Centralized vs Decentralized UX orgs… over 20 years ago

In Org Design for Design Orgs, an entire chapter (Ch. 4, The Centralized Partnership) addresses organization models for design teams, specifically centralized, decentralized and embedded, and the Centralized Partnership, our preferred hybrid approach.

After I spoke at the Big Design Conference in Dallas last week, an attendee, Randolph Bias, pointed me to a paper he co-wrote in 1995 (“Usability Support Inside and Out”, PDF) about about whether or not usability engineering teams should be centralized or embedded. It turns out our book echoes many of the themes in this 20-year-old paper…

On the benefits of centralized:

Regarding the objective of maintaining usability engineering expertise, one primary advantage of the centralized model is it facilitates maintaining a human factors “critical mass.” The “care-and-feeding” of the human factors professional is very likely to be monitored, and good, under such a model. Relatedly, it is easier to hire new human factors professionals into such an organization, and the department manager is likely to know how to evaluate the professional’s contribution.

And the benefits of “mainstreamed” (their term for embedded):

The beauty of being a mainstreamed human factors professional is being, and being seen as, a team member. The clear buy-in of someone in the development group per se helps with communications to and from the rest of the product developers.

Plus ça change, plus c’est la même chose.

Glowing unsolicited feedback about “Org Design for Design Orgs”

In writing Org Design for Design Orgs, Kristin and I had a sense of the kind of impact it could have, in that it addresses a gap in the current conversation around design, specifically on organizational, operational, and managerial issues. And while we consider this topic vital, we weren’t sure how readers would respond.

So, color us thrilled when this unsolicited feedback come our way a couple days ago. Responses like this that affirm our toil was worth it, and suggests the kinds of value you might get from the book.

Thank you so much for writing this book! The minute I started reading it, I couldn’t tear my eyes from it. It was almost 2am when I found myself in the middle of chapter 7, and had to really talk myself into getting some sleep that night.

Your book is so rich with personable, practical substance on how build design culture, that I felt like I was in a cozy fireside chat with a design mentor. It’s like I was listening to stories from someone who has lived through it all. Someone who could relate and empathize deeply with the issues I’m facing with my design organization now, and give smart, actionable advice on how to proceed forward.

I’m a designer and researcher at a startup that grew from 25 to 50 people over the course of 2 years. Our design team suffered through the “drawbacks of centralization” (your description of the “us vs them attitude” and designers rolling their eyes was uncannily accurate), and are now going through the “decentralized and embedded” setup. We haven’t lived in this model long enough to hit the drawbacks, but reading your book felt like it was a warning for what’s to come. Then reading about the “centralized partnership” is where I got seriously excited, and made me itch to propose these changes to my team immediately.

Our design team has been through so many ups and downs. Struggling with getting design to be recognized, getting research to be valued and integrated, designers hitting career ruts and stunted growth, issues with aligning with leadership. This book covered all these issues in well organized manner, supported by clear diagrams and real world stories.

This truly is the book I wish I had since the beginning. The book I would reference with, write proposals and plans with, wallow in sorrow with, get uplifted with hope with. It confronts the reality of being an in-house designer straight on. I always have it with me when I’m at work, showing highlighted sections to my design manager, to product managers and engineering leads I work with.

It’s not only a book that I know will be useful throughout my career in design, it’s the greatest design mentor I could ever ask for, always in my back pocket.

Thank you so much for writing this book.


Levels framework: like Lebowski’s rug for your design org

From Chapter 7 of Org Design for Design Orgs:

Many companies use a framework of levels to chart the seniority of employees. Typically, human resources (HR) teams use levels to calibrate employees across different functions, to make things easier in matters such as compensation. The risk of working with levels is adopting a bureaucratic stance, seeing team members not as people, but as resources within a certain band of experience. Do not let levels define the team. Instead, use levels from the perspective of the team members, who are eager to understand how they can grow and evolve in their careers. Done right, levels are the scaffolding that helps team members elevate.

In the chapter, we share a 5-stage levels framework. Because of printing constraints, we couldn’t put the entire framework on a single legible page, so here it is available for download (PDFformatted for 11 x 17, aka Tabloid, and should work in A3 for our friends outside the US).


I’m about to tread in some deeply organizationally geeky areas here, so be warned.

