1. The document discusses various feedback loops, including in science, open source software, user testing, online communities, free speech, marketing, and evolution.
2. In science, a feedback loop involves forming a theory, making predictions to test it, testing through experiments, and revising the theory based on results.
3. Evolution operates through a feedback loop where an organism's fitness for the environment determines whether it lives to breed and pass on its genes.
Axa Assurance Maroc - Insurer Innovation Award 2024
Â
Better, Stronger, Faster Failures
1. Bett r
Str ng r
F st r
Fa l res
Nat T rki gt n
He H n nga Consulting Partners
Riverside Est t Pr pe y Dev l pme t Tr st
W ll Ut lize Syn rgi s f r $ $! !
2. <t lk g es he e>
Since I spoke at Webstock last year, Iâve had literally twos of people
come up to me on the street and animatedly say how my talk was the
highlight of their conference experience. SpeciďŹcally, it was hilarious
that I had a font goof and several letters were missing on each slide.
Font fail, if you will.
3. <talk goes here>
This year that wonât happen, sorry to disappoint. Again.
Anyway, Iâm going to tell you some stories and develop a thesis that
will hopefully make you feel more comfortable about something you
donât currently like. Your job is to see whether you can ďŹgure out
what that is before I just out and tell you. If you succeed ... well, I
apologize in advance for how boring the rest of the talk will be. HA.
4. Iâm going to talk about virtuous circles, or feedback loops.
9. change?
Ah, well, you see, it doesnât. Not in any systematic way that
progressively improves our understanding of the world weâre in.
10. 1610
This guy was the ďŹrst scientist. He inherited a world view that was a
thousand years old, developed by Aristotle and Ptolemy and reďŹned by
countless others. The view was being questioned by the time he
arrived, by people like
11. 1543
Copernicus. He was an astronomer who like all astronomers of the
time made observations of the night sky using his naked eyes. He
suspected that we didnât have a geocentric cosmos (Earth in the
middle), we had a heliocentric cosmos (Sun in the middle). He and
Kepler, though, couldnât bring themselves to abandon all the trappings
of the old model and so threw themselves into inelegant maths to
make a theory that was both Aristotle-compatible and observation-
compatible.
12. Even Galileo began astronomy by seeing the skies through the eyes of
an Aristotelian. But Galileo, when he looked through one of these (a
âlookerâ, the ďŹrst telescope)
13. saw more detail than anyone had ever seen before. He made detailed
maps of the moon, and tracked the motion of planets through the sky.
He even saw moons on other planets, something that hadnât been
seen before and which deďŹnitely violated Aristotle.
14. Galileoâs observations about planetary motions were hard to explain in any
way but that the Earth and the other planets are in orbit around the Sun.
Tada! Heliocentrism overthrows geocentrism, cosmology takes a giant step
forward, science rises, enter the Enlightenment and itâs a short hop until
weâre exploring those things that Galileo saw.
Not that fast. Kepler and Copernicus had met with opposition when they
mentioned their theories. But Galileo didnât give up. He wrote a book, in
common Italian rather than the Latin of the elites, explaining his theory.
His book had a character called Simplicus in whose mouth were put
15. the Popeâs arguments, made appropriately ridiculous. The pope used
to be his friend. Used to be. There were charges of heresy, and
16. he went before these fellows, the Inquisition. Galileo had to recant,
say âok ok, I donât believe that at allâ, and died still under house
arrest. However, rumour has it he muttered
17. E pur si muove!
âE pur si muoveâ after recanting. âBut still it movesâ. Anyway, this
stubborn bugger is widely acknowledged to be the ďŹrst scientist. He
had begun as an Aristotelian with his theories but changed them
based on evidence, and he went looking for evidence to test his
theories.
18. science = world view?
He got in such a power of strife because he was trying to change the
view of the world. We think science is the view of the world. Itâs not.
19. Science = change
Science is how we change our view of the world. Thatâs what science
is, itâs change. Change is scary if youâre the fat rich bugger with a
slick beard and a lot of sitting on the top of the heap. Hence all that
business with the Inquisition, house arrest, etc.
20. change based on
evidence
Itâs not just random change, though, itâs change based on evidence.
Scientists change their opinions based on observation.
21. In fact, thereâs a nice little loop here. A feedback loop. A scientist has
a theory, which makes predictions. Well, you can design experiments
or observations to test those predictions and see whether theyâre true
or not. If theyâre not, then you have to formulate a new theory and
with it comes a greater predictive power and understanding of the
world around us.
