Thanks for this email, Oliver, it's fantastic! Since I'm one of the people who says 'flat' and 'flatter' a lot, I feel compelled to respond, though I run the risk of painting an already-perfect lily.

One of the first essays we read in the Flat Org group was 'The Tyranny of Structurelessness'[1] which makes a similar point to Oliver's, and I think it's one that everyone is wise to remember. The question that I seek an answer to is not "How can we smash hierarchy?" It is, rather, "How can an institution be less reliant on the competence and benevolence of a small number of people, and less vulnerable to malice or incompetence on the part of a small number of people?" In my experience, traditional top-down management systems are highly vulnerable because they're great at magnifying whims and mistakes.

I'm pretty sure that it's possible to have structure without having a rigid power-based hierarchy. To some extent, that's what democracy is, or at least what it seeks to be. It's definitely what Wikipedia seeks to be. I hope that someday the WMF joins Wikipedia and the other projects in their weird adventure of anarchist/collectivist/self-organizing weirdness. Not because I want the Foundation to be governed more like Wikipedia, but because I want Wikipedia _and_ the Foundation to be governed better, and I think there are lessons we can learn together.

(Top-posting to prevent scroll-wheel-related RSI)


On 2/24/16 12:34 PM, Oliver Keyes wrote:
I would like to clarify a fairly major premise of this conversation:
namely, the comment I made that Yuri quoted in the very first message.

When I say that the hierarchical organisation of the Foundation is
something that is preventing us from doing better, I was not thinking
of how we develop software. Indeed, I suspect that peoples' tendency
to bring things constantly back to "does it improve the measurable
speed at which we right code" is symptomatic of the problematic
dynamics here. What I was thinking about was how we pay attention to
organisational hiring, to how we promote, to how we treat people, what
empathy we have and how we value empathy.

I have consistently found the Foundation to lag in all of these
regards. It is not good at making sure that the recognition of
employees is fair and treated equitably (be that who gets called out
in presentations, who gets given opportunities, or who gets raises).
It is not good at making sure that how we hire is fair. It is not good
at making sure that concerns of employees are given weight. All too
often the people marginalised by our approaches are the people
marginalised outside the Foundation, as well; women, people in
"non-technical" roles, people in roles that we code as "support work"
(and guess what tends to correlate with a role being coded as support
work?) All too often the work marginalised by our approaches is the
work that Doesn't Product Code (again: guess who tends to do the heavy
lifting on things like organisational health and process and

As an organisation I have found the Foundation overly rigid and
resistant to the most conservative change around these problems;
particularly I think of efforts to improve unintentional bias in our
job descriptions. Basically, unless you as an employee go out and do
the damn work yourself, for free, with 0 recognition of the emotional
and temporal cost of that work, it doesn't get done. The organisation
as a whole is not interested.

Switching to a flat organisational structure does not, in any way,
solve for this problem. In fact, in some way it makes it worse,
because it makes us *think* that we have solved for the systemic and
hierarchical power dynamics that make it difficult for low-level or
marginalised people to get things done, or people doing marginalised
work to get things done, when we have only shifted them.

To pick on someone, I pick Trevor (sorry Trevor. For reference this is
an entirely hypothetical example and Trevor is lovely): Trevor's voice
is given a lot more weight in the organisation than mine. Trevor has a
lot more influence than I do. Trevor has a lot more influence than
most WMFers do!

Crucially: this *isn't because he's management*. This was the case
even *before* he was management. Because:

1. He's been here a really really long time and so knows everyone.
2. He's an Engineer, and we give engineers more weight and cachet than
we do, say, administrative staff or people in "support" roles, even
though those people are both as-smart and have an equal interest in
the organisation's success;
3. His background matches what we strongly correlate with Authority Voices.

If we switch to a flat organisational structure where nobody has a
title, or..whatever, all of these things will still be true. We will
switch pronounce systemic biases or uneven power dynamics Done, and we
will have achieved something that's actually worse than not doing
anything at all. Because now, we still *have* all those problems, we
just think we're done and don't have to put any work in and can't talk
about it, and nobody has the responsibility for continuing to fix

The Foundation I would return to is not an organisation with a flat
structure. In fact, it could be an organisation that looks a lot like
this one, because I don't believe reporting lines or titles have as
much of an impact on dynamics as we think they do. What *does* have an
impact is how we recognise the value of emotional labour, how we
recognise our implicit biases and advantages, and how honest we are
with each other: not just in terms of what we *say* but in terms of
how we *listen*. In other words, the litmus test for me is: what
happens when the socially and politically weakest person in the
organisation has an idea?

Anyway; I don't particularly want to go into a long drawn-out
conversation, just correct the initial, fundamental misunderstanding.
Hopefully I've provided a bit of food for thought along with that.

On Wed, Feb 24, 2016 at 3:50 AM, Pau Giner <> wrote:
If I remember correctly, I think that's how the Content Translation project
started -- it was someone's personal project, which got more people and
attention because it's a great idea and showed real success.

That is not accurate. I think Content Translation is a good example of
bottom-up and design-driven project, though.

The Language team identified that users frequently were asking for better
support for translating Wikipedia articles, and decided to learn from
existing translators (and their heavily manual efforts) without a
predefined idea of how the solution would look like. After many iterations
of design, prototyping and research the team started building the tool
iteratively and driven by user behaviour (based on metrics and more user
research). I wrote a more detailed piece about this some time ago if anyone
is interested in more details:

So while this project didn't came from top-down roadmap, it was also not a
solo "cowboy-style" personal project.
I definitely think it followed a good pattern for more projects to consider.


On Wed, Feb 24, 2016 at 4:40 AM, Andrew Lih <> wrote:

On Tue, Feb 23, 2016 at 7:53 PM, Yuri Astrakhan <>

And that got me thinking. WMF, an organization that was built with the
and community-driven principles - why have we became the classic example
a corporate multi-level hierarchy? Should we mimic a living organism
than a human-built pyramid?

This may sound naive and wishful, but could we have a more flat and
flexible team structure, where instead of having large teams with
sub-teams, we would have small self-forming teams "by interest".  For
example, someone decides to dedicate their 20% to building support for
storing 3D models in wiki.

So glad to see this being discussed in the open with smart folks like
Brion, Dario, et al.

3D support would be most welcome – we’re in a holding pattern with
Smithsonian 3D GLAM projects in DC because of that shortcoming in Commons.
It was never known whether anyone was paying attention or going to put 3D
on the radar screen. (

At my keynote talk at Wikiconference USA, I said one of the things WMF
“must do” is multimedia and interactivity. Your work on the interactive
graph is a great step. Brion, myself and others have been working on
collaborative video. Brion’s ogv.js work is a great example of skunkworks
type projects having a huge impact.

And if 3D is given a priority, the three areas would be a great collective
step towards Wikipedia continuing its revolutionary work. Best of all, they
would be technologies developed in service of content and community needs,
and not simply created for tech’s sake.

An organism reacts to the change of its environment by redistributing
resources to the more problematic areas. Would small, flexible, and more
focused teams achieve that better?

Yes, and in a recent meeting you mentioned Bell Labs as a model. As someone
who worked there, it’s a very good ideal to shoot for.

Thanks for opening up this discussion.
Wikimedia-l mailing list, guidelines at:
New messages to:

Pau Giner
Senior User Experience Designer
Wikimedia Foundation
Wikimedia-l mailing list, guidelines at:
New messages to:
Wikimedia-l mailing list, guidelines at:
New messages to:

Wikimedia-l mailing list, guidelines at:
New messages to:

Reply via email to