to the earth and we now have deal with it. We can’t recork the bottle now; it just too damn late. So work with what you have with all of it’s wonderful and UNBELIEVABLY frustrating quirks.
My two cents…I’m now broke.