a

GOTO 2019 • How Thinking Small is Changing Software Development Big Time • Sander Hoogendoorn

[Music]

it's my microphone can you everybody

hear me because even with the windows

open right so I I get it could you maybe

I know it's I know it's warm I'm sorry

about it but if I can't hear myself talk

it's okay

well good afternoon everybody this is

going to be because it's a tough day

it's warm in here it's warmer than usual

in the Netherlands I am from the run so

I know it's usually rains around this

time of year so they say they may be

slightly more lightweight torque than

the other talks you've seen so far maybe

I'm not sure it depends on what you saw

so just my show fan who of you is in a

let's say an agile project or in an

agile team or that's less than I expect

I would say I would like expect like 99

percent to be who have used in a scrum

team who you say certified or a

professional scrum master I mean you

even dare to admit it that's pretty cool

okay so let's start so this is basically

a talk so so I've I've been doing agile

for over 20 years now

so I'm fed up with stand-up meetings

it's nothing they're not good I just

don't like them anymore

the same goes for retrospective the same

goes for lots of these ceremonies that

we tend to have in agile and especially

his current projects so to make it clear

I'm not against scrum I did this talk a

couple of days ago in in Belgrade and

the talker performing was Jeff

Sutherland so I waited for him until he

left the room before I started talking

so I might say big things some things

about scrum dough so yeah my microphones

off so anyway here we go so this is me I

am sundar I'm basically a dad which is

my most important occupation it's in the

end basically biochip ian to start

because they have three kids there are

23 and 19 and 14 two of them live sort

of full-time with me one of them is a

drummer I live in a small house in

Utrecht so having a drummer on board is

not the most best way to make friends

with your neighbors and in my spare time

I'm basically well I'm basically program

I've been writing code since I was 14 so

it's about 62 years now or something

I do agile stuff and I usually take on

CTO chief architect roles at companies

so does my current role is being the

chief architect for a company called

Cuba and Cuba is an IOT company they

make tong the thermostat that an ACO has

and they do lots of smart stuff with

their data that they collect for

instance they can predict when your

boiler breaks down or when you turn on

your washing machine and it's not

efficient stuff like that right those

are the services they deliver by the way

this is my website my Twitter attack my

email address if you have any further

questions what do I else do I I speak a

bit travel around the world a bit I

write a bit I'm actually writing two new

books one on micro services which is a

much more technical book and one about

the subjects that will touch today oh

dear there's even more people coming in

now it's get really really warm right so

as you all know I mean as you all seen

today is that the times are changing

right now the thing is Bob Dylan set

this in 1964 right and that's a long

time ago

time has changed a lot since then and

the thing is it changes faster and

faster and faster I'll show you some

examples of that for instance if you

look at Moore's Law you might think this

is a linear line right Moore's law says

that the number of transistors in a

dense integrated circuits doubles

approximately every two years this is

not a linear line this is an exponential

line right that means the possibilities

we have as technologists as developers

they grow and grow and grow and grow

right now let's say you would order if

you would order a computer in 1954 let's

say at Amazon it had looked a bit like

this right this computer has less power

than my mobile phone today it looks like

a big piano but it's a computer actually

now when I start a programming we had

computers like this does anybody still

recognize these or you're all younger

than I am right so I'm probably the

oldest guy but except for you maybe but

outside of that I'm probably the oldest

guy in the audience that's trapped

basically

I'm only 52 guys this is an IBM 51 what

really so what's your birthday first of

August and I'm older right I'm from the

seven

generally so see anybody older cool so

you know you recognize this one right

it's an IBM 5150 the black thing there

is basically where you put in the floppy

disks well you've never seen the floppy

disk I know they look like the Save icon

but they're actually there to store data

I could right go in this thing at home

right basically I started out with a

Commodore 64 having 36 K bytes free to

write code that's how I started but this

was a big change because it allowed us

not on the five computers that we had in

a country but at home or in our own

department to write code and then

basically everything changed again sort

of like 2006 2007 when Amazon or later

Google and Microsoft came out with this

cloud thing right now this cloud thing

changed everything basically because

what it did it allowed us to write code

or store data on somebody else's

computer we did not have to take care of

me more about these computers where we

ran our software and where our data was

and that changed things big-time right

I'll show you some things for that so it

basically disrupted everything and it

keeps on disrupting everything and as a

result we need to start doing things

differently now let me give you some

example do you know the time it takes

for technologies to reach 50 million

users maybe you've seen this one before

like airplanes for instance took sixty

eight years before 50 million people

flew an airplane television took like 22

years I started out with a black and

