Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Google Podcasts | Stitcher | TuneIn | RSS
Karen McGraneThe web has changed a lot over the past 25 years. Or maybe it hasn’t.
Enterprise content architectures are maturing and finally beginning to separate content from its presentation. But old-fashioned artifacts like PDF files still abound, and authors still expect WYSIWYG editing experiences.
And when Karen McGrane reflects on her work at Razorfish 25 years ago she’s struck by how the team structures and business practices they adopted there are still relevant today.
We talked about:
- her work at Autogram, the agency she founded with Jeff Eaton and Ethan Marcotte
- the content modeling practice at Autogram
- the implications for content strategy practitioners of the emergence of headless CMSs and decoupled content architectures
- the enduring challenges presented by PDF documents and WYSIWYG interfaces
- her take on the benefits for any business of having a niche, like Autogram’s focus on content modeling and design systems
- the evolution and possible merging of content systems and design systems in enterprises
- the uneven adoption of new kinds of authoring experiences for decoupled systems
- the role of language and the importance of using shared terminology
- her take on content personalization
- how they’re using AI at Autogram
- her observation that tagging may be one of the best uses of AI tech in content workflows
- her reflection on the early days Razorfish and the genuinely visionary work practices that emerged there
Karen’s bio
Karen identifies and solves problems with content management and user experience design across print, web, and mobile. She has partnered with some of the world’s largest enterprise businesses to streamline their digital operations and governance.
On a good day, I make the web more awesome. On a bad day, I just make it suck less.
Connect with Karen online
Video
Here’s the video version of our conversation:
Podcast intro transcript
This is the Content Strategy Insights podcast, episode number 151. As web technologies and business practices have evolved over the past 25 years, publishing, user experience design, and enterprise software systems have all converged on each other. Few content strategists have had as good a front-row seat to observe and participate in these developments at Karen McGrane. From her early days at the pioneering agency Razorfish to her current work at Autogram, it’s not an exaggeration to say that Karen has seen it all.
Interview transcript
Larry:
Hey, everyone. Welcome to episode number 151 of the Content Strategy Insights podcast. I am super extra delighted today to welcome to the show Karen McGrane. Karen probably needs no introduction, but I’m going to give her one anyway. She is one of the longest practicing content strategists out there, and she’s gone by other labels, as well, information architect and other titles in the past. But one of the things that distinguishes her, many things distinguish her in this profession, but one is that she came in right out of a human computer interaction program at Rensselaer Polytech. So, she was ready for all this design stuff that we’ve been adding to the profession the last few years, way before any of us.
Larry:
She was, famously, the UX lead at Razorfish in the early days of that famous agency. She’s worked for everybody, the New York Times, Conde Nast, Disney, Citibank, Time Inc., you name it. She wrote a couple books on Content Strategy for Mobile, on Going Responsive on responsive web design. She also did a podcast with Ethan Marcotte, the inventor of responsive web design, and they codified and articulated the whole notion of this now embedded practice. And she taught for many years, and still does teach, at the School of Visual Arts in New York City. I’m exhausted from that intro. Karen, tell the folks what you’re up to nowadays.
Karen:
Well, thank you for that very thorough intro. I think you had things on there that I’d forgotten about. So, these days I am a partner at Autogram. We are a three person boutique consultancy. My business partners are Jeff Eaton, I know he has been on the show, as well as Ethan Marcotte, who also has. And the three of us, we think we have a really unique overlap of mutually reflective skills in content design and technology or front-end and backend work. And we really think we understand how websites get made as much as anybody does. So, our goal is to consult with enterprise clients. We’re also starting to do some work with partnering with CMS integrators, working on helping organizations sort out their content modeling in support of doing a big re-platforming of a CMS or a redesign of a website.
Larry:
Cool. You just said my favorite word, content modeling. That’s my main thing these days. So, I will selfishly ask, in a lot of my modeling work, it’s been driven by the emergence of headless CMSs and decoupled architectures.
