Why pixel perfect isn't so perfect

56
Why pixel perfect isn’t so perfect How to scale agile

description

This is a case study I presented at the HIVE 2014 conference in Seattle about how Yammer, by giving up micro-control enabled their designers and engineers to work directly and collaboratively, ultimately creating an effective workplace.

Transcript of Why pixel perfect isn't so perfect

Page 1: Why pixel perfect isn't so perfect

Why pixel perfect isn’t so perfectHow to scale agile

Page 2: Why pixel perfect isn't so perfect

9%

Hi. I’m Jenny.

Page 3: Why pixel perfect isn't so perfect

9%

I’m a UI Designer at Yammer

Page 4: Why pixel perfect isn't so perfect

9%

Page 5: Why pixel perfect isn't so perfect

This is what we do at Yammer.

9%

Page 6: Why pixel perfect isn't so perfect

Yammer is the enterprise social network that connects coworkers.

9%

Page 7: Why pixel perfect isn't so perfect

We work across platforms. On the go, everywhere.

9%

Page 8: Why pixel perfect isn't so perfect

We started small….

9%

Page 9: Why pixel perfect isn't so perfect

But quickly scaled up.

9%

Page 10: Why pixel perfect isn't so perfect

And then we scaled some more…

9%

Page 11: Why pixel perfect isn't so perfect

What was the problem?

Page 12: Why pixel perfect isn't so perfect

What was the problem?

We were slowing down…

Page 13: Why pixel perfect isn't so perfect

We were slowing down…

9%What happened

Page 14: Why pixel perfect isn't so perfect

We were slowing down…

9%What happened

• LOTS of documentation

Page 15: Why pixel perfect isn't so perfect

We were slowing down…

9%What happened

• LOTS of documentation

• Slow to staff projects

Page 16: Why pixel perfect isn't so perfect

We were slowing down…

9%What happened

• LOTS of documentation

• Slow to staff projects

• Slow to ship

Page 17: Why pixel perfect isn't so perfect

We were slowing down…

9%What happened

• LOTS of documentation

• Slow to staff projects

• Slow to ship

• Lack of enthusiasm and ownership

Page 18: Why pixel perfect isn't so perfect

Why was this happening???

Page 19: Why pixel perfect isn't so perfect

Why was this happening???

Functional silos

Page 20: Why pixel perfect isn't so perfect

Anatomy of functional silos:

Project Managers

Designers

Developers

Page 21: Why pixel perfect isn't so perfect

Anatomy of functional silos:

broken link

Project Managers

Designers

Developers

Page 22: Why pixel perfect isn't so perfect

You know you’re in a functional silo when:

9%

• Designing is painfully slow because it has to be signed off by 10 different people.

Page 23: Why pixel perfect isn't so perfect

You know you’re in a functional silo when:

9%

• Designing is painfully slow because it has to be signed off by 10 different people.

• Design has no real seat at the table when discussing business strategy.

Page 24: Why pixel perfect isn't so perfect

You know you’re in a functional silo when:

9%

• Designing is painfully slow because it has to be signed off by 10 different people.

• Design has no real seat at the table when discussing business strategy.

• Design details are consistently cut back due to “scope creep”.

Page 25: Why pixel perfect isn't so perfect

You know you’re in a functional silo when:

9%

• Designing is painfully slow because it has to be signed off by 10 different people.

• Design has no real seat at the table when discussing business strategy.

• Design details are consistently cut back due to “scope creep”.

• Teams spend more time defending turf than collaborating.

Page 26: Why pixel perfect isn't so perfect

Implementation bottlenecks

9%

Developer

PM Designer

Page 27: Why pixel perfect isn't so perfect

Implementation bottlenecks

9%

PM & Designers are slowed down making specs

Developer

PM Designer

Page 28: Why pixel perfect isn't so perfect

Implementation bottlenecks

9%

PM & Designers are slowed down making specs

Confused coders—decoding specs

Developer

PM Designer

Page 29: Why pixel perfect isn't so perfect

Implementation bottlenecks

9%

PM & Designers are slowed down making specs

Confused coders—decoding specs

Frustration when a spec goes nowhere

Developer

PM Designer

Page 30: Why pixel perfect isn't so perfect

Anatomy of functional silos:

9%

specsmocks

Project Managers

Designers

Developers

Page 31: Why pixel perfect isn't so perfect

Common result=meetings

9%

Battle over who owns who’s time

Page 32: Why pixel perfect isn't so perfect

All these elements make fertile grounds for lack of trust between teams.

9%

Page 33: Why pixel perfect isn't so perfect

So we tried something different…

Page 34: Why pixel perfect isn't so perfect

So we tried something different…

Decentralized decision making.

Page 35: Why pixel perfect isn't so perfect

Before

Project Managers

Designers

Developers

broken link

Page 36: Why pixel perfect isn't so perfect

After Project ManagersDesigner

sDevelopers

AutonomousTeam

Page 37: Why pixel perfect isn't so perfect

Each team was autonomousBusiness strategy

implementation

Autonomous Team

Page 38: Why pixel perfect isn't so perfect

Went from having one big ship to a fleet of smaller, faster ships.

Page 39: Why pixel perfect isn't so perfect

The results?

Page 40: Why pixel perfect isn't so perfect

The results?

Business strategy

implementation

Autonomous Team

Fewer top down decisions.

Page 41: Why pixel perfect isn't so perfect

The results?Sitting together fostered organic sharing and involvement.

Page 42: Why pixel perfect isn't so perfect

The results?Sitting together fostered organic sharing and involvement.

Page 43: Why pixel perfect isn't so perfect

The results?Sitting together fostered organic sharing and involvement.

Page 44: Why pixel perfect isn't so perfect

The results?Got rid of a lot of bottlenecks

9%

specsmocks

Project Managers

Designers

Developers

broken link

Page 45: Why pixel perfect isn't so perfect

Beforespecs & mocks

The results?Got rid of a lot of bottlenecks

Afterspecs & mocks

9%

Page 46: Why pixel perfect isn't so perfect

The results?Balanced input

the product

Project Managers

Designers

Developers

Page 47: Why pixel perfect isn't so perfect

The results?Trust feedback cycle between designers and engineers.

Developer reigns in any unrealistic designs.

Designer helps Developers code the details.

Page 48: Why pixel perfect isn't so perfect

The results?Fewer meetings.

Page 49: Why pixel perfect isn't so perfect

The results?Faster ship dates.

Shipped a lightweight chat app in only 3 months.

Page 50: Why pixel perfect isn't so perfect

The results?

We had fun!

Page 51: Why pixel perfect isn't so perfect

What you can do in your company tomorrow?

Page 52: Why pixel perfect isn't so perfect

What you can do in your company tomorrow:

#1 Cross functional seating

Page 53: Why pixel perfect isn't so perfect

What you can do in your company tomorrow:

#1 Cross functional seating -Try moving your designers and engineers closer together.

-“war rooms”

Page 54: Why pixel perfect isn't so perfect

What you can do in your company tomorrow:

#1 Cross functional seating

#2 Decentralizing decision making

-Try moving your designers and engineers closer together.

-“war rooms”

Page 55: Why pixel perfect isn't so perfect

What you can do in your company tomorrow:

#1 Cross functional seating

#2 Decentralizing decision making

-Try moving your designers and engineers closer together.

-Pick a small, low-risk project and try writing the spec as more of an outline, letting the team fill out details as they go.

-“war rooms”

Page 56: Why pixel perfect isn't so perfect

Questions?

Jenny Mullins@[email protected]