My marketing career started at developer-focused companies like GitHub and Dremio.
Software developer culture had a lasting effect on me and how I think about my own work. It’s a systems mindset, and a useful one for marketers to consider as our work is increasingly integrated with automation and AI tools.
These days, at ércule, our clients primarily come from the developer tools, data infrastructure, and AI spaces. Many of their products are concerned with how to deliver software.
A lot of the problems that we’re grappling with now as marketers are problems that developers solved for themselves more than a decade ago.
So I’m going to highlight some lessons from DevOps and how we might incorporate them as marketers.
Overview: DevOps for newbies
DevOps is a framework that brings together software development (Dev) and IT operations (Ops) teams.
A shift in focus from products to production systems
DevOps signaled a shift to thinking in terms of production systems rather than individual products.
❌ The old way: maintaining individual servers and taking care of each one.
✅ The DevOps way: managing fleets of servers and containers and huge numbers of these things in a way that gets to reliability.
The analogy that is often used to explain this shift is: cattle not pets. You’re dealing with large fleets of things instead of very specific individual things.
You need to manage infrastructure at scale. And because the human resources at any company are finite, scaling up infrastructure like this requires a workflow that is much more collaborative.
Development is agile and continuous.
We want to continuously be updating the application that we have.
❌Old approach: there’s an individual product and I deployed it and I have to take care of it too.
✅DevOps approach: humans work together effectively to maintain and improve the systems that enable production.
DevOps is an organizational and cultural shift with very practical ends in mind: continuous integration and continuous delivery, smaller, faster updates, more frequent updates, etc.
We want new features to come out continuously. We want to ship 500 times a day.
It’s all in service of creating better products
When you implement this systemic approach, the result is not only more releases but better, more creative releases. Quality-assurance is scaled alongside production.
It’s not automating every technical task. It’s automating the rote tasks. In a DevOps model there are still people who are still writing code and thinking about interesting problems. But the point is: you're making your baseline a lot more efficient so that those people can be freed up to work on interesting things instead of managing very specific stuff.
From DevOps to content systems
Many of our ércule clients have thousands of blog posts on their website. The only way to think intelligently about a collection of assets at that scale is to think about groups and how those groups are managed.
So we're working with them in terms of whole libraries and we're thinking in terms of topics.
Yes, we care about individual pieces! But in order to create the most effective pieces, at scale, we’re thinking in terms of system.
The content library on your site is a system.
So we ask systemic questions when we’re trying to improve performance:
- What is the star content in our library?
- What is the declining content of the library?
- How do segments of the library work together?
When we look at how content is working, we’re less concerned with individual pieces of content and more concerned with groups of content that are aligned in a particular way – for example, all of the product-focused content on your site.
Automation practices for production
Automation is a major tenet of DevOps. Marketers can also integrate automation.
❌The old way: individual pieces of content being individually managed.
✅The new way: publishing quality content at scale.
Let’s look at a few automation tools we use.
FAQ finder
This is a workflow we frequently run for clients. The FAQ finder does web research to identify all the different questions that people are asking about a topic.
👆These are all “People Also Asked” queries from Google. That's why they're marked here as related questions. We also pull questions from Reddit, Gartner, user review sites, and social platforms.
Customer research is integrated in the DevOps approach to content
We’re using this automated workflow to collect questions that people have about your product, about your industry, about your topics, about what they're trying to do, what your persona is.
And we’re doing it on a much larger scale than ever before. Five years ago we were manually searching Reddit, spotting an interesting debate there, then (maybe) writing a blog post on the topic. Now, in a matter of minutes, we’ve collected this list with several hundred real questions that people have.
Ideally, we want to answer every question we can for the brand’s target audience. We want to be as useful as possible to the customer base. One nice byproduct of that ability: it’s great for brand awareness and visibility in SEO and LLM search.
Linting content like developers lint code
Our agency spends a lot of time building and working on AI workflows. Here's an example of a very simple AI workflow: the content linter.
The content linter goes through every single article that you have on your blog and or on your site. And it can do things like spellcheck the entire site. It might solve for problems like:
- What needs to be refreshed?
- Does this page’s messaging match up to date?
(This tool is adapted from the linters used by software developers. Their linters check for errors in code.)
So there's lots of AI workflows like that that can automate a big chunk of the work we do.
Automating production + maintenance
Those “What is…?” posts no longer require an intense hiring and production process. The ROI isn’t there. A workflow can help with it.
Content automation is much more practical now than it was a couple years ago. You still need quality-assurance, of course. You can bake that into a lot of these workflows.
- We're going to deliver with speed.
- We're going to deliver with quality.
These are devops goals that we can adapt to our content system maintenance.
Competitive matrix
This competitive matrix is basically a product marketing in a database. It contains information about a variety of competing products.
It contains fields for…
- Core features
- Integrations
- Support
- Pros & cons
This database is an input for a content generation workflow. We're feeding this product information into the systems that then actually do the work. The humans are in charge of setting the parameters, building the flows, and running QA on the output.
More time for marketing creativity + collaboration
Once those baseline pieces are published, the actual (human) creatives on staff can think about things like: this one piece of content is really our marquee piece, so what does it need to really shine?
Any given piece might need…
- Interview material.
- Drama.
- Interest.
- Cleared connection to our POV.
- Data analysis.
- Customer quotes.
All those things that only humans can do. And in a DevOps approach to content that's what the humans are doing. They're working on more interesting problems instead of maintaining lots of fiddly individual bits.
Moving beyond the content calendar
Continuous delivery and integration are key principles of DevOps. They’re now possible for marketers.
DevOps is about a shift in the way of working to a more collaborative, faster deployment, higher quality model that works on larger scales to deliver a reliable outcome. When you’re able to continuously publish and update content, a calendar becomes obsolete.
When we publish a piece that doesn't quite match our brand, or we need to update something from five years ago, there's a big AI component that takes care of that for us as part of the system which is part of what we build with clients.
Humans are working on the harder problems that require individuals to take a piece of their perspective and put them together.
The more you can think of in terms of the system, the better. (Frankly, all of marketing might be going the way of DevOps but don't tell anybody I said that.)