white television at my parents place

right

it had three channels well originally it

had even two right we didn't get any

foreign channels there was no Netflix it

didn't exist and from there it got

shorter and shorter and shorter like

Twitter took two years to get the 50

million users of course they're now back

at like 10,000 users but what active

users that is except for the books but

if you look at modern technologies

Pokemon go to nineteen days right how is

that possible

well basically because we all have

computers in

pockets and our kids have computers in

their pockets as well so the adoption

rate of technology inspired by the cloud

basically because it's running elsewhere

the things you can do

it's humongous right if you look at

Netflix for instance Netflix is a fairly

big company right

they have like 150 million subscribers

they have no data center they have

around ten ups engineers ten right can

you imagine Netflix having only ten

I think that's Cuba you already have ten

right now we're a smaller company than

Netflix while most of us are about

probably right so it changes fast and

you can see that in any industry to give

you an example we'll start in the

FinTech industry this is the FinTech

landscape of startups in the Netherlands

in 2018

there are 430 companies in this picture

right what if one of them decides let's

build an online bank let's deploy it on

Amazon and we'll use a dynamo DB where

rights and nodejs stuff will write a

front end in react done can you do that

become a bank yes you can write this

with a small team let's say five people

and about a year you're basically there

well all you need to get is a permit

that's harder but anyway the rest of

writing code is easy basically right my

girlfriend works for company does a guy

on his own riding an online bank it took

him about a year and he's almost done

including blockchain and everything you

can think of why he's only written the

mid office by the way but so are the

banks to do that yes in the UK this bank

called munzo does that in Germany does a

bank hold on 26 and they're pretty cool

because they have transparent credit

cards isn't that cool so you know what

the so one of the sponsors here their

CEO he says well basically it's not the

startups they were frightened of it's

when the Amazons of this world started

to move in right now Amazon is moving in

because in Germany you can already get a

loan at Amazon I'm not sure if we do it

here already but we don't even have an

Amazon dot NL yet so I guess we're not

important enough but this is where stuff

gets different right and companies like

Apple move him so happening Apple starts

his own credit card

right and it's pretty cool as well

because it doesn't have anything on it

right your name isn't it doesn't have a

number I don't know how to do it but

it's pretty cool actually even though I

don't like Apple so what if other

companies are coming into the industry

as well what if you're in FinTech and

our ACS says you know what we're going

to do payments as well because we do

roads of payments tickets and the coffee

you buy on board I hate it by the way

when I fly in has to buy my own coffee

it's terrible well the coffee's terrible

anyway but so this is anybody can come

into this field and only just now about

an hour ago I read this Facebook

announces a new cryptocurrency called

Libra I think it just announced it today

or yesterday what they're gonna create

their own coin this is probably going to

change as well so the whole industry is

changing right and this basically goes

for any industry you can be in your

industry as well to sum it up basically

you can say that we live in a world

where everybody anywhere everywhere can

enter any market at any point in time

you want to because it's technology it's

us as technologists developers that

change the world basically and if you're

in a company in a market that is about

to be disrupted probably there is

another party doing the same thing you

do but smarter and faster and cheaper or

even disrupting the whole business model

right if you're a taxi driver five years

ago you were okay until rubric came

along right if you were a hotel you were

okay five years ago until Airbnb came

along any market can be disrupted at any

point in time right the company I work

for is in the energy market last month

right after today we had a strategy

session Google announced that they were

going to send this new home hub to the

Netherlands as well how does that change

our company we don't know yet right

that's the thing you need to figure out

continuously as a result when we build

software we need to start doing things

differently even differently than you do

already you might think all right Roger

already and if we're agile we're good

right

most people think most managers England

is there anybody manager here in the

room Oh only two well I explained the

two afterwards again right and I'm just

kidding

that one down easy anyway so so anyway

most people think all we do scrum so a

good no you're not you're not safe even

if you do a scrum and if you do safe

well you're not even agile anyway so

sorry about to save people in the room

so what I think you should do you should

all move beyond them right you're all

pretty much agile you're developers

right so you're good at this already so

what I think we should do is we should

move to actually developing small

services over doing projects the

metaphor project has never worked in our

industry the software industry that is

right because it's the wrong metaphor

I'll try to show you also we need to

move to even shorter cycles like if you

have two-week Sprint's now go beyond

that can you yes you can

you know how by removing most of the

ceremonies don't laugh it's that sexy

this is a true story right so and you

