Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Spotify | Android | TuneIn | RSS
Deborah CarverDeborah Carver sees direct connections between her academic study of literature and composition and her work as a content strategist, content marketer, and SEO.
She also sees similarities between AI engineers and content professionals, both of whom endeavor to create meaning with language.
We talked about:
- her work as a consultant and the creator of The Content Technologist
- her discovery of Google’s knowledge graph in 2013 and how it helped her SEO work
- how her background in literature and mass communication made SEO work come naturally to her
- how grade-school sentence diagramming prepared her to understand entities, natural language processing (NLP), and other tech concepts
- the similarities she sees between LLM engineers and content professionals, both making meaning with language, just coming from different directions
- how her study of information science, library science, linguistics, and other academic disciplines informs her semantic work
- her data-driven approach to keyword research
- her take on the “call and response” nature of search
- how she balances her keyword research with customer and user research
- the ways that her study of poetry helps her discern user intent
- her early interest in natural language processing and AI and how it prepared her for the current tech environment
Deborah’s bio
Deborah Carver is an independent consultant and the publisher of The Content Technologist, a resource for content professionals working in the age of algorithms. She spent the first part of her career working in traditional publishing, then transitioned to working on SEO and digital strategy full-time in 2013. Focused on organic content performance and authentic digital connection, Deborah helps clients navigate what makes “good content” findable, usable, informative, and delightful. She’s worked with businesses of all sizes, from Fortune 500 to independent startups and is an avid trendspotter, a deeply experienced website content analyst, and a massive music fan.
Connect with Deborah online
Video
Here’s the video version of our conversation:
Podcast intro transcript
This is the Content Strategy Insights podcast, episode number 186. Many content professionals studied English in college. More than a few of them have worried about how they’d turn that knowledge into a career. Few have shown as well as Deborah Carver how the study of literature and composition connect with content strategy, content marketing, and SEO. Deborah sees direct links from her study of poetry and rhetoric to the skills she applies to give both her human customers and search engines the content they expect.
Interview transcript
Larry:
Hi everyone. Welcome to episode number 186 of the Content Strategy Insights Podcast. I’m really happy today to welcome to the show Deborah Carver. Deborah is an independent consultant and she’s also the creator of The Content Technologist, a website and newsletter for folks interested in content and technology. Welcome, Deborah. Tell the folks a little bit more about what you’re up to these days.
Deborah:
Hi. Yeah, so I am an independent consultant. I largely help agencies and businesses with analytics and information architecture on large content focused websites. And I am currently working on a series of courses that are launching throughout the year that are based on helping people understand or helping businesses understand how they can be found on the internet and how they can measure that impact of their visibility, so yeah.
Larry:
Everybody wants to be found out there and that’s notoriously difficult. And that’s one of the things you’re known for is your SEO chops, which is sort of how I… Well the way this conversation came about a month or so ago, you made this post on LinkedIn about, “Hey, what’s in your knowledge graph?” And this a kind of nice, provocative, intriguing post and about a workshop that you have. In fact, depending on the timing, this episode might air right around the time of that workshop.
Larry:
So I apologize to folks who don’t hear this in time to go. But anyhow, you discovered the knowledge graph, and specifically Google’s knowledge graph, pretty much when it was invented. Can you talk about the agency work you were doing at the time and how you came to discover this?
Deborah:
Absolutely. Yeah. So I spent the first 10 years of my career working in publishing and doing a lot of writing. I moved to the agency world when I realized I was creating all this content and I had no idea how it was performing, whether people were finding it. I knew all these tools existed, but I didn’t know that how they worked.
Deborah:
And so I went to a performance marketing agency. That agency was very focused on education and I worked with some really smart people. It was back in 2013 to… I worked there 2013 to 2016, and it was just a very fruitful time for search, for social. And the agency was very focused on education. And so from some of my colleagues, we were talking about natural language processing, the knowledge graph and how that worked. And it was all that kind of planted the seed. And as the past 10 years have played out, I’ve definitely seen how that natural language processing comes out in Google search results, now in the AI boom. I am pretty familiar with how it works just by practice and by working with clients and by using some SEO methods to create knowledge graphs for my clients, sort of in a way as much as we can through a website, so yeah.