Karen:
Absolutely.
Larry:
Yep. And tell me, because you were talking about modular content 25 years ago, and you’ve talked for many years about the decoupling of presentation from the content itself, and modeling is the thing that makes all that work. Can you tell me about the modeling practice at Autogram? How you do it there?
Karen:
Yeah. I mean, honestly, I think modeling is a core. It’s one of our core competencies, or most of the time if a client’s going to want to work with us, modeling is going to have some aspect of it. What we’re focused on right now, in particular, we’re aiming at a specific problem around content migrations. And you say content migrations to people, and it inspires horror stories, right? Anybody who has done a big replatforming project has been through some scope explosions, basically. You find things that you weren’t expecting, things are much more spaghetti than you think they’re going to be, something doesn’t manage properly, and it winds up taking a lot more time and honestly, a lot more pain than was expected. So, I think we’re trying to reduce risks for clients and reduce risks for agency vendor partners in scoping and planning for a migration.
Karen:
Part of that is the inventory, and I think we probably talked with you all about Spidergram, which is the web crawler that we built ourselves because we didn’t think Screaming Frog did a good enough job. But I think, more than that, I think it’s an angle to get clients who might not otherwise care about modeling or might not necessarily see how much effort they’re going to put into the modeling or what value they’re going to get from the modeling to be able to commit to spending some time with us up front to do a very focused strategic piece of work to try to outline what that model should be. I think, honestly, it saves a lot of time. You can either pay that at the end of the project in development time overage, or you can do a little bit of work up front and, hopefully, reduce the pain.
Larry:
Yep. No, and the way you just said that, too, I’ve just become more aware lately of the relationship between auditing and modeling. Any modeling project starts with an audit of what you’ve got there, so there’s that. But also, maybe it’s because you put it in the context of a migration, that’s the unique one reason. I just had Paula Land on a few weeks ago and she was talking about-
Karen:
Oh sure, yeah.
Larry:
Yeah-
Karen:
Yeah.
Larry:
… about the various intents of an audit and one of them is that, a content migration. So, that’s really interesting. And then, is the intent to future-proof the content? Well, I guess, does this get back into that decoupling, too, of the content itself and what people are doing with it?
Karen:
Yeah, yeah. I mean, absolutely. I think the move to headless is, I think any of us working in this space see it as, probably, one of the major technological evolutions or trends that we’re all wrestling with right now. And, yeah, it’s a motivator for some organizations to get off WordPress, for example, and go to a headless platform. I know some of the headless work that organizations do is new, experimental. They’re not doing a migration because they’re using it for some new thing that they’re developing. But as headless has matured, I mean, I would say it’s at a fairly mature point in its life cycle, organizations, large organizations are thinking about re-platforming existing content over to headless.
Karen:
And that means, for sure, they need some kind of inventory and having a sense of that model, even if all they want to do is replicate the current state model, which I’m not recommending, but that’s what winds up happening in a lot of cases. Having at least a good version of what that current state model is is useful. But me, personally, I think they should hire Autogram and also have us figure out what the future state vision intended use of both the model and the front-end information architecture should be like. That’s the projects we really want to do.
Larry:
Right. And as you’re talking, too, I’m reminded that it sounds like a lot of your work is just driven by the budding maturity of digital practice. The fact that there are old things out there that need attention, that that’s a your business right there.
Karen:
Yeah. I mean, it’s one that never really gets old. So, organizations having to remediate old content that worked for them in a previous era with previous technological paradigms and constraints that now no longer works. Like PDFs, right? Honestly, I would love it if every financial services company in the world would hire us to remediate their old PDFs. Yeah, we can help you figure out how to get the stuff from your PDFs on the internet. Let us. It’s time.
Larry:
Yeah, I love it. You’re reminding me now, I did a lot of research for this before we talked. And one of the things I didn’t realize before, that you’re a conspiracy theorist, that you’re convinced that Xerox created WYSIWYG interfaces to sell more printers.