need to go to even smaller teams because

the current teams like six plus or minus

three teams they don't cut it anymore

and I'll explain that to the part I

won't explain today is the part about

even moving to smaller components but

there's many talks on micro services

here already so I suggest visiting well

not now but after this talk visiting all

of the micro services talk ok so these

are the three topics that I will try to

discuss first of all let's go to small

services there's an interesting model

and it's this one have you ever seen

this before you're a manager right

you've seen it good good that's good no

I'm just I'm not gonna make jokes about

managers anymore right this is the

connection framework this actually

opened my eyes big time about how we do

our work it's written by a guy called

Dave Snowden he's a professor at Bangor

University in Wales and he wrote this

down while he was working for IBM he

said basically you can distinguish four

or five different contexts or zones that

she could be in right you could be in

what is called the obvious on which is

down here

my pointer works it looks like I'm drunk

but it's actually the thing itself right

so this is the obvious context in the

obvious context if you have a problem

there is a best practice there's no

doubt there's only one practice right if

my sink is full of dishes I have to put

them in the dishwasher that's the only

solution yes I have three kids they're

not gonna do this have to do myself it's

the only solution do it you're done

right this is plan about this you can

say we know what the solution is we'll

go there we'll write it and that's it or

we'll just implement it or whatever you

do right don't write your own ticket

system there is existing ticketing

systems there okay sort of and and

that's it that's the best solution

alternatively you can be in a

complicated context in a complicated

context to a given problem there is a

set of possible solutions and the thing

you need to do is you need to you

analyze each of these solutions and

figure out which one is the best and

from there maybe build a tested

implemented etc etc now these two

contexts which are on your side here are

actually the ones you can manage in the

way that most companies are managed a

linear style this is where waterfall

prevails right but unfortunately most of

us are not in there most of us are

either in complex or in chaotic contexts

now that means in a complex context you

can still set a daata eak arising and

say this is where we are slowly heading

and then experiment towards that

particular goal and when every step you

take every step no I'm not gonna sing

anyway so this is basically it works or

doesn't if it works go in the same

direction if it doesn't change direction

that's the way it works most of us are

in fast-changing worlds meaning there is

not a predefined set of solutions

anymore we don't know how to get where

we're going because every time we do

something Amazon comes up with two new

frameworks or solutions or whatever they

have and you have to rethink everything

again that is where we are right it

could be worse if you don't have adopted

horizon you don't know what your point

what your strategies what your vision is

or your mission or all this other

marketing mumbo-jumbo you can figure out

the why in the how and the what and

whatever right so if you don't have a

particular goal to go to you need to

figure that out the only thing you can

do is take a small step see what happens

and from there I decide on what the next

step will be now most of us are in

either complex okay or X owns a

colleague of mine oh yeah wait I had

this attitude so um just to sum it up

basically a friend of mine came up with

it he said well this is the zone where

you could say duh that's the solution

right and in the complicated zone you

could say ah that's the one we're going

to pick that one now if you're in a

complex tone it's more like were

pronounced right and if you're in a

chaotic zone if you are there we're in a

company that is there it's basically

like we don't know we have to figure it

out right so a colleague of mine says

basically we are operating in a new and

fast-moving market nobody has invented

it's smart energy market a profitable

business model for this market yet

meaning we have no daughter at the

horizon meaning we can only take small

steps and hoping that we get somewhere

meaning it's the end of projects right

projects are out of the door so you have

to stop because if you are here or here

you need to tackle things differently

right you well the good thing is

basically if you look at Dave Snowden he

says well in a chaotic context which is

where most of you will be in searching

for the right answers will be pointless

that's sad basically iteration ships

between cause and effect are impossible

determined because they shift constantly

basically because of new technology and

new insights and whatever you might have

and no manageable patterns exists only

turbulence that is where most of us as

developers are in now the good thing

about this is is that this chaotic

context or domain is nearly always the

best place to imperil innovation so if

you like innovation that most of us

probably do otherwise you wouldn't be

developer I guess that means that if

you're in a chaotic context you can

innovate all right which is pretty cool

that means that projects and what a

foolish stuff this actually comes from

the original white paper by Winston

Royce don't work they don't work in our

industry so if you're in a project get

out that's the main message I guess if

you have detailed planning's

and the company you work for find out a

job right you need to get out here

because this doesn't work it's certified

not to work what you need to do is you

need to move in small steps you need to

