Content Visionaries: Content Technology Insights from Heretto

This interview is part of our Content Visionaries series, asking content leaders across industries for their insights into findings from our 2021 State of Content Operations Study.   

The growing field of content operations brings together people, process, and technology to manage content effectively and efficiently.

In this interview, Patrick Bosek, CEO and co-founder of content operations platform Heretto, gives his views on how this space has evolved, the challenges facing organizations, the state of AI and personalization, and how best to take advantage of the technology available today.

As a leader in the content technology industry, what major changes have you seen in this space in recent years?

BOSEK: This pandemic has been the most disruptive event of our lifetimes, but it’s also probably been one of the more disruptive events in the content industry, especially as it relates to knowledge technologies. So, we’ve seen through many years in this space a slow march towards content technologies, approaches, and information architecture. Over the course of the last couple of years, the technical ecosystems and then the practices to leverage those – to operate them and to create them – frankly have all started to mature. I think some of that is supply-based and I think some of that is demand-based.

I also think that there has been this realization that in 2016 we were still in a person-first world. The reality is that if you looked at the interactions, buying experiences, support experiences, you were still largely dealing with humans. That was the primary mode of customer interaction with brands. 

Recently we’ve had a hard shift toward a digital-first world. What that means is that the ways that we do business, which is also to say the ways that we interact with each other in anything other than personal interaction, has become different. It’s become digital. Most of it is content.  Typically you’re interacting with a system, and the question is whether or not that software has any asynchronous content that is between you and a human being responding directly. I think that we’re starting to see that is becoming much more prevalent. It’s becoming more expected.

People think, ‘it’s the internet, it’s my device’. Well, it’s not digital-first. It means that your communication goes through software, as opposed to going through to a direct team. Because everybody wants their stuff faster, you have to have better, more personalized content.

It’s a tectonic shift in the way that we interact with each other and content is at the center of it. 

What are the biggest barriers facing organizations today?

BOSEK: The barriers today are broadly the same barriers they have been for the last 10 years, but progressively less throughout that time. 

The first barrier is the fact that an organization has to commit to it and that barrier is actually going down. We’re starting to see titles and practices and centers of excellence and professionals, and leadership, groups formed, cross-department collaboration, enterprise-wide strategy, all these kinds of things as the business function starts to mature.

The second barrier is technical.  ‘The right content to the right customer at the right time’ has been around for quite some time. I don’t like that particular slogan, and the reason I don’t like it is because I could hand you a book and I could say, ‘this is the right content’ and I could be right about that. That has been the practice for many, many years. We might think what we need to do is to get them the PDF they’re looking for, and that PDF could be a thousand pages. But you haven’t really accomplished the mission. The mission is to educate at the point of need, to answer the question, to provide the piece of data – to transfer just the piece of information the person needs when they need it – which is slightly different. 

When you take that slightly different lens on things, not only do you need different practices to put that into place, but you also need different technology. The technology that accomplishes those two things can be the same, but in practice, they typically aren’t. We’re trying to answer the question, we’re trying to provide the instruction, we’re trying to give just the measurement. 

The technology is still a barrier, but I think it’s more an aspect of having the organizational will to implement that technology at this point in time. 

If you’re going to do it at enterprise scale, unfortunately, even today, the technology isn’t simple. Don’t get me wrong, it’s become much more user-friendly, but it’s not commoditized. It’s still real, true enterprise stuff. We’re moving to a place where this stuff is available to all organizations, but to really do it at an enterprise level, you still have to go and embark on an enterprise project and that’s not a small thing. The technology is expensive. The teams to implement the technology, and then weave together the different things that are going to actually cause the experience that you’re looking for, they’re expensive. So you have to have real leadership buy-in that has a pretty decent budget and commit to something that’s not going to return stuff overnight. 

How important is content personalization today?

BOSEK: I’m not going to say the thing that everybody expects me to say, which is it’s super important, and it’s the only thing you should care about.  I don’t actually believe that. I think content personalization is something that is different for every organization. The way you implement it and the importance that you place on it really changes organization by organization.