Larry:
Well, right, because if you’re doing analytics and AI, but what’s the point if they don’t find you? And so the SEO fits in perfectly in that. But one of the things I liked about that was the way you were able to… Well, I guess first of all, let’s talk about the implications of that knowledge graph and that necessity that Google famously said at the time that, “We’re talking about things, not strings. We’re talking about entities, not necessarily the keywords that describe them.” Can you talk about how your approach has evolved with that understanding?
Deborah:
Yeah, so it started with the idea that… So I majored in English literature and my master’s is in mass communication, so I’m very familiar with communication, rhetoric, all these structures. And so it was very natural for me in working in SEO to move towards… Instead of keyword stuffing and the math formulas part of it, to move towards, well, all the other websites on this topic seem to be using these sets of words that are ranking, and that is clearly what Google is seeing, is this related set of words rather than just the number of times the keyword was used.
Deborah:
And at the time I was also… The Hummingbird algorithm had just been launched for Google, and that was when natural language processing was introduced. So it was just kind of a natural fit. I wouldn’t have liked the keyword stuffing area era of SEO anyway. I really enjoyed helping clients understand like, “No, if you speak more naturally, if you write more naturally the way that writers write about a topic, you will show up for that topic more often. And here’s the proof of these kinds of words.” But they’re also the words, if you’re talking about cakes, you also want to talk about icing and chocolate and recipe and party and all these words that are related, but not necessarily like cupcake. There’s a balance there and there’s a difference. And helping clients understand that difference was a big part of my job for a long time, and it’s still what I do. It’s helping clients understand what are the words that people are talking about and what are the other words that naturally fit into that writing.
Larry:
And when we talked a while back about this, you talked about some of the things that have come together for you out of your… Because when you studied poetry, you’re doing literary analysis, but also writing and rhetorical composition stuff. And so everything you just said fits in that frame, that you have to understand the analytical part of it, which you still do a ton of. And the composition part of it, the creating it. Can you tell me a little bit like, how was that not leap, but the connection between your academic literary analysis and rhetorical background? It sounds like it almost translates one-to-one into SEO and content marketing.
Deborah:
Well, and it’s funny because yeah, we all… Well, at least I don’t know if they still teach diagramming sentences, but I remember very clearly learning how to diagram sentences in 6th grade where you map out where the subject and the verb is and where the preposition goes. That was something I didn’t like doing, but I was pretty strong at. When I learned about the idea of entities and how they worked in natural language processing, those two kind of connected because it’s about finding the relationship between the words and how they are uniquely used and mapping that out in your head. And that made sense because it also fits with what I learned about writing later on, which is that every sentence connects to the paragraph that it’s in. Every paragraph connects to the first and the next paragraph. You want to make sure that everything is connected and that it makes sense. That translated really naturally to understanding entities and knowledge graphs, I mean, kind of. It took a while. I mean, it’s been 10 years.
Deborah:
And so once you understand the ideas and the foundational ideas, then you start to see how they work in practice. And so that’s how I built that out and still kind of how I see it and how I explain it to other English majors who might not be super familiar with code, but it’s like, “Okay, well you learn to scan poetry, you’re finding the patterns or you learn to close read poetry.” Analyzing content for SEO isn’t all that different from close reading poetry if you think about it as poetry as a code. You can get really nerd out on that if you want, but it is about finding the mathematical patterns or being able to see those and understand how that would be processed by a computer.
Larry:
Yeah, that was really interesting to me as we talked earlier about the… We talked a little bit about vectorization word embeddings and those kind of hover in mass, space, and algorithms, figure out how they’re related and generate or do things with that. But you have this, and I wonder now about these large language models, how many English majors worked on those? Probably not enough I’m going to guess.
Deborah:
There’s a lot of linguists and semantics professionals who I know are working on them. And I know that a lot of… I’m being very heavily advertised to train large language models from certain companies. I’m getting a lot of, “Oh, you’re an English major, you’d be great at this.” So I know that there is some in the process, but it is also strange how we see these as all wholly different disciplines and it’s like, “No, we’re just coming at the same idea of making meaning with language in different directions.” Yeah, that’s kind of how I see it. Yeah.
Larry:
Yeah. And you’re reminding me now of another thing we talked about, the history of that in your personal professional development. You were studying computer science and library science and data science around that same time, that agency time. Can you tell me how that came together?