deliver small features you need to

deliver small services because if you

only deploy into production like once

every three months it's big that means

you need to do a lot and a lot of

testing meaning you go slower and slower

and slower in the end your innovation

will stop the only thing you need to do

is to move into delivering small

services again and again and again and

again I have a lot of these actually so

this is basically to illustrate my point

is that move in small services that also

means that you will end up with I don't

want hundreds maybe well tens of maybe

hundreds of repositories and pipelines

and lots of automated testing as well

and also it means that you can finally

after doing them wrong for fifty years

in this field stop doing projects right

Wow you're all very quiet now is it too

warm what it is right so let's see what

it's in the second step the second step

is about shorter cycles now I suppose

most of you are in like scrum or agile

project having to us to experience are

there people with three week Sprint's

four week Sprint's people who don't have

Sprint's well I don't have Sprint's

either but that's because I sort of

abolished them basically but I'll tell

you the story so what about shorter

cycles now most people I meet in this

industry and I go to a lot of events I

go to a lot of clients to do coaching

and if you go to all of these let's say

agile events my tip don't go there by

the way they're horrible it's sort of

like this people say yes agile means

scrum right

and and a bit of JIRA and that's it and

it isn't right

agile was not about doing scrum right

agile was about adding value

continuously it was about learning stuff

continuously about improving

continuously that is what agile in my

mind is about right so it doesn't mean

if you do scrum right that you're agile

and that you're equipped for this new

world that is changing faster than we

can think of right also what I see is

that a lot of people come in from

outside of the industry into our field

and and do mastery stuff scrum mastery

stuff right now if you would ask

somebody who does I don't wash a large

like karate so my girlfriend did karate

for quite a while and she was quite good

at it she was like 17 times Dutch

champion by the way I didn't knew that

when I started dating her she's kind of

dangerous and I said well how long did

it take you to become a masters she said

well actually I'm not a master 17 times

Dutch champion right I'm not a master

but I did train for ten years five times

a week three hours a day that's when you

get to become a master now there's this

other mastery that we occasionally meet

is that why you get the point right so

that of scrum master er you do the two

day training course you do 30

multiple-choice exam questions and

whether you are a history teacher a

physiologist or an archaeologist you can

become a scrum master and then you get

to solve all the problems we have in the

industry right even though you've never

seen a line of code in your life before

now I'm not sure that works because what

will happen if you start doing things

like that it becomes really dogmatic

right you see people like that solving

problems in a way that's proven not the

best rate so I'll read it to you the

dogma is the established belief or

doctrine so scrum help our agents from

the scrum work or a particular group or

organization its authority and not to be

disputed ever doubted or diverged from

by the practitioners that means you have

to do stuff exactly as a

in the scrum guide now the scrum guide

is far from complete by the way if you

want to do projects or even move beyond

projects because it's basically meant to

do products so the thing is if you have

become a scrum master or whatever

certification you have that doesn't mean

that you're equipped to solve all the

problems in the industry these are my

girlfriend and my son there we went

skiing and I said well I'm going skiing

cuz I know how to do that I'm I'm old

I'm not gonna learn snowboarding anymore

they started snowboarding took him three

days and they couldn't even go down hill

and they said you know what for the last

couple days we'll just ski because it is

really hard to learn new stuff right the

same goes for coaching projects or

coaching teams it's hard basically right

so that means if you do alpine skiing or

agile coaching or scrum mastery or

writing code or whatever you do it takes

a lot of knowledge and experience so

that means that if you have a project

that says oh we have a sprint we're at

the end of the Sprint we had some work

left and this is your second sprint and

your third and your fourth on your fifth

and your sixth and you have work left at

the end of every sprint and your project

manager this is an actual project from

Brussels basically in Belgium and and

the project manager says you're terrible

you don't know how to estimate your work

true also you don't know how to do your

work because you have work left every

sprint and by the way we only have three

Sprint's left that means for the

remaining three Sprint's you need to be

three times as productive as you are now

an Annie concluded that the teams that

were on the project would never gain

that productivity that's true by the way

so he hired three hundred developers

from India this is a true story by the

way and of course the project failed

miserably not because they're from India

because adding three hundred people to a

project that's already failing doesn't

really work right that's Brookes law

adding more resources to relay project

makes the project even later and then he

got fired and the project failed of

course so that's it right so the

question is do

these prints actually matter do you

actually need to do sprints if you want

to be agile and the answer is no because