22. Of course, if we just test and ďŹnd itâs true, we never get o the loop.
We never get to revise and thus improve. So itâs important to, at some
point, have a test come back that shows the theoryâs predictions are
wrong. In fact, that is THE most important thing. To be science,
23. FalsiďŹable
theories must be make predictions that you can measure and show to
be false. It took nearly 500 years after Galileo for that to be ďŹgured
out--500 years during which gravity, magnetism, electricity, atomic
theory, and even quantum theory were ďŹgured out, so obviously not
understanding what they were doing didnât entirely stop them making
progress. This theory of falsiďŹability was put forward by
24. Karl Popper
1934
Karl Popper. Good book, âLogic of ScientiďŹc Discoveryâ, well worth
reading. Kiwi connection too--he came to Christchurch from Vienna,
fell in love with the place as a socialist haven, and only left to go to
England to Cambridge, be knighted, and die famous. He managed to
live two Kiwi stories: the creative class immigrant and the brain drain.
25. knowledge increases
when someone is
wrong
Anyway, the key point to make is this: our knowledge about the world
only increases when you ďŹnd out that someone was wrong. Thatâs
what science is.
26. Scientists Seek Failure
All that scientists want to do is to ďŹnd a failure. Theyâre human, of
course, so preferably itâs someone elseâs failure, but they can only get
a name for themselves if they ďŹnd that someone was wrong.
27. a theory with
predictions
revise test
how did it hold up?
Because thatâs what the feedback loop of science needs.
28. 1. A/B Testing
Whew. Letâs look at some other feedback loops. I talked about A/B
testing at Webstock last year.
30. A
?B
two dierent versions of the same web page. Well, not you obviously,
they donât show you two pages at the same time. They show 10% of
their visitors the shiny new version, with dierent ads or a bigger
Kindle picture or whatever, and the other 90% get the standard home
page. Then they see who buys more per capita--those who saw old A
or new B. If the new page gets more sales, they whack it in for
everyone.
32. we assume that the current way of doing things is the best possible.
Then we test that by running a dierent ad. Does it result in better
sales? If so, our assumption was wrong and we have a new candidate
for The Best Way Of Doing Things. Which we test by running a
dierent ad .... This is continuous improvement, each time around the
loop makes us better.
33. 2. Open source
Or another example, Open Source software like Linux and Rails.
34. âscratch your own itchâ
The guiding philosophy of open source is that you can scratch your
own itch. Whether itâs a printer driver that should mail you when you
job is done, or a new button in your mail app to unsubscribe from
mailing lists, you can add the feature if you want it.
35. Hereâs the feedback loop. You use the software, if it doesnât work for
you, you can make it better. In proprietary software, you rely on
market forces to direct the vendor to scratch your itch for you. Not
always going to happen and not always most eficient.
Notice how this isnât reďŹning a hypothesis, this reďŹnes software.
36. 3. User testing
Related to this is user-testing. Remember science? User-testing is
the experiment and evidence part of the loop: if you take it out,
thereâs no way to go around the loop and get better! Wouldnât it be
horrible if we lived in a world where software didnât get better because
it was tested on users?
37.
38. 4. Community
Another example, from web sites. Can you build feedback loops into
sites of community contributions to steer the participation in
directions you want? (e.g., away from goatse and swastikas)
39. One way is to let people tell you whatâs good and whatâs bad. Hereâs
Digg.
40. Hereâs Slashdot, with thumbs-up and thumbs-down. Slashdotâs been
in this space a long time--they have a very functional and elaborate
system for managing comment threads using âkarmaâ points that you
give to comments you like.
42. So this is the feedback loop. We show some news items or photos, we
get the user to apply critical thought and tell us whether theyâre good
or not and thumbs-up/thumbs-down them. This then informs our
choice of news items and photos to show the next person.
43. homophily
The danger is that you only show people things that they like, or you
only show people things that everyone likes, and nobody is ever
surprised or discovers things they didnât know they wanted. This
situation, where we only see more things like the things we like, is
called homophily. The opposite of homophily is
44. serendipity
serendipity. We can break our feedback loop, deliberately, to add
surprises and new things. Just because we know whatâs popular and
what people like, doesnât mean we have to show just that to them.
45. 5. Free Speech
Another feedback loop is in free speech. You see, there are two
dierent approaches to the problem of
46. âwhat if people say bad
things that are wrong?!â
people saying things you donât like. The ďŹrst, as exempliďŹed
47. Germany
by Germany, is to forbid it. There are things you canât say in Germany,
like âthose Nazis were lovely chaps, just a bit misunderstoodâ.