Let me give you an example of that. So as it relates to our company, we do content personalization, but we don’t do a lot of it because we don’t need a lot of it. I should caveat that by saying we’re going to do a lot more of it next year. 

We have materials that go both to our customers and to our employees, and those are personalized. There’s information that is stripped out when it goes to customers. Then we also have some materials that are personalized for different deliverables that will go to different customer types that have different educational resources. But broadly speaking, that’s as far as we’ve gone with personalization.

Whereas our customers will very often filter website content in real time based on who you are. We don’t need that because the piece of software that we sell is, broadly speaking, pretty consistent for everybody who uses it. There are two versions: one version does more than the other one, but most of the stuff is pretty logically separated. We’d like people to see all that stuff. Maybe then they’ll buy the bigger version, and we don’t really want to filter that out. Despite the fact it’s useful in certain circumstances, it’s not an organizational imperative like in a really advanced system.

But some of our customers have things they can only show to certain people in certain geographies. Or their software systems are so complex with so many configurations and so many different variations that if they just gave out one version to their customers, it would either be so difficult to parse because it would never be able to find answers, or it would be so generic that it wouldn’t be helpful. 

So for them, content personalization becomes an absolute imperative. If your customer either can’t get the specifics of their situation or they have to mentally parse out a large portion of your content in order to get to that, then you have to personalize. Otherwise, they can’t actually access the things that you’re creating. Then obviously there’s regulation, there’s regional things, and all that stuff really matters too.

There’s a lot of different cases where it really makes sense, but that doesn’t mean it’s a true necessity in business for all organizations. It’s extremely important for some and less important for others. 

What are companies that are most successful in providing content personalization doing right?

BOSEK: The number one thing they’re doing right is they’re taking the right approach to their implementations. Go and do your homework and make sure that you’ve got good help coming into this, because otherwise you’re going to waste a lot of time and money. If you want to go fast, you gotta start slow. That is 100% how this works in this industry every single time. If you don’t take the time to really outline your goals and plan your content, then go and plan your content systems to match the capabilities you want from your content – you’re going to fail or you’re going to waste a lot of time. 

If you’re looking for knowledge-content systems – that is the learning, reference, product answers, support, content, self-service, policy, procedure, all that kind of stuff… the tried and true path is to:

• Select your content standard, which realistically means just go use DITA because there just really isn’t anything competitive at this point in time.
• Find somebody who’s an expert in DITA, if you don’t have one on your team. And if you do have one on your team, maybe still find one because they probably also have another job that they’re trying to do, and they don’t have time to work on your DITA project unless that’s literally what they were hired to do. Get some outside help.
• Work through your objectives as it relates to your content architecture and then select your tools.
• Design the content operation infrastructure that you want, that is going to provide the ability to put the content where you want to put it, and support your information architecture.
• Then implement that and ensure you can make all the connections that need to be made. 

That’s the pattern that everybody follows, that you see successful companies doing. It’s not an easy line, but it’s a straight line with a lot of work. It works every time and you get good results and then it’s iterative after that.

Almost every unforeseen business case when you’re implementing content systems is resolvable with developers and APIs. If you have an API to access your content and you have access to developers to leverage those APIs, you can solve the problem almost every time. Making sure that you plan for that upfront is really important. That’s basically what successful organizations do and it works every time.

How should teams approach structuring content to take advantage of technology?

BOSEK: Use DITA if you are looking for the forms of content I have been talking about. It’s not necessarily an everything solution, but none of the content technologies are. There’s really three major approaches to a content ecosystem or a content ops infrastructure today. 