if you look at the agile manifesto as

old as it may be it basically says

continuously delivery of valuable

software right

it doesn't say well every three weeks

every three weeks every month it says

continuously that means you have to put

stuff into production as fast as you can

as often as you can right not like in

this particular graph that I have here

this is a nice reaches like in 2014 13

percent of the people who voluntarily

filled in this inquiry said they never

went into production with anything I

have no idea how this works by the way

but the thing is you have to move into

delivering as fast as you can as often

as you can

that also means that all of the ceremony

that you add having Sprint's like these

in or planning sessions and sprint kick

of them whatever you do and refinement

sessions they can go out of the door

because the only thing you need to do is

figure out what are the items that we

want to work on and what's the most

important item to do now work on it move

it from left to right on the board by

the way I'm not saying here use JIRA

this is an example right it could've

been post-its on a wall as well but they

fell off so the only thing you need to

measure is how fast can we get the items

from left to right because that's the

only value that actually matters

how fast can I get stuff into production

and the faster you go the more agile you

are basically that changes perspectives

bit and by the way if you think hey it

looks like Kanban oh it looks like him

but except that Kanban only says

visualize your process and improve from

there it doesn't tell you what the

process is so you need to figure out

what the process is yourself by the way

I find this board game in a shop called

Kanban the description on the back is

horrible go out and read it anyway so

and if you want to move to continues

delivery

actually continuous delivery that means

you have to automate everything

everything in the kitchen sink this is a

Jenkins pipeline and I don't care if you

use Jenkins or go CD or bamboo or

teamcity or whatever you do all I'm

trying to figure out and say automate

everything this is a literal pipeline

from one of my clients saying okay as

soon as I check in my code the build

kicks off my unit tests run if they're

okay we'll send it off to Sona q that

repeats the unit test basically and

checks for the amount of code coverage

checks for my cyclomatic complexity if

that's okay we'll package into a darker

darker image which is deployed onto an

environment where we run API tests or

integration testing from there it's

deployed again to the next environment

and they're all the tests between these

components and all the other components

take place from there goes to acceptance

they're automatically performance tests

are executed and if they're okay it goes

into production this whole process takes

about five minutes right and so from the

time I check in my code including the

unit tests of course because everybody

writes unit tests right I'm not going to

do a show of hands because it will be

terrible anyway so that basically means

it takes five minutes to put something

in production actually in production

right now if you can go at this speed

that means you can actually deploy

hundreds of times per day having tested

everything automatically you're very

very agile for instance this example

this is one of the marketers that a

particular company has like three

thousand lines of code and a code

coverage of ninety seven seventy nine

percent two percent right 342 unit tests

on that particular piece of code that's

quite a lot that's one for every ten

lines of code I guess right so you need

to automate everything if you go in this

direction and you need to go in this

direction because you want to stop do

projects right now it's getting really

warm you're like oh I'm not good I'm at

the wrong company now maybe you are

maybe you're not well let's investigate

the third part which is the part about

the teams now agile was actually so the

word agile came from the agile manifesto

which was written down in 2001 which is

18 years old right in those 18 years the

world has changed quite a bit and

there's lots of issues that the agile

manifesto although it's a brilliant

piece of work and I love it to death

doesn't solve current day issues we now

have that we didn't have back then for

instance there is a low availability of

what managers like to call resources

I mean people right so so if you look at

so this is a Dutch resource being

conducted last year it says basically in

Dutch so if you're Dutch you're okay if

you're not well I can translate it for

you it says very tight market right it

is very hard to find the right people

company women I work for we've been

trying to find Java developers for for

over two years I guess now it's hard

right we cannot get to the right people

as a consequence all of us need to

become t-shaped or by shape that's

implied with too strong that's of our

expertise but you need to know one thing

really good

it's your sweet spot and a lot of other

things you need to take into account as

well you need to learn about how AWS

works about how Azure works about domain

driven design about micro services about

continuous delivery about automated

testing about testing frameworks

whatever right there is lots of stuff

you need to know now as a developer that

you could say well there's this

specialist there ten years ago they're

not there anymore we need to know

everything that means we need to learn

continuously also if you live in this

particular country or most countries

actually this is a picture I took on the

highway to Amsterdam a couple of weeks

ago right if you look at the other side

of the road I'm in traffic this is five

lanes going from you tech transfer them