48. America
In America, however, you can say almost anything. Speech,
particularly political speech, is constitutionally protected. Why?
Because the founders of the country believed in a feedback loop
49. whereby the things people said could be subjected to critical thought,
determined to be bullshit, and corrected. If people werenât allowed to
say those things, the correction from the feedback loop could never
happen.
Of course, it all falls down because it assumes most people can bring
critical thought to bear upon bad arguments.
51. iPhone vs cafe training
or
Why Iâm not Steve Jobs
Hereâs a simple story to show you how they work.
52. Cafe Training Web Site
⢠market: cafes and restaurants with unskilled
labour
⢠product: web site with training videos
⢠beneďŹts: better staff, happier customers,
greater income
I had a brilliant idea for a web site. NZ has a lot of cafes, but high
employment. So the only wait sta are generally untrained high-
turnover people. That perhaps explains why the service is shit here.
So I thought I could run a web site with training videos for cafe sta:
the cafe owners would pay me $50 per person and the new waitron
would learn how to do it. The cafe owner would be happy, Iâd be
happy.
53. My friend Marc Hedlund shot this down with one question.
54. âWill your users do
your marketing for
you?â
How do you get your users to do your marketing for you? The cafe
owner would treat my web site like a competitive edge, and wouldnât
want to tell other cafe owners about it. Compare this to
55. the iPod. Everyone who had an iPod wanted to tell their friends about
it. Even if you didnât ďŹash it about like this guy advertising that you
were hip to the iPod, you had the little white ear buds and people even
bought little white ear buds so they could pretend their huckory Zune
was an iPod so they could get laid. Now thatâs your customers doing
your marketing.
56. Hereâs the ďŹowchart for the iPod feedback loop. The see/share-covet
bit is where we get the growth each time through the loop. If you
replace âpurchaseâ with âcontributeâ, youâve got Flickr, Slashdot, Digg,
and so on.
57. âviralâ marketing
With the iPod, one user infects ten others, and each of those infect ten
others and .... Youâve probably heard the popular term for this.
58. Itâs an analogy to the spread of a disease, or the growth in a
population. You start with one happy iPod user, in this case Thomas
Hobbes.
59. They tell nine other people. Now you have ten happy iPod users. Each
of those people tells another nine.
61. And now youâve got a thousand, and so on. Notice how with science
we have a ďŹxed number of scientists and weâre improving the model
one bit each time through the loop, but now weâre changing the
numbers and not the model? Weâre getting ten times as many
customers each time through the loop, exponential growth. These
self-reinforcing positive feedback loops are powerful things.
65. This is obviously based on the lobster. Good for getting around on
uneven surface underwater: tail pushes you down and forwards, legs
deal with bumps and holes without unbalancing the body. An
Australian biomimeticist was asked why he turned to nature. He said
âevery technique you see in nature is a success. The failures have all
died o over millions of years.â
66. Thatâs the evolutionary feedback loop: how well you ďŹt the
environment determines whether you live or die, and thus whether you
breed and pass your genes on.
67. ?
Question: what type of feedback loop is this? Is it a steady rate of
improvement, like science, or an exponential growth like customers?
68. M
UL
TI
PL
Y
Exponential growth: each individual makes many babies, each baby
makes many more babies. In only a few generations dramatic changes
can be seen in populations.
69. This is the light-coloured pepper moth, common in England before the
Industrial Revolution. So common, in fact, only a handful were
reported before 1850. They were an unusual mutation. When coal-
burning smokestacks started covering the landscape with soot, these
buggers were easy to see on the now-dark trees. Consequently, the
white mothâs population decreased while
70. this fellowâs numbers grew. By 1895, the dark moth was 98% of the
population.
Once coal went out of favour and trees slowly lost their carbon coat,
the ratio of dark moths in the population declined and has been
steadily declining.
71. M
UL
TI
PL
Y
This is because each surviving dark moth was able to pass on its
melanin-producing genes, and each of those dark baby moths could
make more dark moths. Thatâs how they drove out the light moths in
only 50 years. Exponential growth, powered by a feedback loop.
73. cit
eďŹ
d
e
d
tra
GD
P
inďŹation
ma
rgin
al t
ax
rat
e
unemployment
e
rate at
hr
cas
You might think of economics as the source of boring phrases in the
business section of the paper.
