This is the second part of my discussion of Author Experience with Noz Urbina. Here, we venture into the realm of adaptive content, and the future systems that can expand the scope of how for it adapts.
Content Strategist, Urbina Consulting
Noz: There’s a thread that’s been running through a lot of my thinking lately. For the last ten years or so, many of us have known that we need to up our game in terms of content personalisation. We need to know a hell of a lot more about our customers than we did before.
We need to understand their lives. We need to understand where they are, what they’re doing and thinking.
We need to draft our use cases in more detail than ever before.
Rick: Absolutely. Though I would go a step further and say that we need to create personae in a very different way than we have before.
Noz: Please elaborate. How is it different?
Rick: As you say: where are they now? What are they doing? We need to consider that level of granularity, as it relates to our communication.
Taking the example from the first post in your adaptive content series, you were at a wine-tasting event, at a vineyard. That’s a very specific use-case, but one that is relevant for a vineyard. Those criteria impact how they would want to interact with you,
We’re talking about a shift in perception. This isn’t the old generic marketing segmentation of age, gender and socio-economic status. This is about context.
Noz: Yes. Socio-demographics are mattering a lot less for adaptive segmentation. What matters is what the user is doing, and what they want.
It comes down to understanding the user, from the user’s perspective. Why are they interacting with us?
When developing use-cases and scenarios, we have a tendency to frame the user in the window of our medium. We’ve said: I’m making a web site, so the user wants to do this or that with my web site. From the get-go, we boxed the user’s desires, their ambitions and goals, into our frame of reference.
Rick: Can you give me an example?
Noz: I’ve asked clients, about their customers, “How do they buy?” The clients would tell me “They go here; they click on that.” My response: “You don’t start a purchase with a click.” What is their motivation to buy in the first place? How did they get to the point of transaction? What research are they doing? What factors drive them to the site…?
Rick: Let’s go to the core: you start a purchase with a thought.
Noz: Exactly. What initiated that thought? When does that thought come to people? Those are the questions people fail to ask. And they’re basic sales questions.
I think the best practices from customer interactions in non-digital channels are finally making their way into the digital consciousness. We’re not just responsible for the digital channels any more. We’re responsible for the entire customer lifecycle – or at least providing digital integration to those who are. We’re learning to support the demands of those other channels.
Rick: I have to ask: how well are those other channels doing? Have they been negatively influenced by digital’s poor marketing mind set? Has digital had enough quick success that the other channels have started doing things our way, forgetting their better practices?
Noz: That’s an interesting question. Some of them, potentially yes.
Though I think that the bar for what counts as a decent sales team is fairly well set. Non-digital sales teams have been measured against different criteria than their digital counterparts. Some of our sloppiness may have rubbed off, but for the most part, it’s the digital bar that is being raised. We are going to have to address things that have long been standard for non-digital teams.
Rick: Good. We need a target to move towards.
Of course, this raises the interesting question of how we present these adaptive option to the author. We’ll cover the details of the interfaces themselves in our third session, but for now, let’s deal with the issue as you raised it in our first session – when we’re writing for adaptive content, as when we’re writing for more than one delivery channel, we must stop writing in the presentation format. This means we are writing multiple channels and scenarios at once.
Noz: And that’s quite a challenge. Even with great author experience tools, there is a conceptualisation problem. How do you write for everybody and nobody at once?
Rick: I agree. It’s difficult. I covered that in the book (p42): Thinking about thinking.
If I want to get a concept across to you (say this one I am explaining) I need to come up with a series of sentences that you can understand. I need them to be in a certain order. I need particular thoughts and ideas to be in the correct sequence within those sentences so they get into your head and you go: “ah, yes, that makes sense to me.”
Noz: OK, I like this stuff. We’re getting back to the fundamentals of communication.
Rick: Now, the way you receive this message, and the way you analyse it to understand it, is in its finished form. You can do something with the complete sentences. But how do I create those sentences?
On a basic level, I string words together. But how do I choose the words? How do I choose which synonyms, analogies and idioms, and whatever, I’m going to use, to create this story for you?
Noz: So what’s the answer?
Rick: How many people can actually think about how their mind works on that level?
Noz: Not many. But that’s kind of a dead end.
Rick: Yes and no. If we want computers to create hundreds or thousands of outputs for us, depending on segmentation, there’s no way we can write them all. And therefore, the computers need to think like that. They need to pull these bits together and actually build the sentences.
Noz: I see what you’re getting at.
Rick: That is some way off. But there is as interim step. Instead of the thousand options, let’s work with fifty scenarios. We don’t want to write the full script for each; we’re not that good. And maintaining consistency is too much work. We want to write the elements, the partial sentences, and pull those together to meet the demands of each scenario.
Noz: You’re reminding me of a bigger idea. Or maybe another way of wording it. We need to unload more of the brain work into modern content systems, not just the manual work.
Noz: For years, we unload the grunt work of punching numbers into a calculator and computers. Then we unloaded the grunt work of sorting and formatting, of copying and pasting. What we’re trying to do now, across the world of computing, is actually make them smarter.
So, whether we’re talking about search or query engines, Apple’s Siri, or IBM’s Watson, the whole industry is working to make computers more intelligent, in human terms. Effectively, a form of artificial intelligence.
Rick: You’re going to warn me, now, that we’re broaching a subject that will scare most people.
Noz: I am. People aren’t generally ready to talk about artificial intelligence. Because the general perception is of the Hollywood bill of goods: that one day, some genius company will show up with a fully-functional artificial intelligence.
But that’s not how these thing work. That’s not how technologies come to market. No technology has ever appeared fully-formed. The reality is gradual, over decades, with little puzzle-pieces of the concept coming together. Sometimes they’re useful; sometimes useless. Sometimes they leap forwards, but are so buggy that only the most dedicated geeks will try to wrangle them into shape.
Rick: And sometimes they take us backwards.
Noz: That too. The principle is incremental adoption. Siri is not a very good artificial intelligence. But it’s a stepping stone in the right direction. Google Now is a stepping stone in the right direction. It’s via these steps – one every few months – that we’ll eventually arrive at something most people would call artificial intelligence.
But all of these steps are forms of artificial intelligence – albeit basic ones. Once we have more robust artificial intelligences, we’ll be able to look back at the evolutionary complexity of the field, just as we can compare the evolutionary complexity of different life-forms, from amoeba, to jellyfish, to dogs, to humans.
Rick: So we already have artificial intelligence, but from an evolutionary completeness perspective, it’s somewhere between the amoeba and the jellyfish.
Noz: Exactly. In the content world, we can add semantics to the artificial intelligence mix, adding human meaning to our content so computers can put it back together for us.
Rick: And so they can take what we’ve written, mash it up, and reassemble it, creating adaptive output.
Noz: Yes. And that’s what we need now. We need to move that aspect of the workload to computers, get ourselves accustomed to it, and then we can move to more complex systems.
We need to think ahead, to the way these system will work when they are more evolved, and put in place the elements they will need to do the adaptive grunt-work for us. We need to embed a high level of semantics. And this is not only to make it work in the distant future; it’s the same data we need to manage and manipulate it with today’s targeting algorithms.
If we want our content to adapt, now and in future, we must add semantics.