on the other side of the road there's

the same amount of people getting stuck

in traffic in the opposite direction and

you can ask yourself what is the point

in that right could we just stay at home

and read our emails at home and then go

to work what if 20% of us would do that

right we would actually solve all

traffic problems in the country I guess

and the third thing is

despite that most managers think that we

work best from nine to five our minds

doing creative work they don't work like

this they work a bit like this right

managers might think you work from 9:00

to 5:00 and taking a braid and don't

think during your break but the blue

line is actually basically how we think

I get most of my ideas when I'm in my

car or in public transport or even in

the shower or while I'm having breakfast

that means I need to be able to figure

out when to do my work that's the best

time that I can do it basically when I

have these ideas which is not always

between 9:00 and 5:00 so we should

change basically our work ethic and it

changed a lot also well you might know

that as developers we find communication

often hard right it is hard basically it

is very hard to understand if people if

you listen to Bruce Dickinson this

morning he basically said the same right

he said it's not always easy finding of

an accurate answer to a question you ask

or the other way around to give an

accurate answer and that's because

communication is hard so what do these

people in agile projects do they come up

with all sorts of meanings why why I'm

the best architect in my company he's

always gone he sits next to me

officially but I never see him because

he's always in a meeting

why is he doing actual work right the

thing is we have too much of these

ceremonies even in agile right so we

should cut down on this stuff right now

if you work in smaller teams instead of

saying teams of 9 to 12 to 15 people if

you would just be able to work with two

other people you would save a lot on

this communication you could save a lot

on all these terrible meetings actually

that means in my point of view we don't

need more communication as a lot of

people propose we need better

collaboration that means we need smaller

teams to do that

smaller than we are used to right now

that also means like stuff like these

full-day refinements sessions do you

like them do you have them do you like

them not particularly much right

none of us usually do right they're

boring a lot of people in the room a lot

of people in the room have nothing

say and just sit there until they get to

do the t-shirt sizing exercise or the

planning program then they wake up right

and I think we should abandon this if

you move to a continuous flow doing item

by item the only thing you need to take

into account with let's say 2 to 3

people who work on it is that particular

item from the point in time that you

pick it up the backlog which could be

any point in time at any day you only

need to worry about that item that's the

item you need to discuss with the two

other people you're working on it not

with the whole team that really doesn't

add too much value all right and then

there's one thing I really have to

debate you are all in open floor plans

right do you like it no we don't you

know what happens this is the golden

opportunity if you are producing noise

cancellation headphones right if you're

Sony you love open floor plans because

everybody is buying these and

everybody's sitting there the only way

left to communicate if I have to talk to

somebody who's sitting next to me so bad

called Yerevan he was always had his

headphones on I messaged him almost like

and he's sitting next to me right that's

not a proper way to communicate

basically open floor plans kill your

productivity there's plenty of research

being done so if you are in an open

office space find the research go to

your manager and say dude this is our

girl this is not working I need to be

able to concentrate I need to be able to

focus on my work right and the next big

thing of course it's autonomy

now autonomy is interesting this is my

19 year old drumming son and after high

school he took a year off and he's like

that I have no clue what to do so go

figure it out it's your life good

this is giving autonomy as a parent

right which is really hard it's

basically the same as giving autonomy to

team maybe a bit harder even some

cigarette out and he found a study he

went to study here in Amsterdam doing

politics philosophy in economics after a

month he said that I'm gonna quit like

okay and it's not no you cannot quit

after at least do it

fear and I'm like dude follow your heart

right so he agrees with me of course and

he stops and he said I want to become a

professional drummer now that was like

nine months ago and in the meantime he

got admitted to a conservatory so he's

he's practiced really hard and we gave

him the freedom to figure it out himself

that is really hard to give somebody

just enough autonomy so they can rule

their own life it's not easy because for

most people being autonomous or

self-organizing teams means you have to

go out of your comfort zone right this

is not the zone that we want to live in

because we have to get out a figure I do

stuff usually if I say the teams

yourself organizing now you're

autonomous the first question they'll

ask me is what do you want us to do

which is not the right question right so

what I figured out is this is extremely

hard it's basically I can tell you if

you want ask me how to draw an owl I can

do the first two circles but that's it

the rest of it you have to figure out

yourself

that means autonomy it's good because

having people make decisions at the

right level in organizations that's

extremely good but it doesn't mean this

this is that Zappos actually in the u.s.