Karen:
They did.
Larry:
And I wonder if there some analogy-
Karen:
No, they did. That’s not even a conspiracy.
Larry:
That’s not even a conspiracy?
Karen:
Yeah.
Larry:
That’s an actual documented fact, isn’t it?.
Karen:
That’s an actual cited source, yeah.
Larry:
But a PDF has got to be the most WYSIWYG kind of content format out there, too.
Karen:
I understand why it happens, but I really just think it’s time. It’s time. Especially, we have banks and financial services firms as clients, and sometimes I go poke around and see what they’re still managing as PDFs on their website. And it’s like, “Come on. I know you need them. I know some of them are going to still exist, but some of this is just laziness. You haven’t bothered to fix a problem, and you’re still acting like it’s 10 years ago.”
Larry:
Right. I don’t want to get too far in the weeds on the actual technical part of it, but it can’t be that hard to get the content that’s in the PDF out of there and into a proper CMS.
Karen:
I mean, it’s harder than you would think just because of the way PDFs store the content. I mean, it’s a WYSIWYG problem. I mean, it’s not structured in there, and so you just get a lot of garbled text out of it. There’s things you can do, but it’s not always a slam dunk to be able to just convert straight out of a PDF into HTML.
Larry:
Yeah. Hey, the other thing is, as we talk about this, is a couple of things you said, and I don’t know if you’ve always said them in tandem. I have them together in my head now, but that every company is now a technology company, and that every business is in the user experience business. And so, if that’s true, why are they still hiring Autogram? They’re not all mature enough yet, right? Or tell me about that, yeah.
Karen:
I am a huge fan in business of having a very specific niche. When I started out, sure, user experience was a small enough field that I could say I’m an information architect. I used to make this joke. If you knew enough to apply for a job as an information architect at Razorfish, you probably were qualified for the job because nobody knew what that was. Now, I would not want to go to market saying, “I hope businesses fix their user experience problems.” Too generic. What I want to do is help businesses solve this specific problem. Particularly, I mean, I guess I would say broadly it’s around content modeling and design systems and the relationship between them. And a lot of times that comes to a head during a CMS replatforming or a redesign of a website or some enterprise-wide design system initiative. Not very many people or organizations are going to know that they have that specific problem. But when they do have that specific problem, I want them to go, “Oh, we have to hire Autogram because that’s what they do.
Larry:
Yeah, no. And now, I’m thinking about does the role of design systems, what percentage of your projects now are in enterprises that have a design system? And a follow on to that, your content work. How does that get integrated into their design systems?
Karen:
Yeah, yeah, yeah. I would say a hundred percent of the organizations we work with have a design system. I’d be really surprised, I mean, at varying levels of maturity, by all means. And, yeah, I mean, we’re usually being hired in some capacity to talk about how the content, structure, and the components and the naming of the components and the meaning of the components, how that’s going to interrelate. And, to me, that’s the most fascinating problem in the business right now. I joke about it. I honestly have been trying to solve what I think is this exact problem for 25 years now, and I will probably be trying to solve it on the day I finally retire. But I think it’s a fascinating problem. Mobile was another version of the problem. How do you make sure that your content isn’t attached to a particular screen size? And now, we’re dealing with it with headless CMS and component-based design systems, and how do these two things feed each other? And how do the people working on both systems make sure that they work together?
Larry:
Yeah. Give me your top level, because there’s so much to that. But are they merging? Or are they becoming co-equal thingies in enterprises? How is that unfolding?
Karen:
I talked to somebody recently who said that they have a systems team, and the systems team is responsible for both content and design systems, which I thought was pretty clever, and I thought that was great. No, I think, again, it’s a level of maturity. And to be fair, in one classic scenario, what you have is a development team that’s using headless as a content repository, and they’re also the people who are creating whatever’s going to be on the front end. So, it’s not like there’s a separation between the two. But as the systems get rolled out enterprise wide, as you have other teams being expected to use the design system in some sense, or create content that is then going to be laid out with this design system, it becomes a language problem.