Deborah:
Well, actually it was… For this course, I’ve just launched a course about keyword research and understanding how language is read by algorithms. And so I was researching the history of keywords and I was looking at the Ngram for Google. If you type in the Ngram for keyword, it spikes in the ’70s. And the reason that that happened is because in information science, so library science, they were trying to figure out the world of computers. Computer science was trying to figure out how to make recommender systems and how to get people to use Boolean searches. And also at the same time in the cultural studies world, which is closer to where my graduate schoolwork was, Raymond Williams, who’s a cultural theorist, wrote this book about keywords and how certain words in a language have more meaning than other words depending on the context and depending on how they are used in discourse in communications. So that was kind of all happening at the same time in the mid ’70s when the word keyword became just more commonly used.
Deborah:
And so it kind of surfaced to the top of linguistics, library science, computer science, data science and cultural studies all at the same time, which is I think is very interesting. Again, it’s like how we’re all coming at the same types of ideas from different directions. I think that’s also just really cool to see how much computer science borrows from these other disciplines, borrows from library science and information science, which has been a longstanding academic institution long before computers were mainstream or invented. So it’s definitely very interesting to dive into where that confluence happens when we’re talking about keywords or we’re talking about processing language and how that works with context, where that came from and the roots of that and the foundations of that.
Larry:
Yeah. You remind me, I was at the Information Architecture Conference last week and did a workshop with Ren Pope. One of the things, he had these tables kind of showing different terminology in different contexts. It’s very much like what you’re just talking about. He didn’t do it specifically about what you were just talking about, but you could probably take that same little framework that he had and drop in like, “Hey, how are you talking about semantics and meaning and structure, Mr. Computer scientist, Ms. Data scientist?” whoever it is. Again, kind of back to the connection between that insight, it led currently to this workshop you’re developing, but how do you use it in day-to-day practice?
Deborah:
Well, so I do a lot of actual keyword research for my clients. I mine the advertising algorithms from Google to find out which patterns people are typing into the search bar most often and how those relate either semantically or just in Google’s database to other patterns and other language, what comes up most often, what words seem to be associated with each other. And I look a lot at that to help guide clients in specific directions like, “This particular framing of language aligns more with your brand,” or “This is what people are looking for. If you didn’t know that that is attached to a brand like yours” and understanding how these large bits of data that are being mined by computers can also be mined by people and be understood by people maybe in a more natural way because it’s like, “Oh, I understand…” SEOs talk about things like search intent. It’s like search intent and context. “I understand that because I was an English major, I know how to read. I was very much trained how to read and analyze very quickly.”
Deborah:
So that comes out in the work that I do with clients and then advise them on how to build that into the actual information architecture of their website so they can be found by search more often. It’s a lot of, yeah, I feel like I’m like, sometimes I just do the same thing as the computer in processing all this data. It’s just that I’m putting my own brain and experiences into it and OpenAI has a different brain, so yeah, it’s-
Larry:
Well, I made a note-
Deborah:
It’s really fun, yeah.
Larry:
I made a note the last time we talked that you said, I think this is a direct quote, “We are all cyborgs now.” And you just supported that case right there. But one of the things about that you said, I think in that same context you said something about call and response, but is it conversations with machines? Is that what you meant?
Deborah:
Well, it’s more like, I mean, we’re all typing in search all the time. People do not realize how much they type into the search bar and/or what terms they’re using. It’s just such a natural part of our information retrieval and even conversation that search is this crutch that we lean on to find information or at least to find the baseline of information. And the internet is made up of that call and response. It’s made up of people publishing websites that are then scanned by search engines and that are processed and then spat back out to them through a recommender system in a search engine. And the search engine responds to how people use that information via clicks and via the time spent on the site. It is a conversation between humans and computers all the time.
Deborah:
Now, people think they can manipulate that conversation by poking at the computer’s weaknesses, but also it’s one of those, the more you make a pattern of doing that, the more the computer sees the pattern and then the computer can say, “Nope, that pattern is bad. And we don’t look at that pattern anymore.” It’s funny, I am not an engineer at all. I use computers creatively and for creative expression, but I just see it as like, optimizing an information architecture, it’s like preparing a group of magazines to be bulk mailed, like, this is just the distribution system, but the distribution system happens through the language that we use instead of through postage and it’s just a different system. It’s interesting to dive into because people search in ways that… Understanding keyword research, people search in ways that you would not believe the search terms they use. I mean, most search terms are pretty normal, but you always find some gems in there that it’s like, “People look for that? Okay.”