Since writing the book, and consulting with companies on design org matters, I’ve realized that a solid levels framework acts as a ‘core service,’ plugging into other aspects of running the org. Without this framework, design orgs flail, particularly those with more than 20 members.

Performance review and professional development

Most obviously, a strong levels framework assists performance review processes. As team members and managers work together to chart professional growth, having a clear understanding of expectations, and what it means to be promoted, proves essential. And to make sure that this is fair across the whole team, it’s necessary to make these expectations explicit and public, so everyone can see where they fit.

Many design managers are cavalier about levels, operating as if they’re not that important. If a team member wants to be called “Senior,” well, what’s the harm in that? As someone who has inherited teams over-leveled people, the harm is that the team member is not being honestly supported in their professional growth. Accelerating people through through levels before they are ready is a set up for failure. And if someone like me comes in and tries to rectify a chaotic situation with a clear framework, these team members feel frustrated as they are inhibited from further promotion until they appropriately shore up their experience and authority.


In the Levels framework we developed, progress for designers is tied to skills building. This feels appropriate for a role that is about practice and craft. As stated at the beginning of this post, it’s important to approach your levels framework from the perspective of a team member, particularly one earlier in their career who is figuring out how to grow. This can then tie into a more formal professional development or training program, as designers take classes, attend conference, read books, and practice, either deepening existing skills or acquiring new ones.

Team creation

A levels framework can assist with team creation. You will likely want to have a mix of junior and senior people, including someone who can serve as a strong lead. And through the skills-building previously mentioned, you’ll understand how can do what, and how to match them in a complementary way. If you find a skills or leadership deficit, that can inform…

Recruiting and Hiring

When opening a requisition for a new team member, how do you know the level of seniority you’re looking for? Many teams either guess, or they don’t bother specifying.

More importantly, how do you judge the seniority of a candidate? You can’t rely on their job title, as those are notoriously subject to inflation. Without a levels framework, this assessment is done relative to those who are already in the org – if the candidate seems to be about as senior as Jane, and Jane is a Lead Product Designer, than the candidate is brought in at that level.

However, in large design orgs comprised of many teams, that relative ranking becomes dodgy. It might not be apparent that one team’s senior-level candidate might be another team’s mid-level, and if that person moves between teams, it can be awkward. A clear, shared levels framework prevents this confusion.


This gets us back to compensation. Without a clear and shared understanding of levels, an organization may have two people doing the same thing but ranked differently, or two people ranked the same but with differing expectations. Since compensation is typically tied to levels, team members, particularly those who are expected to do more, but are earning less, will feel the unfairness. A clarified levels structure makes sure that folks are appropriately  ranked, and helps a company’s compensation consultants do better benchmarking when comparing a company with the broader market.

Ties it all together…

As I’ve engaged with more and more companies, I’ve found that a Levels Framework is like the rug in the Big Lebowski


Without it, all these important operations are untethered, with arbitrary decisions made in the moment that, when combined lead to an incoherent mess. Levels work as a ‘core service’ that ties together these elements into a cohesive whole.

Design orgs get fucked over by relative team size

Typically, design orgs are significantly smaller than the engineering orgs they support. Designer:developer ratios are usually between 1:4 and 1:10. In the diagram here, we’re looking at 1:6.


In the book, we discuss the benefit of this ratio—design is a highly leveraged function, where a relatively small number of designers can have an outsized impact.

There’s a dark side to this though, in terms of perceived authority and influence.

In this example, the figure in black is the design Team Lead, overseeing the work of 5 others.

On the engineering side, that gray-ish figure oversees 30 engineers working with those designers.

Companies typically assign levels and seniority based on the number of people in someone’s organization. And this is where design orgs get fucked. With 5 people in their org, the Team Lead would be considered a Manager, but the Engineering Lead, overseeing the work of 30, would be considered a Director, maybe even a Sr. Director.

And so, while the Team Lead and Director of Engineering are peers when it comes to matters of scope, most companies would invest greater authority with the role overseeing more people.

As long as this persists, design will remain the kid sibling of other larger functions, and have trouble being seen as a true peer.

To address this, we need companies to rethink how they assign seniority, moving away from “how many people report up through you?” and towards “what is your scope of influence on the service we deliver?”