Karen:
It becomes really a communication problem within the organization. And a design system that that’s got its start to solve a technical problem for a development team or for a smaller team can often stumble if it gets traction and there’s this sense of, “Oh, we should roll it out more.” And you don’t have the shared language and you don’t have the shared understanding and vision of what the system’s supposed to do.
Larry:
Right. And that shared understanding that’s so crucial in anything. But-
Karen:
Absolutely.
Larry:
… I’m just trying to picture just one lens on that, like the authoring experience. In CMSs, Dean Barker and Greg Dunlap famously talk about how everybody hates their CMSs because they hate the authoring thing.
Karen:
Oh, sure, of course.
Larry:
Yeah. And I don’t know if there’s an analogous thing in design systems. Anyhow, is there an analogy there in how the components are built, or how do you assemble the content elements so they go into a design system? And conversely, the design system people, they’ve always had to account for content in design systems. I’m just wondering how that’s maturing and evolving.
Karen:
Yeah. Honestly, I think it’s all over the map. And I think one of the big questions for me around, and I think this is a very common theme in headless CMS, is how much of the content is structured, the content author is expected to be in there filling out content templates, and the goal is to have that separate from presentation. So, either the content has low enough variability that they don’t really need to be considering the design, or their role is set up in such a way that it’s not expected that they should care about the design that much or have to worry about it. But that’s just not how human beings operate. So, then you get into the whole situation of like, okay, well, then what do you do with preview?
Karen:
But it then just starts to bleed over into the question of, well, do we need a page builder? And what is that page builder going to do? And are we going to have a WYSIWYG? Are we going to give people the capacity to have Squarespace? So, whatever’s going on, you may have a headless structured presentation-independent content repository, but then you may also have people out there who are building pages out of stuff. So, then, you start to explore the idea of, okay, well what if we had some thing that we could do it with a design system? So, rather than just having Squarespace spaghetti, you could have some structured components to it. But then, it really becomes a question of, okay, well what does that authoring experience look like? And how do you convey to the people who are actually going to be constructing those pages what the rules are? How do you make sure that the pieces all get put together in appropriate ways? It’s not an easy problem, but it’s a fun problem.
Larry:
No, it’s really interesting too because as you talk about that, the idea of a page builder, I just bristle at it because, come on, people, were in the 21st century. But the way you just said that, I’m almost wondering, one thing that occurred to me is can harvest structured content out of some page builder experience that accounts for them. But also, I think the bigger thing about that is that a lot of that’s driven by, I guess, what some people call the need for an omnichannel content strategy now. When did you write the mobile book? That’s like 10 or 12 years ago, at least.
Karen:
2012. In 2012.
Larry:
2012. So, 11 years ago.
Karen:
Yeah.
Larry:
And so, this has been a known issue for at least that long and maybe a little longer.
Karen:
Oh, sure.
Larry:
Yeah.
Karen:
Yeah, yeah, yeah.
Larry:
And so, that need to accommodate different delivery places with the same or very similar content, and then that need to accommodate the human beings who are authoring it, you just talked about that a little bit. But have you figured out any systematic ways or authoring environments or authoring strategies to streamline that?
Karen:
Yeah, I think so much of that is going to be tool dependent. What capabilities does your CMS or your digital experience platform have? Is it a composable system where you may be using one tool for the headless piece, and a different tool for a page builder, a WYSIWYG. And then, what’s happening on the front end? What’s going on with the front end developers that are actually working on the design system? That’s a whole workflow that would need to be mapped out for an organization to try to understand what are the tools capable of? What are the content people capable of? What is the development team capable of? How much time do they have? And again, I’ll just keep coming back to it’s a language problem. You have to have a shared language for what it is that you’re doing and why you’re doing it. Otherwise, it falls apart.