• Digitally modeled content (headless CMS) is really good in certain circumstances. It’s very good when what you want to do is create a digital model of your business or some aspects of your business, and then you want the ability to use connections between different models of things to represent content experiences. There’s a talk that describes using digitally modeled content for personalizing content snippets inside of games at EA. That’s a really good case because you’re taking discrete pieces of content and displaying them to the user in a very personalized way. But the key there is that when you go digital modeled, the whole piece of content is selected based on the attributes of the user and it’s provided to them. The digital model works really well for anything in bite-size snippets, totally standalone. 
• DITA works really well for anything which is naturally longer form, but could also be delivered in a shorter form. So a tutorial might come in a bunch of flavors that get filtered, changed, modified, and rearranged, but you still want a tutorial. A course is the same thing, a user manual, a set of policies and procedures. You don’t want these to be just little snippets; they’re not going to be programmatically arranged onto a page based on your preferences. They have flow and context, which is why DITA is a really good format for those forms of content. When you go with something like XML, that content is condensed, changed, modified, filtered, and then provided to the end user. 
• Markdown has a lot of problems that don’t become evident until you get further into it, but can work for some organizations. I can probably count on one hand the number of times I’ve told an organization that they should really look at markdown as part of the main portion of their content ops infrastructure. It’s very select.

What role do you see artificial intelligence playing in content production and delivery in the near future?

BOSEK: It’s a tool and I think everybody wants it to be more of a panacea and it’s not. It’s an effective tool for certain circumstances. If you can get AI systems that will effectively tag content at a very high confidence level, that could be really important. If you can get to AI systems that can effectively take certain forms of consistent content and then convert them reliably to something which is more structured, that could also be really valuable. 

One of the things that still blows my mind about Amazon is that, frankly, the filters aren’t that good. Wayfair’s filters for furniture are fantastic. So you go in there and you’re like, ‘I want a bed that is this big and this color and this style’. But you can’t really do that on Amazon yet, although it is slowly getting a little bit better. And the reason is that they can’t go through and structure the content for the billions or however many products they have, whereas Wayfair has a much smaller set. 

It’s humans behind that stuff. Every time you see something that doesn’t scale up to a certain level, that’s because the automation isn’t there. So if we were able to effectively take the information that’s coming in from all those suppliers at Amazon, and structure it in such a way where it was all perfectly normalized for the different ways they slice those product searches, we would have far better filter capabilities. We would have a better experience, like what we get on Wayfair. But they haven’t done it yet. They’re one of the biggest companies in the world and they haven’t gotten it right. They can’t get this right with AI. And if you don’t think that they’ve got a bigger budget than you’ve got for AI… 

There are a lot of limitations in what we’re going to get out of AI today. So when you approach anything that relates to AI, you should really think through the state of the technology and what you’re going to get out of it. What is your real business objective? Think of AI as a tool; validate that tool does what it says it’s going to do. 

Are there any other exciting technologies you have your eyes on?

BOSEK:  One of the things that’s interesting, and I don’t think that we’ve gotten right entirely yet but I think we’re getting closer to, is that we haven’t figured out the right way to have a bidirectional conversation in the new digital world. I think the challenge with that is the technologies that exist to do that today are either too restrictive on the incoming communication from the customer or too unrestrictive. When it’s too unrestrictive, it becomes a burden on the organization to manage it; to police everything that comes through. All content that comes through into your organization creates overheads, creates cost, or it’s ignored. 

Every organization is trying to pick technologies that change the size of that pipe. Are we going to be able to get to a pipe that is the right size and adjust for the organization, but also has enough automation in it that it brings the cost of managing incoming communication down? 

Maybe this is where AI can play a role because AI is good at classifying things. That’s really the only thing that it does astoundingly well: classify and route, filter and sort content in such a way where an organization can handle efficiently and cost-effectively a larger quantity of incoming customer communication. 

Outside of just support, I think it’s going to be really interesting when we get to a point where our side of the conversation is done asynchronously through content, and the customer side of the communication is done asynchronously through feedback, questions, thoughts on that content. And it’s done in such a way where it’s comfortable for both sides and productive for both sides. That’s where we’re moving to, but I don’t think we’re there yet.