would you want to work in an office like

that no I wouldn't basically but you

think this is only in the US where do

you see the next picture you know why

this is it's not IKEA by the way this is

my girlfriend slag in a job interview

for a Sears management role at a Dutch

retailer I'm not going to tell you which

one but Durham Rotterdam and I have a

blue logo no I'm not going this is she's

in a meeting room up to her knees fill

their balls right that has nothing to do

with atonium yourself organization this

is overdone even though I like the

company and order lots of stuff there

the only thing I figured out to be more

autonomous it means if I put less rules

on my team's just enough to set a little

bit of boundary they're better off and I

came to that conclusion when I was in an

in an easy action last year so I visited

the city called Maidan which is you

chilly left three million people in it

and I was in a taxi or in a car actually

that picked me up because I have to

drive all across the island to go

somewhere else and and

this is where we are crossing the road

remember this is about autonomy and

self-organization right do you seal the

traffic lines all the traffic signs and

the lines on the road you see them

they're not there right there isn't a

single sign on the road actually if you

look at the traffic light it says both

red and green at the same time what does

that mean right still we cross the road

so what happens here right the question

is what do these people do what do they

do to communicate right they pay

attention that is what they do and they

do that by themselves and they have to

because there's no rules and still we

gotta cross the crossroads safely which

is probably coincidence but still the

idea is good right I'll show you the

crossroads around the corner from where

I live it looks like this not every day

but it happens because there's lots of

regulations and lots of traffic lines

and signs if you go through the country

and in edlund's you'll find signs on the

roads every 10 meters

we're over regulated which means people

stop to think and that is basically what

we need to do we continuously to think

about how we are doing stuff and what we

are doing also the next thing is you

notice right you know there's a good

solution to this there is no estimation

right the thing is and I'll explain it

to you by showing you the law of large

numbers and this is a very interesting

theorem it basically says the law of

large numbers describes the result of

performing the same experiment or a

large number of times according to this

law the average of the results obtained

from a large number of trials should be

close to the expected value and you

might think what on earth that does mean

late in the afternoon in a warm room in

Amsterdam it basically means that if you

estimate let's say on a scale from one

to five and you estimate a hundred items

and you take the average what do you

think it will almost certainly be around

three so if you know that already

why spend all the effort in doing stuff

like I don't know what planning poke

t-shirt sizing right it does not add any

value it's nice to do the first three

times but after having seen this a

hundred times and every sponsor coming

up with their own deck of cards new

planning poker I sort of say well just

count the stuff right if you discount

the stories it's the same as doing

esters because you get to the average it

means you're estimating on a scale from

one to one right so it just adds up to

the efforts so you can basically stop

doing all these estimation techniques at

low level because they don't add value

so you can quit doing these also again

there's the red sprints of course there

was a company actually I work for that

actually called them Brett's prints that

is just terrible right so you know what

to do you can just stop doing these

because this is a ceremony that we think

is mandatory in being agile and it's not

you can go beyond this stuff you can go

figure it out yourselves right also the

thing is that software development has

became way too complex having a team of

six people plus or minus two or three

whatever you think is current doesn't

cut it anymore if you look at what

that's kadaj successed famous dutch

computer scientist he says the program

has to be able to think in terms of

conceptual arrow keys that are much

deeper than a single mind ever needed to

face before he said is in 1986 he

basically says this field is the most

complex field you can ever be in that

means building suffer does not fit

anymore in a single person's head that

means you have to collaborate with other

people but the last people you can

collaborate with having a small team the

better it is right working with two

people or three people is much more

efficient than working with 20 people so

yes small teams are good right but still

we need to have a lot of knowledge in

place if you look at one of my clients

previous code bases they have like nine

million lines of COBOL and they have

three million lines of PHP I don't know

which one is worse but

I could take a guess though but I'm not

gonna do it so they have to maintain

this with twenty developers I asked the

developers of my current client to

actually come up with the list of

technologies that they are using it's

this they have like 40 developers on

board max right how do you maintain all

of this stuff

well you don't basically because it's

too hard and then well the good thing is

if you start working in smaller teams

even smaller teams any work now every

work item that you might pick up from

your backlog continuously actually only

requires a small subset so if we are all

t-shaped

you can do a single work item with two

or three people and then move on to the

next item and then change maybe the way

your team or your small team is composed

change it all the time for every work

item which is a much more effective way