Larry:
Yeah. And that language thing, all of a sudden, I’m super jealous that you get to nerd out on language with Jeff Eaton all day.
Karen:
Oh, my God. It’s the greatest thing. Our Slack is the best place. Honestly, I have wanted to run this company with them from as long as I’ve known them. I have thought it’d be really great if the three of us had a company. And safe to say it is.
Larry:
Yeah.
Karen:
Why would you hire us? Just because we are so great to work with.
Larry:
Yeah. You’ve all collaborated. I know you’ve had multiple collaborations with both of them prior. Was it just a matter of time being right? Is that how you…
Karen:
Timing. I had to get Eaton to quit his job at Lullabot, but I was emotionally successful.
Larry:
Nice. Yeah, and they seem to be doing okay without him, but then-
Karen:
I think that, yeah. No, I mean, Lullabot’s a great company. I have lots of friends from there, but he was there a long time, and eventually you want to do something a little different.
Larry:
Yeah, I know that. Hey, and for some reason, I think Jeff and I talked about this at Confab. But, anyway, the notion of personalization, personalized content experiences. That’s one of those things that, I can’t remember. It was somebody at Confab I was talking to about this. But just that it seems like there’s way more people talking about it than doing it, than actually delivering.
Karen:
Oh, yeah.
Larry:
Yeah.
Karen:
Yeah.
Larry:
Do you have theories on that or thoughts or any success stories?
Karen:
I mean, yeah. Personalization is not my thing. Again, I think it’s like there’s specialized tools and technology and approaches that I’m just not up on. I mean, fundamentally, I think there’s an underlying need for structure and taxonomy and metadata and everything that goes into personalization, but the types of projects we work on tend to be motivated by other goals. But, yeah, I’ve always thought of personalization as a marketing shiny object. It’s the thing people get excited about, and they think they can throw some money at it and see what happens. I think we’re going to be hearing a lot about AI used for personalization. I mean, we’re going to be hearing a lot about AI used for everything, but personalization, I think, especially is going to get on that bandwagon.
Larry:
Yeah. Hey, I guess I would be remiss to not get your take on AI, given that it’s the spring of 2023.
Karen:
Yeah.
Larry:
How is that affecting your work at the agency? Or is it?
Karen:
Yeah.
Larry:
Tell me how all you are approaching the AI.
Karen:
Yeah, yeah. In fact, a little plug for our newsletter. So, Autogram has a newsletter, and the issue that just came out yesterday was about AI. So, if anyone wants to subscribe to that, they can go to our website. To your specific question, no, I wouldn’t say it has affected the work that we do. I mean, there’s things we use it for. AI transcription technology is delightful. And I have not personally gotten a lot of value out of using Bing Chat or ChatGPT, but there certainly are scenarios where I can imagine it might be useful for doing relatively routine tasks, like editing or reviewing something. I mean, it’s not something where I would ask it to do work for me, but I might ask it questions.
Eaton plays around with that. I know he’s experimented with all kinds of things. The most value I could probably see us getting out of it in the near term would be some of the AI machine learning tools that we can use with Spidergram when we’re producing inventories. There’s a lot of things, like grammar checkers or a linter, like Grammarly. The text you could do automate tagging. Automated tagging, honestly, if you put a gun to my head and said, “Tell me what value you’re getting out of AI,? I’d probably pick automated tagging.
Larry:
That’s interesting. That reminds me, again, of the authoring experience. Maybe it doesn’t need to happen there now. Maybe you just…
Karen:
Right. Yeah. I mean, I have long said you want to use computers for the things computers are good at, and there’s some things that humans are bad at that computers are good at. And tagging is, honestly, probably one of those things. And I’m not saying that there’s not going to be a human being still involved in the process, because, of course, there has to be. You need somebody to set up the conceptual space, the taxonomy, the terms that are used, you’ve got to train the robots. But once you’ve done that, and you have a process for reviewing things, let the machines go solve that problem. It’s complex and time-consuming, and people are bad at it.