Larry:
Now human beings are always messing up our best-laid plans.
Deborah:
I know. I know. And it’s great. And it’s like, “That has that much search volume? What are they even trying to find?” But it’s great. I think being curious and fascinated by human nature also… Oh, if you just remember that it’s not a computer that’s spitting out these terms, it’s actually people entering in that data and then the data comes back to them and gets spit out back at them. So yeah, it’s very much a conversation, and a weird one.
Larry:
But it also reminds me of something else you said earlier. I think it was in our prior conversation about how all of this is about humans understanding how other humans think through this whole machine dialogue and everything else that’s going on. So that kind of gets at the fundamental intent. I’m wondering if you ever try triangulate on that with market research, user experience research?
Deborah:
Oh yeah.
Larry:
Oh yeah, I assume so.
Deborah:
Oh, yeah. Oh, yeah. Do that all the time. If you just use keyword data, you’re not going to get a full picture. You really do need to do your audience interviews, your other types of research, your user research. You need to look at your analytics. There’s a lot of different steps that kind of get you those seed keywords that you then throw into the algorithm and see what comes back at you. And yes, see what other people are doing. It is a weird dance. There’s triangulation, but it is triangulation and then there’s the call and response and you tweak things. But not too much because if you tweak things too much, then your users will get lost. So again, it’s the principles of user experience always come back. It’s like you don’t want to use words that are too out of the norm because then no one knows what you’re talking about. So it’s a weird dance.
Larry:
It’s like the Goldilocks just write thing at every juncture. Because one of the… You also talked about Paul Ford’s quote, I forget something about people want things organized but not too organized.
Deborah:
Yeah. Yeah. And that’s where personalization kind of comes in. People want to find their own way and their own path. Well, some people like things to be served to them. Yeah, there’s different types of people. It’s called uses and gratifications theory in communications. People use different media for different purposes and they like to have it organized kind of, but they also want to do their own organization. And that’s interesting. But yeah, searches and natural language processing is a way of organizing information that puts structure on it through the language that we use instead of through the buckets that we physically put it in.
Larry:
Interesting. So more like scaffolding than neatly organized warehouse shelves or something?
Deborah:
Yes. Yeah. I say it’s more like the foundations and not necessarily how not overly clean or overly clean or overly spic and span. I don’t know. It’s funny, I live with… My partner is a very tidy person and I am a cluttered person. And so we find this nice happy medium between clutter and tidy, and that’s kind of what search is. That’s kind of how I think about it. It’s like it’s not too tidy, but it’s tidy enough that we can understand what’s there and why.
Larry:
Nice. I want to circle back to something, we talked about it a little bit earlier, but one of the things I was really hoping to get from a former poet, a poetry student. So that notion of how that helped you move from the sort of literal word thinking, and this kind of goes back obviously to the keyword and entity, but getting to that conceptual thinking about entities. Tell me your process of getting… Or was that just an easy process for you because of your literature background or there’s still some work for you to do around that?
Deborah:
I mean, it’s always entities are weird because it is the conversation. So it’s always evolving and it’s always changing. Whereas a poem is set in stone. We are still studying The Flea by John Donne all these centuries later. But as an English major, yeah, we spent a lot of time close reading poetry and scanning it for the meter and scansion as well as understanding why the poet used this certain word and not that word. And sometimes there’s a poetic reason for it and sometimes there’s a meaning reason for it and a semantic reason for it.
Deborah:
And so this started to make more sense to me, especially the knowledge graph thinking started to make more sense to me when I compared it to how I used to analyze poems and how I used to have to write 10-page papers on analyzing a poem and why the artist’s intent or the artist’s poet’s intent or their meaning, what they meant by the poem and how that’s open to interpretation and how computers kind of do that too based on… I mean it’s the words that are the most specific. You don’t understand meaning from a sentence like, “I was there,” but you can understand that I was at the baths in Budapest or something like that. And it’s like Budapest and the baths go together.