of doing stuff I'll give you one more

problem we have is that the framework

and I don't mean spring I mean the agile

frameworks they don't actually cut it

for you because the thing is my

conclusion is you always have to think

for yourself right so yes small teams

are good really small teams it doesn't

mean that you actually have to do all

the ceremonies that your agile framework

whatever you might think probably scrum

is telling you to do because they were

put there in place how can something be

a best practice that was written down 20

years ago that doesn't mean we haven't

learned anything since right and we did

we learned a lot of things so we can

move beyond that and if you think that

oh yeah we need all this stuff because

we need to scale agile I couldn't care

less basically that's why I don't like

working for big companies so if you are

in the field for saying our we need

discipline agile delivery which is one

of the enterprise agile approaches how

many roles do you think are in there

well a lot actually that's terrible

right

how can you deliver suffer if this is

your setup or it could be even worse you

could do safe is anybody of you in a

transition transformation towards safe

quit

I'll show you why right this is my this

is my annual quiz let's go where's the

customer I'll show you the picture from

safe I think it's four to zero and you

have to answer why the customers right

this is the picture where is the

customer I found it by the way it's on

the right side there in the middle right

it's the dude being hit by train so dust

it out a dirty no it doesn't right

I don't so des companies and I know

companies that are also here that are

trying to copy the Spotify model who of

you is you know that the Spotify model

was written specifically for a company

called Spotify right that grows like

eight hundred percent every year doubles

their number of people every year for

every six months or whatever they do

also they're wrote this down in 2014

that is five years ago right that means

if you are implementing something they

wrote down it was the way they work then

and you're saying are we going to do

this now because it fits perfectly for

us it won't because basically you're not

Spotify well unless you work for Spotify

but then again if you work at Spotify it

has changed already because it's five

years down the road they have learned

right and so should you

you shouldn't be standing still at some

framework that you might implement a

thing always safe now oh wow that's not

a good example oh we're doing strum so

we good for the rest of our lives

you're not going to be good for the rest

of your lives what you need to figure

out is that you cannot just copy

somebody else's model and think it works

for you you have to figure out your own

way of working right and that is

probably different from any one of these

existing frameworks it's up to you and

in the end this is probably the most the

smallest approach I can think of is

meaning you get to work you have a

conversation about the one item you're

working on you do some work you deliver

it into production automating the

out of it go home have dinner with your

family or your friends and come back the

next day that's about it alright this is

the only approach you need F by the way

if you do that you can basically stop

doing projects I got to the

retrospective right which is on well

every project needs to have a reach

spective so this is it so the thing is

we have never been in this field in

complicated review zones right because

everything in this field changes

continuously and it's changing faster

and faster tomorrow it will change

faster than it did today that means we

are in complex or chaotic zones meaning

there's no projects there's no waterfall

that helps you there's no framework that

if you implemented will work for you you

need to figure out how to do this

yourselves so there is a conference I

was in Romania last year and I could win

a t-shirt basically so I went in to join

this multiple-choice question net Stark

is a senior product on which approach do

you think he will recommend when

refining the product backlog very much

tuned towards scrum right and I look at

the answers and said well basically

anyone is good answer

depending on your situation and they're

like no it's B or C I don't remember

anymore right

they were very stuck on what they were

doing and you need to get beyond it you

need to stop implementing somebody

else's model and if you think as a

developer you cannot do that you cannot

move beyond where you are now right

because you're only a developer or

you're only QA engineer you're only your

manager right it's not much okay so it

could be you if you want to change its

you that's got to do it it's your life

right if you want to change you need to

I actually tried this out on the parking

spot of my son's football team and I was

the first one to arrive that never

happens by the way so I parked my car in

the wrong direction

and everybody followed right so as

technologists right even though we might

look a bit like this usually this is

probably a Java conference too we are

the ones that change the world right

it's because we change things it's

because technology evolves so fast and

we are basically the only ones who can

keep up with that so that means as a

result you will never can stop learning

right have to figure out your own way in

doing this for every team every

organization is going to be different

than for any other one also we're in the

best industry

right because this is the most fun

industry I've ever known that's my

conclusion I'm like two minutes overtime

so for me that's pretty good actually by

the way stop doing projects they don't

work in our industry I have a QR code in

here if you want to go to the handouts

which are actually longer than I did

today thank you for indulging me in this

warm water and I hope you enjoyed the

rest of the conference

[Applause]