Larry:
Yeah. And I think I’m not particularly lazy, but if a machine can do it for me, I’m all for it.
Karen:
Yeah. I mean, there’s some things machines are better at doing, and I think that’s weird about AI is that now it seems to be blurring. I mean, back in the day, it was like, “Well, computers are really great at doing math. That’s why we invented them is because math is hard and machines are better at it.” And over the years, there’s been increasing numbers of things that computers are good at, and now it’s starting to get into some weird territory where it’s like, “Is it good at things that we think we’re good at? That’s supposed to be our stuff to do.” But I don’t think it’s really true.
Larry:
No, I’ve been bingeing a ton of like HBR and other podcasts, and the AI gurus out there, and the business ones always say the same thing. That every tech innovation like this that’s come along has created more jobs and made work, well, arguably more interesting. That might be some propaganda, but it is generally a positive thing, but it’s always stressful in the transition.
Karen:
Yeah. I mean, there’s that old truism that any technology invented before you were born is just the nature of things and has always existed. And any technology invented between the time you were born and age 30 is a cool new thing that you can stake your fortune on. And any technology invented after you’re 30 is the devil and should be feared. And I feel like I have this sense of myself, as now a person in their fifties who’s been doing this for 25 years, where AI has me doing my grandpa old man shakes a fist at cloud thing. But the thing is, the kids these days? I guarantee there’s a bunch of 20 year olds right now that are going to stake their career doing entry-level AI work at some company. And I don’t know what that’s going to be, but they’ll do it. And kids who haven’t been born yet, who knows what’s going to happen.
Larry:
Yeah. Hey, Karen, I can’t believe it. We’re already coming up close to time, but I want to make sure before we wrap, is there anything last, anything that’s come up that you want to elaborate on, or that you just want to make sure we talk about before we wrap up?
Karen:
It’s funny. I find myself musing over my career, and I feel like I always want to give a shout-out to the early days of Razorfish. I started there in 1998, and we had an information architecture practice. We had a content strategy practice. They were not the same thing. We had a robust user-research practice. I mean, this was all back at a time when the people seemed to think was pre-history. And I was like, no, we were doing it back then. I mean, I was literally making spreadsheets of websites 25 years ago. And it’s fascinating to see both how much the web has changed, but yet how much is still the same.
Karen:
I think you mentioned as we were chatting, I gave a talk at the first Information Architecture Summit, IA Summit in Boston in 2000, and talked about component-based content. And I feel like I have been solving that problem for my entire career. But to give a special shout-out to Razorfish, I got to work back in those days with a lot of really very, very smart, talented people who all went on to successful careers in the field. And it was fun to be around other people who cared about information architecture when nobody else knew what information architecture was.
Larry:
Nice. Yeah, and you look around our landscape, and there’s still a lot of those alumni who are leading really brilliant, big organizations now.
Karen:
Yeah. Yep.
Larry:
So, it stuck. Yeah. Well, thank you so much, Karen. Oh, one very last thing. If people want to follow you online or stay connected, what’s the best way to connect with you?
Karen:
That’s such a fraught question these days, because normally I would say Twitter. So, you could follow me on Twitter. I am still there. I am on Mastodon. I am on LinkedIn. But I don’t want to talk to you on LinkedIn. You can connect with me on LinkedIn, but that’s not where I want to be friends. I am on Reddit. I am a mod on the /r/UX design sub. So, if you want to check that out, that would be cool. Also, when Twitter collapsed, I shifted my social media energy over to Reddit, and now Reddit is collapsing. So, I don’t know. Mastodon might be your best bet in the future or go to Autogram.is. That’s our URL, Autogram.is, and sign up for our newsletter.
Larry:
Great. Well, thanks so much, Karen. This was such a treat to get to talk with you this morning.
Karen:
Yes, it’s great to talk to you, too. I had a lot of fun.
Leave a Reply