Deborah:
It’s about being intentional with the meaning and understanding what you’re trying to say. Content strategy is about what is the business trying to say and what are the best words to do it, but also how do they match with the words that other people are using. So figuring that out is interesting. But yeah, it’s a lot like understanding a poem. And that’s how when I work with English majors and writers who are like, “I don’t understand this code stuff,” it’s like, “All right. But you do know how to analyze a poem so let’s go back to that because it’s that same thing. It’s just applied differently.”
Larry:
Mm-hmm. Yeah, as you’re talking about this, it’s occurring to me that there’s all these specific things you do. And a lot of the questions I’ve asked have been very specific about particular background things. But I’m wondering now if this is as much about a mindset that you established, that you’ve cultivated that kind of becomes then like not a heuristic and not a shortcut exactly, but some kind of fast onboarding into this ontological understanding. Does that make sense?
Deborah:
Yeah, I mean it is about how we learn things in school and the subjects that are defined in school are not how they actually play out in the real world. There’s a lot of blurry boundaries. It helps to be able to see how the same concepts apply across those blurry boundaries and to kind of question a little bit what you’ve learned and be curious about how what you’ve learned applies. And also, I mean maybe it’s just me trying to justify how my English major has eventually helped my career because everyone’s like, “You won’t get a job with that major.” And actually, I think it brings a lot of creative thinking to my career and helps me come at it at a different way than many of my colleagues. So it just helps me see things differently. And that diversity of thought is what makes it good. Having a lot of different people weigh in on it from different perspectives is that’s a good thing. It’s not just about your college major, it’s about how you use what you learned and how you make sense of it in your work.
Larry:
Nice. Yeah. I had a minor in English. My major was in journalism, but I still had the same concerns. But one thing you mentioned before too is that you were writing about and studying and working with the AI stuff like five years ago. And at the time, nobody cared.
Deborah:
No one cared, yeah.
Larry:
Yeah. And now people care. Has this helped you cope with all the current fuss around large language models and ChatGPT and that whole world?
Deborah:
I think cope is an interesting… It’s a good word. It’s just funny because yeah, I have a post on natural language processing algorithms that dates to 2020. And I was just looking at it and I was like, “Okay.” This post, it was not a very popular post. And now when I look at it, I’m like, “This was so ahead of its time.” It also helps you see as a professional communicator what actually changes conversation, and a lot of it is big splashy product releases and not the actual knowledge itself. I guess it helps me understand it, although it does increase my frustration, I would say with press coverage of it. Although the press, now that they’ve had about a year and a half with OpenAI and ChatGPT, it’s like you can see the understanding is starting to evolve and the just general knowledge of how LLM’s work is starting to evolve. So yeah, it helped a lot.
Deborah:
I have worked with Writer, the LLM and enterprise AI tool. They have also been really smart, and they reached out shortly after that post and we’re like, “Oh, we seem to be on the same page.” So it has been helpful to kind of understand what other companies are doing, not just Google and not just just IBM Watson, and understanding how that plays out among different companies, not just ChatGPT. Seeing how the different ways that something can be and the same concept can exist, that helps a lot in my understanding of it as well.
Larry:
Yeah, and I think when I used the word cope, I was kind of projecting my own wrestling with all this, just all the hype coming out of Silicon Valley. Anyhow, there’s just so much going on.
Larry:
But hey Deborah, I can’t believe we’re coming up close to time, but before we wrap up, is there anything last, anything that you want to share with the folks or revisit from the conversation?
Deborah:
I would just say, if you’re interested in this, if you’re interested in learning more about this for your career or for your business, I am hosting this workshop at the end of the month on April 30th. It will be recorded. I’ll have it available, but you can also find out more about the full course at keyword.school is the domain where that’s sitting right now. And you can also find me at content-technologist.com or just Google The Content Technologist. I hope I come up first in my search results. I do when I do it. So I’m trying, trying to take my own medicine on the whole SEO thing, so yeah.
Larry:
I’ll do that search as soon as we wrap up. A lot of people have LinkedIn profile. Is LinkedIn also a good way to connect?
Deborah:
LinkedIn is a good place to find me, yeah. I’m there most days. And yeah, email is always open, deborah@content-technologist.com. Sign up for the newsletter. There’s lots more thoughts and digressions where this came from for sure.
Larry:
Nice. Well, I’ll put all that in the show notes as well. Well, thank you so much, Deborah. I really enjoyed the conversation.
Deborah:
You too, Larry. Yeah, this was really fun. Thank you.
Leave a Reply