Issue 26

Reader Q & A

As teased in Issue 25, I’m celebrating Issue 26 as the one-year anniversary of the Model Thinking newsletter. It’s hard for me to believe that the time has gone so quickly!

For this milestone, I put out a call for questions from readers, and I’m going to answer them in this week’s issue. Thanks to those who sent something. If you like this idea, I’m opening an ongoing form for reader questions that I’ll occasionally answer.

Managing internal knowledge

Both Brooks and Michael are essentially asking about how to manage internal knowledge—the knowledge that grows within organizations.

In many organizations, that knowledge is generated organically in sanctioned and unsanctioned tools. Few of the tools are richly connected to each other, and the content within them is rarely structured and often out of date.

It’s a gnarly problem space.

Interestingly, I saw a job posting earlier this week for an “internal content strategist” that would have a big role in trying to bring order to the chaos at some company. (For the life of me, I can’t find the listing again.)

In my opinion, organizations using file-based tools (OneDrive, Microsoft Office, Google Docs, spreadsheets, slide decks, and so on) are at a disadvantage. The knowledge in those systems is rarely inherently connected to other knowledge. Knowledge is not very findable or discoverable.

My preference would be to use systems like Confluence (it’s in my Atlassian alumnus DNA) or Notion where there’s a more inherent relationship between knowledge artifacts. (And I’ll be the first to say that Confluence isn’t perfect at this.)

My tips:

• Curate carefully. Retire or update content that is outdated. Glean (essentially an AI search tool) is very good at finding content in the systems it’s connected to, but it won’t know if what it finds is outdated.
• Use structure to help search tools find and understand your content. It’s no secret that I like structured content. Most of the workplace tools mentioned are pretty unstructured, but even using good heading styles helps. That’s not really new in the last 5-10 years.
• Move away from file-based systems as much as possible. You may not be able to completely get away from spreadsheets and slide decks, so consider embedding them in tools like Confluence along with context about them.
• Be open by default. Confluence makes all content available to anyone in an organization by default, but allows for content to be restricted if desired. The Google apps suite is the opposite in my experience. By default, only the creator has access to a file, so sharing knowledge requires an overt act of sharing.
• Hire content strategists or knowledge managers to shape and tend your internal knowledge systems.
• Beware the AI implosion Michael Iantosca warns of that I covered in Issue 23.

What is content strategy’s place?

The problem is that there’s not a clear understanding of content strategy and that there are too many people deciding on content strategy at many organizations.

I worked on a project where we developed the idea of “global” roles like content strategist and content architect who would oversee an org-wide approach while “local” roles might own their silo’s content strategy within the enterprise content strategy—what Colleen Jones would call content vision.

The question of when content strategists get brought in probably depends on what definition of content strategy the org favors and what their maturity is. Obviously, I’d say the sooner, the better, because honestly, a good content strategist is really a business strategist.

I see a little pick-up in organizations looking for content strategy expertise in this time of year, and other consultants I talk to indicate this is a bit of a pattern. Whether that’s because there’s the mental fresh start coming in January or if it’s because of fiscal budgeting cycles, I’m not sure.

For the content strategists out there, you might be headed into a window where you have time to breathe a bit. Consider how you might use that time to show, not tell stakeholders of the potential you can unleash.

Help with an AI style guide

I’m not really sure, to be honest. It sounds like you’re wanting to know how to approach the maintenance of a style guide related to AI. Is the style guide for how users should interact with AI? Or is it for how AI should operate in terms of brand voice, glossary, etc.?

My assumption is the latter, so I think you’re on the right track: audits, feedback loops, providing and maintaining quality source documentation (style guides, glossaries, etc.), clarity on what underlying models are used and for what purposes, how are updates to underlying models handled, who can train the model and who can’t, measurement criteria and if there are triggers for intervention based on the criteria.

If you’re staring this down, you’re probably further down the path than me. If any readers have thoughts, I’m happy to make the connection.

Bringing structure to marketing content

This is a meaty one. There seems to be a perception that marketing content has less structure or needs to be more fluid. To a degree, that’s true, but I also think there’s a chance for more structure in marketing content.

It’s probably helpful to think not of an entire deliverable but of the pieces that make up a deliverable.

I’d come at this from several angles.

1. From an ephemerality perspective
   ​Some marketing content is very ephemeral; it is very temporary. But other marketing content has a pretty good lifespan. Consider focusing your efforts to structure the more permanent marketing content.
2. From a reuse perspective
   ​If you look across your entire organization, do you see other silos creating the same—or nearly the same—content? No doubt you will. When you find duplicate content, look to bring structure there so everyone can share.
3. From a personalization perspective
   ​Personalization feels like a place where you have to rely on manually managing everything. In reality, the more structure and metadata you have, the better-equipped you are to drill down and provide personalized content. The book The Personalization Paradox: Why Companies Fail (and How To Succeed) at Delivering Personalized Experiences at Scale covers this really thoroughly.
4. From a scale perspective
   ​No doubt you’re expected to churn out more and more content with fewer and fewer resources. Having granular, structured content may enable you to scale your content fairly easily, even without AI. Of course, if AI is part of the solution or the mandate to generate more content, having an undergirding structure is going to be helpful. (But as a good content strategist, I’m contractually obligated to remind you that sometimes less is more.)

What does it take for Austin FC to improve the quality of its soccer?

Ah, the important stuff.

For those who don’t know, Austin FC is a 5-year-old team in MLS.

Matt asks about a striker, which is traditionally the player most expected to score goals for a team. The most MLS goals that a striker has scored for Austin in a season was back in 2022 when Maximiliano Urruti scored 9. By comparison, 5 players in Austin’s Western Conference scored twice that amount in the 2025 season (and there’s still playoffs happening).

This year, Austin had two strikers, Myrto Uzuni (6 goals) and Brandon Vázquez (5 goals in MLS, 9 total before a season-ending injury).

While all the strikers in the team’s history have missed good scoring chances, I think Uzuni and Vázquez are better players than previous strikers. They may not be the right strikers for Austin, but I think the biggest problem is that the team has not had a playmaker or what I call a “bias toward the goal” to get the ball to the strikers in situations where they have a high likelihood of scoring.

Owen Wolff is a 20-year-old midfielder who is arguably the best player on the team, leading the team in scoring this year with 7 goals. The combination of his youthful age and his year-over-year talent growth make him a very valuable player.

The question for the team is if that value is best converted to dollars by selling him to another team or if his value is best retained to build a team around his abilities. MLS has a bunch of complicated rules around how teams build their rosters, but the short answer is that I think they should hold on to Wolff and sell his midfield colleague Dani Pereira instead.

John Collins

Thanks for reading! Did someone forward you this email? Subscribe here​ If you’re already a subscriber and you found value in something here, tell your friends and colleagues to subscribe!

Welcome to the 1 new subscriber who joined us since the last issue of Model Thinking.