74. âWealth of Nationsâ,
Adam Smith (1776)
Itâs not boring. Borrow this book from the library and read the ďŹrst
few chapters. I found it electrifying: he nails, in an easy-to-
understand way, how the world works.
75. feedback loops of money
Money has feedback loops, and weâre starting to go through some of
them now as the economy contracts.
76. Businesses hire workers for money. Businesses sell things like
washing machines to workers for money.
Too many workers? Business wonât need to pay as much for them.
Those jobs that pay more will attract workers. Now businesses have to
pay more to get the supply of workers.
77. Similarly, if there are too many washing machines for sale, the price
will be low, perhaps too low to make a proďŹt, and businesses will stop
making them until the price returns to a reasonable level. This means
they need fewer sta, so there are more sta for hire, so wage costs
go down, so perhaps the cost of making them comes down ...
And so on.
78. simulatable
The point is that money ďŹows in a predictable way. So predictable that
you can build economic models: workers behave like this, businesses
behave like this, government behaves like this, consumers behave like
this, so the money ďŹows like this.
79. Thatâs what this is: MONIAC, a simulation of the New Zealand
economy ... in water. Itâs in the lobby of the Reserve Bank building,
not far from here. Iâm going to see it later today. Built by a NZ
economist, money moves between reservoirs (corresponding to govt,
banks, consumers, etc.) according to rules (valves) that you can
change.
80. price = supply demand
So economics turns the mystery of money into rules. Makes it
predictable, so you can look at the world around you and ďŹgure out
whatâs really going on.
The biggest rule is that supply and demand together determine prices.
Prices of washing machines, prices of food, prices of the software that
youâre building in your companies.
81. economy = b0rked
But, I hear you thinking, the economyâs crashing to pieces. Rules, my
arse!
82. But rules arenât the same as predictability. Hereâs a picture of that.
Each point on this picture is a feedback loop. You start with the (x,y)
coordinates, go through a very simple feedback loop a bunch of times
to compute a value, and colour the point based on what the value
converges to (or doesnât converge to). What you can tell is that there
are huge variations in outcome for tiny little variations in input. Those
exponential feedback loops, like evolution and money, donât give
predictable results. Thatâs chaos theory. Youâve heard about it on tele
and it sounds hard, so it must be true.
83. 9. Invention
Letâs bring it back to something thatâs near and dear to Webstockiansâ
hearts: new stu.
85. innovation = invention
you can sell
an invention that you can sell. Itâs easy to imagine an invention you
canât sell (shoes made of lava, an exploding toilet, a web site to teach
French to cats). Inventions arenât innovation. Itâs the useful ones, the
ones that you can ďŹnd a market for, those are the innovations that
change the world.
86. Joseph Schumpeter
(1883-1950)
this economist, Joseph Schumpeter, gave us some of the best thinking
about innovation. He coined the term
88. Thereâs a company with a product, a great product, it makes lots of
money. Think Sony and the walkman. Kids, a walkman is a portable
cassette player with headphones. Like an iPod but from the Dark
Ages. Anyway, Sony invented them and made a shitload of money
from them.
89. However, eventually a new innovation came along, the MP3 player.
Companies made those. The products werenât as reliable, they didnât
have all the music available that Walkmen did, they only appealed to
geeks and not the mainstream. Sony ďŹgured they were safe.
90. After all, the MP3 players were only earning a little bit of money from
their crappy devices that played only a few bits of music and that you
had to be a geek to drive. Sony focused on the Walkmen, selling
customers on the advantages of quality tape products.
91. But as time went on, MP3 players got better. More music was available
for them, and you didnât have to be a technohead to drive them. You
could do things with them that you couldnât with a Walkman. Apple
entered the market and ate Sonyâs lunch. Sony has tried repeatedly to
enter the digital music playing market, and has failed repeatedly.
92. Microsoft Linux
Its not just Sony and MP3 players. Think Microsoft vs Linux, or
93. Word Google Docs
feature-rich Microsoft Word facing down feature-poor but free Google
Docs
94. Creative Apple
even in digital music markets, Creative had their huge market
headstart shot to pieces from the expensive iPod that was Mac-only at
ďŹrst.
95. Kodak Canon
photographic ďŹlm used to be how you made photos, and Kodak ruled
the roost. Theyâve nearly gone under now, because digital cameras
ate away at their market and Kodak wasted its headstart and market
advantage just blathering about how much high quality ďŹlm was to
digital.
96. Newspapers Web
Newspapers are being eaten by the Web. Nobody knows how it will
turn out, except that the future of print newspapers is grim.
97. Big mills mini mills
not just computers, the classic example is vertically integrated steel
mills that could do huge amounts of work being replaced by the
smaller mini mills.
98. Archery Guns
even ďŹve hundred years ago, guns replaced bows and arrows. You
might think guns are naturally better than bows and arrows, but thatâs
hindsight talking. The ďŹrst guns were ghastly inaccurate things. But
you didnât need trained sta, so the archery industry withered as the
ďŹrearms industry bloomed.
99. ?
The question you should be asking is: what the fuck were these people
thinking? Why did they just sit back and let their huge proďŹts and
huge market share evaporate?
100. âThe Innovatorâs
Dilemmaâ, Clayton
Christenson
This book tells us why this happens. Big companies are invested (ďŹnancially
and emotionally) in the products that made them big. They must defend
those products against competition. They see it as stupid to adopt a new
technology product thatâs crapper than the great thing theyâre oering--itâd
eat away at their business, they donât know how to reach the people who
might be interested in that new technology anyway, maybe the economics
of the new technology are so dodgy that only a small company would want
to chase the ďŹrst customers. Lots of reasons why not to. They miss the big
reason: because if they donât kill their cash cow, someone else will.
101. âI used to be afraid of a
big company stealing
my idea ....â
A friend at a large American web company said, âI used to be afraid of
a big company stealing my idea. Now I realize that I have nothing to
worry about. They canât do anything with the ideas inside their
company, let alone the ones outside.â
103. constantly ask themselves whether theyâre meeting every customerâs need
as best they can with the technology available, and then introduce new
products to test this hypothesis. If the hypothesis is proven false with
sales, they update the product line. The problem with this is that itâs
largely bollocks. Itâs all big companies can do to breathe and service the
customers they have with the products they have. Donât expect them to
think agile and nimbly react to new possibilities. The best they can hope
for is to buy the upstart company before the legacy productâs marketshare
drops all the way to zero.
104. 10. You
Which brings me to the end, and to the ďŹnal feedback loop. Itâs not
about big companies, itâs about you.
105. 9.5. S92
Guilt on accusation.
Mashups
Questions in Parliament.
Protest at Parliament today, 12-12.30
Be a protest in Auckland.
Web site blackout on Monday.
Feedback loop for government.
106. Hereâs the genesis for this talk. My kids didnât want to run around the schoolyard in PE because the other kids were
faster than them. I remembered that pain well. But what I know now, now that Iâm not in school any more, is that itâs not
about racing other kids, itâs just about getting personal best times. I explained that this loop is how you get better: after
doing something, you measure to see how you did and then decide what to do to next. So you might practice more or
change your ďŹnger positions if you were learning the violin, or if youâre applying for a job and donât get it then you might
try dressing dierent or doing your paperwork dierent. If you donât run around the ďŹeld in PE, I said, youâll never know
whether youâre getting ďŹtter.
It worked, a testament to my kidsâ smarts rather than my motivational brilliance.
But I want you take this to heart. I want you to live your own feedback loop so you can get what you want from life. But,
you might ask, if theyâre so useful why doesnât everyone do it?
107. A Challenge!
Let me show you one reason why itâs not easy. Hereâs a challenge.
108. If a card has a vowel on one
side, then it has an even
number on the other side.
A B 4 7
Letâs say someone tells you that cards with a vowel on them have an
even number on the other side. How do you test this hypothesis?
(take answers)
109. If a person is drinking beer,
then the person is over 18.
Beer Coke 22 16
Is it any easier if we change the domain to something closer to your
heart?
How would we test the hypothesis that beer drinkers are all over 18?
110. If a person is drinking beer,
then the person is over 18.
Beer Coke 22 16
Weâd have to look at the Beer card to make sure the beer drinker
wasnât under 18, and weâd have to look at the 16 card to make sure
the young person wasnât drinking beer.
111. If a card has a vowel on one
side, then it has an even
number on the other side.
A B 4 7
The same thing is true for the vowel/number hypothesis.
112. Not Wired That Way
We humans evolved to deal well with speciďŹc situations--we learn well
from example, and we can estimate numbers well. But there are lots
of modern situations where we havenât evolved well. Falsifying is an
example: we learn from example by enjoying conďŹrmation. But to test
a hypothesis, we must seek out counterexamples and our brains just
arenât good at that unless itâs a situation weâre already very familiar
with.
113. For example, that exponential growth we saw earlier--thatâs
compound interest. Everyone raves at you to put your money in a
bank account with compound interest, thatâs why--every dollar works
for you, even the ones you earn in interest, and they all just keep
working for you making more dollars. But weâre lousy at estimating
compounding.
Hereâs another example.
114. Question
Linda is 31 years old, single, outspoken, and very bright. She
majored in philosophy. As a student, she was deeply
concerned with issues of discrimination and social justice,
and also participated in anti-nuclear demonstrations.
Which is more probable?
A. Linda is a bank teller.
B. Linda is a bank teller and active in the feminist
movement.
115. Question
Linda is 31 years old, single, outspoken, and very bright. She
majored in philosophy. As a student, she was deeply
concerned with issues of discrimination and social justice,
and also participated in anti-nuclear demonstrations.
Which is more probable?
A. Linda is a bank teller.
B. Linda is a bank teller and active in the feminist
movement.
Weâre bad at dealing with probabilities and subsets. Many people will
tell you that B is more likely, because Linda sounds likely to be a
feminist. But if you look at the probabilities, fewer people are bank
tellers AND feminists so the probability of B is lower than the
probability of A.
116. Practice
Because it doesnât come natural, practice it. Look for situations where
you can test one of your hypotheses. Think hard to make sure youâre
observing the right things!
118. because we, as a society, donât understand failure. If youâre like me,
you worry that when you fail youâll be thought less of. Maybe youâll be
ďŹred, maybe youâll lose friends, maybe itâll be harder to get the next
pay rise. Weâre taught from an early age to seek success but that
121. failure = TEH AWESUM!
Failure is awesome! Failure is fantastic! Failure is how we learn!
Youâre laughing. Thatâs because youâre conditioned to believe failure
is bad. Snap out of it!
Of course, failure is only good ...
122. Donât be afraid to fail,
so long as you learn
if you learn from your mistakes. The problem is that weâre so afraid of
failure, we never inspect our failures to learn. Weâve become a Far
Side strip.
123.
124. failure without learning
=
bad
Itâs failure without learning thatâs bad. You canât succeed unless you
take risks. Theyâre not called risks for nothing--with any risk you take
thereâs a chance of failure as well of success. But failure can teach you
things and prepare you for success.
Donât just take my word for it:
125. âIt is on our failures
that we base a new and
different and better
success.â
-Havelock Ellis
126. âFailure is instructive.
The person who really
thinks learns quite as
much from his failures
as from his successes.â
-John Dewey
127. âIâve missed more than 9,000
shots in my career. Iâve lost
almost 300 games. 26 times
Iâve been trusted to take the
game winning shot and
missed. Iâve failed over and
over and over again in my life
and that is why I succeed.â
-Michael Jordan
129. âIt is a mistake to suppose that
men succeed through success;
they much oftener succeed
through failures. Precept,
study, advice, and example
could never have taught them
so well as failure has done.â
- Samuel Smiles
130. âI have not failed. I've
just found 10,000 ways
that won't work.â
-Thomas Alva Edison
131. âSuccess consists of
going from failure to
failure without loss of
enthusiasm.â
-Winston Churchill
133. fail small
Itâs simple: fail small. Donât start by failing at huge things. Huge
failures have huge costs and itâs not always possible to get a
correspondingly huge amount of learning from them. Fail small: learn
a little at a time.
134. fail before itâs big
This doesnât mean you canât tackle big projects, just that you tackle
them small bit by small bit, and always keep your eye open for the
situation where you can say âthis isnât going to work, the project has
failed but hasnât yet cost us a huge amountâ. Learn as much as you
can for the lowest cost that you can.
135. the gate
someone recently explained it to me like this: projects go through
phases: scoping, requirements, development, testing, maintenance,
etc. At each phase, thereâs a gate: we could kill it after scoping, or we
go onto requirements. We could kill it after the requirements, or we go
on to development, etc. All too often, though, we view a project as a
race through the gate: we have to make it through that gate, not get
the project killed.
Thatâs so backwards. The projects that fail small are wonderful! Just
ask anyone who has been on a project that failed big!
136. learn from the failures
So fail early and fail often, but always learn from the failures. Thatâs
my thesis. Thatâs the dierent approach to something unpleasant. I
want to leave you with a simple catchphrase that sums up how to build
your own personal feedback loop, three short words that I hope you
will take back and try to apply in your life, in your work, and in your
works: