Jeff Bezos talking about entrepreneurship – 2005

Jeff Bezos, telling the story of multiple innovations and companies during a business school conference. Focusing on learned helplessness.
Some of the companies and ideas covered:
– The wiffle ball
– Liquid paper
– WD-40
– Windshield wiper
– Toilet paper …

Transcript:

I want to talk a little bit about how we
00:02
think about innovation at amazon.com and
00:05
give you a couple of examples from the
00:07
world this is the wiffle ball and the
00:10
guy his name is David Nelson Mulaney and
00:15
in 1953 he took a Cody perfume package
00:20
and out of frustration because his son
00:22
had broken a window with a regular
00:24
baseball and created the wiffle ball his
00:27
son named it this woman is named Betty
00:33
Nesmith Graham and she is you will know
00:38
of her accomplishment in just a moment
00:40
if you don’t recognize her name but she
00:43
was an executive assistant in the time
00:46
when typewriters transitioned to film
00:50
cartridges from ink ribbons and and she
00:55
was very annoyed she wasn’t an
00:57
especially was a great executive
00:59
assistant but she was a rather poor
01:00
typist and she was annoyed by her
01:03
inability to erase her mistakes with the
01:06
film cartridges so she invented liquid
01:09
paper which after let’s see by night she
01:14
did this in 1956 in 1975 she sold liquid
01:18
paper to the Gillette corporation for
01:20
forty seven point five million dollars
01:22
so she did pretty well with her
01:26
innovation and by the way it was just
01:29
white paint so
01:35
there was no very very clever person
01:38
because there was no you know special
01:42
chemistry involved here this this was
01:46
not a high-tech solution and but
01:50
extraordinary observation of a problem
01:54
and a solution now so sometimes people
01:59
see the problem and the problem is
02:01
really annoying them and then they
02:03
invent a solution sometimes you can work
02:06
this from the backwards direction and
02:09
and in fact in high tech I think a lot
02:11
of the innovation sometimes comes from
02:13
this direction you see a new technology
02:15
or there’s something out there some new
02:16
understanding in the world and you work
02:18
backwards from a solution to find the
02:21
appropriate problem carbon dating is a
02:24
little bit like this so you know they’d
02:26
always wanted this solution to be able
02:28
to date things but until they really
02:30
understood the radioactive decay that
02:34
wasn’t possible to come up with that
02:35
solution so some new understanding some
02:38
new concept enters the world and that
02:40
there are a lot of things like that in
02:42
high-tech now their critical ingredient
02:44
for anybody who would be innovative is
02:47
persistence and there are many examples
02:49
of people being persistent in the world
02:52
in order to make something work this is
02:57
this is one of my favorites wd-40 was
03:02
this a small team a small company of
03:05
three people got a government contract
03:08
to develop a some kind of coating that
03:14
they could put on the skin of Atlas
03:17
missiles while they sat in their silos
03:20
to keep them from rusting and they they
03:26
worked on this for a long time to find
03:29
this right compound that would do this
03:31
job and and finally did it now it turned
03:34
out that the Atlas missile market was a
03:36
fairly small one
03:38
so that the company wasn’t really able
03:41
to make much progress now what does
03:44
wd-40 stand for it stands for water
03:48
displacement 40th to 10 and that’s the
03:52
name right out of the lab book when they
03:55
did all this work to invent wd-40 so
03:59
when they properly recognized to the
04:02
size of the Atlas missile market they
04:05
started to think about a different
04:07
business plan and the company was called
04:08
the rocket Chemical Corporation and it
04:11
was their only chemical it was their
04:13
only product and so about 10 years later
04:16
they renamed the company the wd-40
04:18
company and you know the rest is history
04:21
one of the most pernicious obstacles to
04:26
invention is learned helplessness so
04:29
people get the problems that you
04:33
encounter
04:34
if you encounter them for a long enough
04:36
duration humans and actually biology in
04:40
general but humans in particular so
04:42
remarkably adaptive that we pretty soon
04:44
see right through the problems we don’t
04:47
even notice them and so great inventors
04:50
and people who they’re very good at you
04:52
know ordinary things bother them you
04:55
know they wake up every morning and take
04:57
a shower and they think this shower is
04:58
terrible you know and but if but it’s
05:01
very difficult to do that
05:03
to kind of push through this learned
05:04
helplessness you get used to something
05:06
here’s an example this woman is named
05:09
Mary Anderson and Mary invented the
05:12
windshield wiper
05:13
let me get Marius paper out here oh well
05:17
I don’t know the story well enough but
05:20
Mary was this was around 1913 and cars
05:27
didn’t have windshield wipers in fact
05:28
when it rained people would pull off to
05:30
the side of the road and they would take
05:33
a rag and clean the windshield off and
05:36
they’d drive another mile and and repeat
05:39
this process and they thought nothing of
05:41
this and Mary you know looked at this
05:45
and she said that it’s ridiculous and
05:47
that there must be a better solution
05:48
than stopping every so often and
05:50
cleaning off your windshield with a rag
05:52
and so she invented the windshield wiper
05:55
and she was roundly criticized that this
05:58
was a ridiculous thing and people had
05:59
all sorts of reasons why it wouldn’t
06:01
work primarily that it would be a huge
06:03
distraction for drivers to have this
06:05
thing going back and forth in front of
06:07
them and but she persisted and and and
06:12
this happened very quickly
06:13
within ten years of her invention of the
06:16
windshield wiper
06:16
they were standard equipment on all cars
06:18
so with one of these things that you
06:20
know and actually even though people
06:22
pooh-poohed it once they tried it
06:24
they’re like no it’s stopping not having
06:26
to stop every mile with a rag it’s
06:28
actually a pretty good idea and so Mary
06:32
definitely got the last laugh and this
06:35
is one of my favorite inventions and a
06:37
fantastic example of learned
06:39
helplessness as you know toilet paper
06:41
was invented and people said the year is
06:45
in the mid 1800s guy named Joseph Getty
06:51
1857 invented toilet paper this is the
06:54
first toilet paper it’s hard to read but
06:57
it says many people have wooed their own
06:59
destruction physical and mental by
07:02
neglecting to pay attention to ordinary
07:05
matters
07:09
rarely an advertisement just one find
07:12
such a profound statement and so but
07:19
this is one of those things people
07:21
didn’t know they needed toilet paper
07:22
until Joseph guiity invented it and then
07:26
very shortly thereafter they began to
07:28
realize that it didn’t know how they
07:29
lived without it and it’s certainly
07:32
difficult to consider going back now
07:39
let me talk about it let me make one
07:41
very important to disclaimer right up
07:43
front that amazon.com does not claim any
07:46
inventions nearly as important as toilet
07:48
paper this is this is clear but we have
07:53
been working hard on innovation and
07:54
we’ve done a bunch of different stuff
07:56
and I want to I want to and we have been
07:58
for the whole nine and a half years of
08:00
our of our history and I want to talk a
08:02
little bit about that so first of all
08:04
it’s sometimes useful to take a little
08:06
time travel trip and move backwards this
08:09
is what amazon.com looked like in July
08:12
of 1995 I think this is actually a
08:13
screenshot from August and and it you
08:19
know it’s changed a lot since then this
08:23
is by the way I wrote all of this HTML
08:26
myself no really please hold your
08:29
applause there’s no need
08:31
really you’re you’re embarrassing me and
08:36
the this notice some of the very obvious
08:39
problems that this has for example
08:41
there’s no search box on this page the
08:45
search box is a click away you have to
08:48
click on I think you know 1 million
08:51
titles will take you there or search
08:53
Amazon coms a million title catalog so
08:57
this is what the website looks like
09:00
today and we invest about 250 million
09:05
dollars in the last year but about 20
09:07
million hours a year for several years
09:08
now 250 million in the last year on
09:11
technology and content development and
09:13
so the website should be better
09:17
sorry significant sums of money the
09:19
company has generated over the last
09:21
three years nine hundred and fifty
09:23
million dollars in free cash flow in the
09:26
context of investing six hundred and
09:28
fifty million dollars in technology in
09:30
that same period of time so we’ve we’ve
09:33
you know we’ve gotten smarter about a
09:35
few things the search box is now you
09:38
don’t have to click it’s not one click
09:39
away it’s right there and there are a
09:42
bunch of other things going on here but
09:45
but the point is the the the you lose
09:49
track in the moment of how much
09:51
innovation there is over a nine year
09:53
period one of the things that is very
09:56
important at Amazon that we’re
09:58
constantly trying to do is reject this
10:01
either-or thinking that really is at the
10:03
heart one of the big impediments to
10:06
innovation I’ll give you a simple
10:09
example which is in customer service
10:11
it what the what we want to do is make
10:16
the service better so that we can have
10:18
the ACSI rating that you’re talking
10:21
about of eighty eight which we are very
10:22
very proud of and that’s a very they do
10:26
they have a panel of like sixty thousand
10:28
customers that they pull so it’s a very
10:30
statistically significant kind of result
10:32
and something that we we do you know
10:36
it’s it’s a metric that people have
10:38
worked very hard not specifically on
10:40
that metric but on the things that lead
10:41
to it so it’s a great for us it’s
10:45
something that we you know are very
10:47
proud of this we get a lot of customer
10:50
contacts and one of the things that we
10:53
would like to do is minimize customer
10:55
contacts and in order to save money so
10:58
they can offer people lower prices but
11:00
still offer great customer service and
11:02
so how do you solve that problem well
11:04
you do it a couple of ways one is you
11:07
eliminate defects and eliminate root
11:09
causes of defects because if customers
11:11
are contacting you they probably didn’t
11:13
want to I think actually the number four
11:16
reason we get contacted it is to say
11:18
thank you and so that’s kind of nice we
11:21
don’t need to eliminate those contacts
11:22
but most of the contacts that we get or
11:26
not to say thank you
11:27
they you know sometimes come in all
11:29
capital letters
11:32
you know and so if you can eliminate the
11:35
defects you’re gonna save money because
11:38
you’re not going to have to handle the
11:39
customer contact at all and you’re gonna
11:40
improve the customer experience so
11:42
there’s no trade-off in reducing defects
11:44
and the more the closer you reduce
11:47
defects to the root to the root cause
11:50
the the more that’s going to be true
11:52
you’re going to both improve the
11:53
experience and save money no trade off
11:56
but the second thing that we did is
11:58
build customer self-service and this is
12:03
our you know this is the page where you
12:05
go to to do customer self-service and we
12:08
let people cancel their own orders we
12:10
let people you know consolidate orders
12:14
and do all sorts of things since change
12:16
their shipping address and so on and so
12:18
on and check on the status and and when
12:21
is my thing going to arrive etcetera etc
12:23
and that has really empowered customers
12:26
and what you find is when other in other
12:28
industries like the hotel industry
12:30
Embassy Suites often wins the best
12:33
customer service award in in hotels when
12:37
they pull heavy hotel users and of
12:40
course they the the secret to Embassy
12:43
Suites winning that is that don’t
12:44
actually have any service it’s all
12:46
self-service and so people when they’re
12:49
serving themselves think the service is
12:50
terrific and so you have the buffet out
12:56
and if the buffet is out and the food is
12:58
free so you just take the food and you
13:00
leave this is just not much that can go
13:02
wrong and and and that that really helps
13:06
so that’s something that that we’ve
13:08
worked hard on it it’s an example of
13:10
rejecting either-or thinking and we have
13:12
been able to reduce contacts over the
13:14
last seven or eight years by eighty-five
13:16
percent which is a huge reduction and
13:18
that that level of that driving down
13:21
contacts per unit continues the other
13:25
thing you have to do if you’re going to
13:27
innovate and that we try to do at Amazon
13:29
accom is maximize the rate of
13:32
experimentation so and if you’re going
13:35
to do that you have to make sure that
13:37
your cost of doing experiments is low if
13:41
your cost of doing experiments is high
13:44
then you’re only going to get to do I
13:46
think the company is only going to do a
13:48
few experiments per year and if you’re
13:50
doing a few experiments per year the
13:53
you’re going to have to do some kind of
13:55
global prioritization to see which kind
13:57
of experiments you can do and that’s
13:59
going to have a couple of downsides one
14:01
is you’re not going to do much
14:02
experimentation but the second is that
14:04
the best most inventive people and your
14:08
organization are going to become
14:10
frustrated because they’re gonna have to
14:12
get ultimately if we can one can only do
14:14
three experiments a year then if you
14:16
wanted to do an experiment you’d have to
14:18
get my permission to do it and I’m not
14:20
scalable so so that is going to put a
14:23
huge impediment in the way of invention
14:26
and if you want on the other hand if you
14:29
want to make experimentation that
14:31
everybody can do it and it doesn’t take
14:33
you know sort of the you know the the
14:34
institutional apparatus doesn’t have to
14:36
approve every experiment then you need
14:40
to make the cost of experimentation low
14:42
if you’re still going to make it a
14:44
reasonable thing to do and and and not
14:46
irresponsible
14:47
you can’t just let if the cost of
14:48
experimentation is high you can’t just
14:49
say yeah this is gonna cost hundred
14:51
million dollars to do this experiment
14:52
but go for it so the getting that cost
14:56
low is is key and one of the things that
14:59
we’ve done over the years has put a lot
15:01
of effort into making sure that our
15:03
infrastructure are in the framework that
15:05
operates our website in other parts of
15:07
our business really make it easy for
15:09
people to do experiments in a
15:11
self-service way without coordination
15:13
with the institutional apparatus one of
15:16
the great things about operating on the
15:17
web is that you have you have so much
15:22
the ability to collect data and all many
15:25
things that in the physical world
15:26
because the cost of experiment would be
15:28
high you have to use judgment to decide
15:30
those things you have to use intuition
15:32
to judge those things because you can’t
15:33
do the experiment at realistic cost so
15:36
you end up having to argue over it and
15:38
on the web you can often just know the
15:40
answer so you know a simple example this
15:43
was a control this page and you’ll see
15:45
the difference here a new line is going
15:47
to pop up one it delivered tomorrow May
15:50
5th order in the next two hours and 53
15:52
minutes and choose one day shipping at
15:53
checkout so this is experiment we did
15:55
about a year ago and
15:57
percent of the customers saw it this way
15:59
and 50% of the customers saw it this way
16:02
and you can just literally measure the
16:05
impact that this has on all sorts of
16:07
important metrics you know first and
16:09
foremost sales so you can see this
16:11
telling people that they should order in
16:13
the next two hours and 53 minutes
16:15
increase sales or decrease sales and and
16:19
and and and and buy how much is this
16:21
worth doing and is it worth pushing
16:24
harder on and so on and so on so being
16:26
able to do that kind of experiment and
16:28
not having it be a big deal is really
16:32
wonderful here’s another example an even
16:34
simpler example this is you know
16:36
customers who bought this also bought
16:38
something that we’re well known for and
16:41
this has been improved over the years
16:43
dramatically here’s a small example this
16:45
was tested – oh that’s not the one I’m
16:46
showing here is we added series to it so
16:49
that you’re you know you’re here’s an
16:52
actual individual DVD but here’s series
16:54
from different things that was a
16:56
successful experiment so it stayed one
16:58
of the experiments that we ran with this
17:01
customers who bought this also bought
17:02
was it seemed very intuitive that if
17:05
this works well then adding images might
17:08
make it work even better so we added
17:10
images turned out that it didn’t make it
17:13
work better it made it work worse and
17:15
even though that’s a little
17:17
counterintuitive I mean it to me I can
17:19
see it both ways but that’s the that’s
17:22
kind of the whole point here you know
17:24
you have you have your a priori guesses
17:26
but then you actually validate them with
17:28
real data
17:30
here’s another try at images since the
17:33
this the the when this one didn’t work
17:35
it was oh maybe it’s too clunky maybe
17:37
it’s pushing things down below the fold
17:39
so this one this one was tried but it
17:43
also didn’t work as well so we stuck
17:46
with this one the other thing we try to
17:51
do at amazon.com is have a customer
17:54
centric obsession instead of a
17:57
competitor obsessed a point of view and
18:01
there are by the way that’s not clear in
18:04
all companies at all times that this is
18:06
the right strategy competitor focus and
18:08
in fact close following strategy
18:11
can be very effective in certain
18:13
business environments and there’s
18:14
certainly nothing wrong with it with the
18:16
with with that but in our industry and
18:19
and given the DNA that we have inside
18:21
the company the kind of people that
18:23
we’ve attracted over time our people who
18:25
like to invent people who like to
18:26
pioneer so you know we’d have we have
18:29
the wrong culture to be close followers
18:32
so there’s part of this that you know we
18:33
are what we are and we and and and and
18:36
we kind of have to stick with that and
18:38
make the best of that I actually think
18:40
in our space this is also very handy
18:43
because the rate of change in the online
18:46
world is so rapid that closed following
18:48
doesn’t work as well as it might in a
18:51
more stable industry that the change is
18:55
more slowly over time but one of the
18:57
things about these fast changing
18:59
industries is I often get asked by by
19:04
audiences if I take you native period
19:07
you know what’s gonna change in five to
19:10
ten years one of the questions that I
19:12
rarely get asked and I think is at least
19:14
as important of a question and maybe in
19:15
some ways an even more important
19:17
question is what is not going to change
19:20
over the next five to ten years and
19:22
that’s another reason why customer
19:24
centric strategies can be so powerful
19:27
because competitors focus strategies
19:30
have to change law the competitive set
19:32
changes so rapidly over the last ten
19:34
years our competitive set has changed
19:35
and the technologies have changed so
19:37
many things change but the basic drivers
19:40
of our business the core things that
19:42
customers want do not change you know we
19:46
know that what customers want our
19:47
selection low prices and convenience and
19:50
I guarantee you that ten years from now
19:54
they’re still gonna want selection low
19:57
prices and convenience we are not going
19:59
to wake up ten years from now and have
20:01
customers say this is all well and good
20:02
I love this thing Amazon to come but
20:05
could you make it a little less
20:06
convenient it’s just not going to happen
20:09
and so so when you work on low prices
20:13
you have to figure out inventive ways to
20:15
do that and there are inventions a
20:17
difference at different granularities
20:19
there are the a lot of the most
20:21
important invention happens at the
20:23
finest granularity
20:24
so incremental improvements in unit
20:27
productivity are critical so can you you
20:30
in in in our fulfillment centers you
20:33
know which are about so we have about
20:35
seven million square feet of fulfillment
20:37
centers each you know the each one is
20:38
about 700,000 square feet these are huge
20:41
buildings you know kind of corner to
20:43
corner there half a mile on a side and
20:44
so how do you pick you can order any two
20:48
items we have a million more than a
20:50
million items in stock which to put that
20:52
in context you know a Barnes & Noble
20:55
superstore has about a hundred and
20:56
twenty-five thousand items a Costco or
20:59
Sam’s Club has about four thousand items
21:01
a big big-box electronics store has
21:05
about seven thousand items so over a
21:08
million in stock items is a lot of items
21:10
and you can order any two of those items
21:14
and before 6:30 p.m. Eastern Time if you
21:18
if you could if you order them before
21:20
6:30 p.m. Eastern Time we will ship them
21:23
that same night so and those two items
21:25
might be on opposite sides of this big
21:27
fulfillment center so we have to there’s
21:29
a tremendous amount of work and very
21:31
sophisticated computer algorithms that
21:34
go into optimizing the pick paths so
21:36
that we can pick those items and get
21:38
them into the same box at low cost it’s
21:41
easy to do that if you’re not worried
21:42
about optimizing cost it’s very
21:44
difficult to do that if you are
21:46
concerned about optimizing cost so those
21:48
kinds of incremental improvements taking
21:50
those pick path algorithms and making
21:52
them modestly better and doing that
21:54
every day and and the key to doing that
21:57
kind of invention is to make sure that
21:59
you have small separate empowered teams
22:01
that aren’t subject to a bunch of
22:04
dependencies in the rest of the
22:05
organization they know what they’re
22:06
trying to achieve and they go about
22:09
making those incremental improvements
22:11
day in day out month in month out they
22:14
have to be able to answer the question
22:15
which in it’s amazing you’ll find that
22:18
this is very common in in in the
22:20
business world that people often cannot
22:22
answer the question how do you know
22:24
whether you’re getting better or not at
22:26
this particular thing and the picking
22:28
the granularity of where you ask that
22:29
question is the trick I mean it’s very
22:31
easy you know on the distant outputs if
22:34
you look at you know something like free
22:36
cash flow or revenues
22:38
those kinds of things people know
22:39
whether they’re getting better but those
22:41
things are not directly actionable you
22:42
can’t you know assign some team and say
22:45
look go drive free cash flow it’s just
22:48
too easy and it’s it’s – it’s easy to
22:51
command them to do it but they won’t be
22:53
able to do it I mean it’s a useless kind
22:55
of things to ask for and then on bigger
22:59
granularities so one of the things that
23:01
we also observe is that used products
23:04
can have much more value proposition for
23:07
customers so we invite third-party
23:08
sellers to compete against us on our
23:11
prime real estate which is our detail
23:14
pages that’s an invention that also not
23:16
only adds a lower price but selection –
23:18
so you go through all these things
23:20
selection one of the things that we’ve
23:23
done recently to increase Bret this
23:25
lecture at Amazon is again invite these
23:27
third-party sellers in this works in all
23:29
the categories where we operate and in
23:32
some of the newer categories like
23:33
apparel it’s a huge part of the strategy
23:38
convenience the most recent thing that
23:40
we’ve done to increase convenience is
23:42
for our heavy customers which is
23:45
something called Amazon Prime and Amazon
23:48
Prime is a we do only launched this a
23:52
week ago and it is a $79 flat fee
23:57
membership you pay $79 a year and you
24:00
get two day shipping for free and that
24:04
is a the idea here is to take what might
24:10
be an indulgence for people getting next
24:13
day air shipment or second day area next
24:15
day next day air is just three dollars
24:17
and ninety nine cents an item once you
24:19
remember and take something that they
24:22
might perceive as an indulgence
24:24
transformed into a fixed cost and then
24:27
once people can view it as a fixed cost
24:28
so well why not have two day shipping on
24:31
my items and so for a certain segment of
24:35
customers this is a very attractive
24:37
proposition and it’s for it specifically
24:40
for that segment of customers where
24:42
convenience is most valued so you know
24:46
there’s another segment of customers who
24:48
are who are completely happy to
24:51
wait you know eight business days for
24:54
the products they’ll use our super saver
24:55
shipping offer which is free all they
24:57
have to do is meet a $25 order hurdle so
25:01
they’re perfectly willing to save a few
25:03
things in their shopping cart until they
25:05
can meet that order hurdle they sort of
25:06
managed that it creates a little bit of
25:08
mental overhead for them but they’re
25:10
they’re okay with that and so that
25:11
becomes a great value proposition for
25:13
them and there’s another set of
25:15
customers who think about this
25:17
completely differently they hate the
25:19
fact that there’s this twenty five
25:20
dollar order hurdle they want to be able
25:23
to buy you know one thing when they see
25:25
it get it done have it common they want
25:27
it to come fast they don’t want to have
25:28
to think about it they don’t want to
25:30
have to try and optimize shipping costs
25:32
and in fact one of the things that
25:33
happens for that kind of customer too is
25:35
they create there’s a kind of cognitive
25:38
dissonance or guilt that gets set up in
25:40
their head you know it they kind of like
25:43
since they can get the shipping for free
25:45
they feel like they should be getting
25:47
the shipping for free and so even though
25:50
you know maybe for them that’s not
25:53
actually a rational economic
25:55
consideration this is another
26:01
convenience oriented feature that’s been
26:03
on the website for a couple of years now
26:04
instant order update and this reminds
26:08
you that you’ve already bought something
26:09
and we get feedback from customers all
26:11
the time this is something that we
26:12
measured tested very very carefully and
26:15
it reduces sales there’s just no
26:18
question about it that in the short term
26:20
is statistically significant way reduces
26:23
sales and we so we overruled that with
26:26
judgment and said even though we can’t
26:29
do long duration longitudinal tests in
26:31
time like like a drug company would do
26:34
to test for toxicity and that kind of
26:36
thing where they do like a 10-year test
26:37
that those tests are just so expensive
26:38
to do we’re gonna make it we’re gonna
26:41
make a intuitive bat that having this
26:44
thing that reminds you that you already
26:47
bought something will actually will be
26:49
so beloved by customers that they’ll be
26:53
one of the reasons they like amazon.com
26:55
and so and and in fact we do get
26:58
tremendous anecdotal response back from
27:01
customers saying they love this feature
27:02
little you know thank you note saying
27:04
about to buy the same music CD that I
27:06
bought a year ago and this is very real
27:08
I don’t know you know some people are
27:10
more organized and have better memories
27:12
than others I definitely fall prey to
27:15
this in fact just yesterday I was
27:17
reading a blog that mentioned a walt
27:21
disney DVD on the back in the late 50s
27:25
or early 60s I can’t remember where and
27:26
Walt Disney was very interested in space
27:29
and he hired Wernher von Braun and did a
27:33
whole series of television specials
27:36
about the exploration of space
27:38
starring Wernher von Braun and these
27:39
things have been collected on a DVD and
27:41
I read about this on this blog and I
27:45
thought my god this is so cool
27:47
I have to buy this and so I went to
27:50
amazon.com actually they were an
27:51
associate and so they had a link handy
27:53
for me and I clicked right through to
27:54
the detail page and there was my instant
27:56
order update I bought it a year ago and
28:01
I’ve never watched it I bought it and I
28:02
must have like you know put it somewhere
28:04
and so now I’m gonna go search for it
28:09
this is a real one – drew who’s sitting
28:12
down here and told me that he almost
28:14
bought these earrings for his wife a
28:18
second time and it’s a true story and so
28:24
I think in that case we that was really
28:27
a valuable service another convenience
28:34
oriented feature is giving people high
28:36
quality product information right on the
28:40
detail page and one of the great things
28:42
about this is an exempt this is a piece
28:44
of the detail page for the Segway and it
28:46
explains in great detail how this Segway
28:48
works and one of the great things about
28:52
the online model that we use is that a
28:57
lot of our best customer experience is a
29:00
fixed cost so if you have enough scale
29:02
if you have you know in so in the last
29:05
12 months we have these you know 47
29:09
million customers who have purchased
29:11
from amazon.com and if you have 47
29:15
million customers you get to amortize
29:17
that the cost of all this content across
29:21
that very large customer base and so
29:23
there’s a sense in which you can have
29:25
your cake and eat it too if you look in
29:27
the physical world at physical world
29:29
retail stores the kind of high-touch
29:33
customer experience stores cannot have
29:35
the lowest cost and the reason for that
29:37
is that those stores the things that
29:40
generate high-touch customer experience
29:43
in the physical world are variable cost
29:46
things so when you double your sales you
29:48
double those costs our model when you
29:51
look at this kind of content or any of
29:53
the software features that we provide at
29:55
instant order update it’s you know
29:57
somebody has to develop that and
29:59
maintain it and make it work but what it
30:02
would cost us the same to do that
30:03
development if we had a million
30:04
customers as it does if we have 47
30:07
million customers so that that that that
30:11
aspect of our model that we get to
30:12
transform customer experience into a
30:15
fixed cost instead of a variable cost is
30:17
really a key it’s underappreciated
30:20
aspect of of what we do now this is also
30:27
for any company that wants to be
30:29
innovative absolutely critical thing to
30:32
do which is to not be distracted when
30:34
people tell you that this is stupid and
30:36
isn’t going to work and people are
30:39
well-meaning and entrepreneurs and and
30:42
and and I’ve used that term in the most
30:44
broad way you know people who are
30:45
building people who are inventing have
30:49
to have this combination of stubbornness
30:52
and flexibility and the trick of course
30:55
is knowing which to use win but but what
31:00
you cannot do is be distracted by the
31:04
outside world you have to stay heads
31:05
down focused on customers and not let
31:08
external events distract you and Amazon
31:11
is a very good case of this this is
31:16
December of 1998 but and actually this
31:20
is a phrase that and I think Rob is in
31:23
the audience but this is not you Rob you
31:25
didn’t write this
31:29
but this is the this is a case of in
31:35
1997 forced to research coined this term
31:39
and and Barnes & Noble had just launched
31:42
their online store and I thought that
31:47
Forrester had a very good point they
31:50
were saying and you know Amazon is this
31:52
two year old company they’ve had this
31:54
nice to your run but you know so now the
31:57
the big the big gorilla has come to town
32:00
and at the time we had annual revenues
32:03
of sixty million dollars a year we had a
32:05
hundred and twenty five employees and
32:08
Barnes & Noble had you know thirty
32:10
thousand employees and at that time I
32:12
don’t know three billion dollars a year
32:13
in sales and so it didn’t seem like an
32:16
even fight and and they coined this term
32:19
Amazon toast and and it was widely
32:23
picked up and you know the the and it
32:26
did bother the there’s one feedback loop
32:28
here that has to be managed which is you
32:30
have to make sure that the the that you
32:33
have some all-hands meetings and that
32:35
you communicate to employees what this
32:37
you know should they be worried about
32:38
this or shouldn’t they be worried about
32:39
this and in our case I thought this was
32:41
a very simple situation because the I
32:45
didn’t think there was anything we could
32:46
do about what our competitors were going
32:47
to do and so you know we did have the
32:50
All Hands meeting and I asked all of our
32:52
125 employees to be terrified and to
32:55
wake up with their sheets drenched in
32:58
sweat every morning and I did I
33:01
carefully specified sweat and and but
33:06
that they should be afraid not of our
33:08
competitors but they should be afraid of
33:10
our customers because our customers are
33:12
the only ones who are going to ever give
33:14
us money and so that helps it actually
33:18
helps have that kind of customer focus
33:21
if your because it helps you stay heads
33:22
down you don’t get worried about your
33:24
circular competitors are going to do
33:25
what your competitors are going to do
33:26
you can watch them you can learn from
33:29
them but you can’t let them set your
33:32
strategy you have to do what you’re
33:34
gonna do and you also can’t let the
33:35
media set your strategy and and but you
33:39
do see that happen you can’t
33:41
wallstreet set your strategy either this
33:44
is a year after Amazon ptosis is 19 May
33:48
of 1999 and this I have always been
33:54
amused by this but I’m uncomfortable
33:58
being misunderstood but this I have to
34:01
tell you my mom really hated this this
34:03
really bothered her she was so pissed
34:08
off about this this just that just
34:12
didn’t like that caricature it really
34:13
bothered her so but anyway you just have
34:17
to ignore those outside influences you
34:21
can listen and kind of try to see if
34:23
there’s any any any kernels of truth
34:25
that you can take and adapt there’s no
34:28
reason to have a siege mentality but
34:30
that’s those are the kind of reactions
34:32
you see you either see that people enter
34:33
this sort of siege mentality with
34:35
respect to external factors or they or
34:40
they’re too responsive to these external
34:41
factors and the reality is you have to
34:45
take it with a grain of salt and you
34:47
know reexamine your strategy make sure
34:49
you really believe in it it’s the right
34:50
one and stick to it that’s the only way
34:52
you can invent because invention always
34:56
leads you down paths that people are
34:58
gonna think are weird it’s sort of
35:01
obvious this is this is this is that a
35:06
CSI score and it’s actually four years
35:09
running that we’ve had this highest
35:10
score so one of the things I mentioned
35:16
is the carbon dating and what happens in
35:18
the online is you in the well UNTAC in
35:23
general is you get you get so much
35:25
change and you have to sometimes look at
35:27
the change and see what you can invent
35:29
so why is things as Moore’s Law which is
35:31
remarkable and it’s variants for
35:33
bandwidth and and and storage but if you
35:37
look at disk space costs as against
35:39
example
35:40
disk space is 30 times cheaper today
35:42
than it was five years ago and so you
35:45
know that’s just what happens when
35:46
things get twice as cheap every year for
35:48
five years and humans aren’t good at
35:51
thinking in the kind of you know
35:54
exponential terms but this is really a
35:58
big deal and if you if you were to
36:00
picture a physical store it created that
36:02
when things change that rapidly it
36:05
actually creates a problem you have to
36:06
figure out what are we gonna do with all
36:08
that disk space because it’s not you
36:10
know you can’t just pocket the cost
36:11
savings you don’t get enough value out
36:13
of that you have to say well if we can
36:14
use 30 terabytes of disk and that’s a
36:17
reasonable thing to spend money on now
36:20
what could you do that would actually
36:22
benefit customers that use this 30 but
36:24
terabytes of disk and the one of the
36:29
things that we did was search inside the
36:31
book which uses about 30 terabytes of
36:33
disk because we keep the full images of
36:36
all the pages and we now have over
36:38
200,000 different books that are
36:41
completely searchable and you can view
36:42
the actual images of the books and and
36:46
and that it kind of the good news about
36:49
that is it’s very beneficial to
36:50
customers and it soaks up a lot of this
36:52
now almost free disk space that you can
36:54
buy out in the world
36:56
another example is a nine let me just go
37:05
here which just launched again about a
37:09
week ago get a nine yellow pages we do a
37:13
search on optical this regular mouse
37:18
works too it’s a lot easier to use and
37:23
one of the innovative things that the a
37:26
nine team did for their yellow pages is
37:30
take 20 million photos of over a million
37:33
different businesses in ten different
37:35
cities and they did that in a very cool
37:38
way I mean this in something you know
37:40
here if you go down the list here this
37:42
is optical in Seattle but you know the
37:47
map changes the little numbers
37:48
highlighted as you go over which think
37:49
it’s a nice user interface to touch but
37:51
if we go down to optical illusions Inc
37:57
click on that
38:00
takes you to a detail page for optical
38:05
illusions Inc and you can kind of stroll
38:09
along the street here so here is you
38:16
know and you can zoom in on this picture
38:20
and next way this is kind of a cool
38:23
picture because this is you can see in
38:24
the reflection of this window that’s the
38:27
vehicle that we took the picture from
38:29
and I also I like the way this guy is in
38:34
motion it’s just kind of fun and that
38:37
you can’t barely see her but there’s a
38:39
woman behind him you can just sort of
38:42
see her see her shoes we have it’s only
38:48
been up a week and you know but if you
38:49
take 20 million 20 million photographs
38:52
of a million different businesses we’ve
38:55
already gotten lots of contacts from
38:56
people you know who have found
38:59
themselves in these pictures and if they
39:03
request to be removed we remove them but
39:06
but this is this is a pretty cool use of
39:11
technology and again it’s looking at
39:14
what’s changed in the world because one
39:16
of the most interesting things about
39:17
this is is how they did it and you can
39:21
keep going by the way I can click here
39:23
and it will take me you know down the
39:25
next street I can just keep walking down
39:28
this street okay so how did they do this
39:34
should I go to the back to the
39:37
PowerPoint alright here’s the truck
39:43
they’re there more than one of these
39:46
that is a a camera a digital camera
39:51
mounted on the top of the truck and when
39:54
they first started doing it their first
39:55
tests were done with not with SUVs but
39:58
with regular cars and and the the camera
40:02
was not in any kind of enclosure but
40:05
they attracted too much attention and
40:06
people would like smile for the camera
40:10
or do more profane things for the camera
40:16
so they’ve solved that problem by
40:19
getting a taller vehicle and you know
40:22
putting the camera in an enclosure so it
40:23
just sort of looks like some luggage on
40:25
the roof and people don’t notice it and
40:28
the so let’s see here’s one of the you
40:34
guys if you’ve been to Times Square you
40:35
may have seen this guy he’s out there
40:37
all the time
40:37
I think he’s called The Naked Cowboy
40:39
he’s been doing this for years and years
40:41
and years he was there
40:43
he’s been there at least for eight years
40:44
but maybe longer I don’t know and I
40:46
think TGI Fridays has hired him now as a
40:50
sort of mascot but anyway they captured
40:55
the naked cowboy in Times Square
40:58
let me is this when I go to the video
41:01
all right let me see we’re gonna do this
41:05
the hard way because it’s not working
41:08
the easy way all right watch this I
41:14
should actually let me pause this for a
41:16
sec this let me give you a little setup
41:20
so this is some some of the actual
41:23
footage from this camera this is a just
41:27
to give you the this is a camera that is
41:29
connected to a laptop that has a GPS in
41:33
it and so the Jeep the car just drives
41:36
around and the driver doesn’t have to do
41:39
anything except make sure all the
41:40
equipment is running properly and it
41:43
just takes you know a bunch of frames
41:45
per second and and and and and and the
41:49
car knows where it is at all times so
41:52
this now we one of the cities that we
41:56
did we did over ten cities but one of
41:59
them was Washington DC and and and the
42:03
driver whose name is Josh in this case
42:06
had a fairly interesting experience
42:08
because the camera broke and needed some
42:11
adjustments
42:12
it just got jiggled or something and
42:14
needed to be fixed and remounted so
42:17
which happens occasionally so he didn’t
42:19
think anything of it he stopped the car
42:21
he got out and he started to fix the
42:23
camera but unfortunately for Josh he
42:28
happened to stop right in front of the
42:29
State Department and and unfortunately
42:35
just to make matters a little more
42:37
complicated it was the day before the
42:39
election and so Washington was at its
42:43
highest state of alert and fortunately
42:47
this entire incident was captured on
42:51
video and it’s kind of fun to watch so
42:58
there’s Josh he’s getting out of the car
43:01
gonna work on the camera little but he
43:04
sees somebody’s coming he’s reaching for
43:06
his cell phone now trying to call Amazon
43:08
become legal now there are two of them
43:16
they’re very curious why josh is
43:18
photographing the State Department
43:20
here’s the third one
43:28
this here’s the here’s the fourth one
43:32
and if you can catch glimpses of Josh’s
43:34
face you can see that he is a little
43:36
stressed by the way to at a certain
43:40
point here you can see the woman in the
43:41
red hair I think it’s the one sort of in
43:45
charge she’s trying to calm Josh down
43:46
and she says something to him like it’s
43:49
not like we’re accusing you of having a
43:50
bomb or anything which didn’t really
43:53
work oh yeah here’s the great they take
43:58
a picture of Josh and the and they they
44:03
actually were were
44:07
they may actually were very friendly to
44:11
Josh and he explained the whole thing
44:13
and they made some phone calls and
44:15
confirmed his strange but true story
44:18
about what was going on and the whole
44:22
incident only took like 45 minutes and
44:25
and and Josh was sent on his way to
44:29
photograph the rest of Washington DC
44:36
let’s go back okay but all of those
44:45
things okay all those things are if you
44:48
look at what all the problems that had
44:50
to be solved to do something like get
44:52
that you know the GPS integrated with
44:55
the digital camera and get those in
44:57
trucks and then solve all these little
44:59
point problems that people were noticing
45:01
the camera and so on
45:02
that’s what innovating is all about it
45:05
really is those kinds of problems and
45:07
the great thing about this is at
45:09
amazon.com is I do very little of this
45:13
so you know we have a culture where
45:15
people in small teams can go off and do
45:17
these relatively low-cost experiments
45:19
and try to build neat things if you want
45:23
to have an innovative company the single
45:26
most important thing even ahead of
45:29
minimizing the cost of experimentation
45:32
is to make sure that you’re selecting
45:34
the people correctly on the way in and
45:37
so you need to hire people who like to
45:39
build who like to invent and you need to
45:42
make sure that they’d like to do that at
45:43
all granularities you know sometimes you
45:45
come across these people who are only
45:46
interested in inventing it the at the
45:49
grandest sort of whiteboard level and
45:52
they actually can’t make progress in the
45:53
real world you know they’re unwilling to
45:55
figure out how to mount the camera on
45:57
top of the truck and mounting the camera
45:59
on top of the truck turns out to be
46:01
incredibly important this is I’ve mostly
46:05
included this slide to demonstrate my
46:07
true pedigree as a geek this is me in
46:10
fourth grade and most of you in this
46:16
room are far too young to recognize that
46:18
device that I’m sitting in front of
46:21
but I was I’m one of the only people I
46:23
know who had access to a combined who
46:26
had access to a computer in fourth grade
46:27
I went the school that I went to in
46:29
Houston got this a company donated this
46:33
teletype and also donated some excess
46:36
mainframe time sharing time that they
46:39
had and so we got this thing one day and
46:43
none of the teachers knew how to use it
46:44
and nobody knew how to use it but had
46:46
manuals and that that phone receiver is
46:49
it called an acoustic modem and you
46:52
actually put the it’s a 300 baud modem
46:55
and the all the programs were stored on
46:59
paper tape which you’d punch holes in
47:01
and we me and about three other kids
47:05
started staying after school and and
47:08
learned how to program this this
47:12
mainframe computer and store our
47:14
programs and our very primitive programs
47:16
on paper tape and that went on for
47:18
several months until we discovered that
47:23
the mainframe was pre-programmed
47:26
to play Star Trek and pretty much that’s
47:31
all we did with the computer after that
47:33
we would stay after school the teachers
47:35
were like boy these kids just love
47:37
computers it was this really cool Star
47:41
Trek game where you know the it’s all
47:44
done it’s teletype on paper and it would
47:46
print a little I think it was a nine by
47:49
nine grid and somewhere in that nine by
47:51
nine grid like you know an asterisk
47:52
would be a star and a plus sign would be
47:54
a Klingon ship and you had to do it was
47:57
a simulation game and you had to like
47:59
you know you had limited resources like
48:01
you had only so much power and gonna
48:03
decide how to allocate that power
48:05
between shields and engines and phasers
48:06
and it was actually unbelievably fun and
48:12
so but you’ve got it this is you’ve got
48:17
to find people who are passionate about
48:18
building this is our first employment ad
48:21
at amazon.com we post this on Usenet and
48:26
and we asked for exactly what we wanted
48:29
and and you know the the technology in
48:32
computer science people
48:34
and businesspeople who are interested in
48:36
technology have always been most
48:37
important so you must have experience
48:40
designing and building large and complex
48:42
yet maintainable systems and you should
48:43
be able to do so in about one third of
48:45
time that most competent people think
48:47
possible and by the way as a total aside
48:52
you’ll see that the name of the company
48:54
at the time was Cadabra that was the
48:57
first name of the company and Cadabra
49:00
was a terrible name
49:02
because while driving across the country
49:05
is my wife driving I called an attorney
49:08
in Seattle and asked the attorney to if
49:12
he could incorporate the company so it
49:14
would already be ready by the time we
49:16
arrived and he said sure this is no
49:20
problem by the way this attorney was my
49:23
friend’s divorce attorney so it wasn’t
49:25
this wasn’t like Wilson Sonsini or
49:28
anything you know this was and and we
49:31
said sure you know what do you want the
49:32
name of the company to be and I you know
49:36
was well proud of myself for having
49:38
anticipated that question and I said
49:41
Cadabra and he said cadaver I knew right
49:48
then that my name choice was not gonna
49:51
survive for long either no no no Cadabra
49:54
like abracadabra okay so he about three
50:03
months later we changed the name to
50:05
amazon.com and that was chosen because
50:09
it started with the letter A and at that
50:11
time all of the lists online were
50:14
alphabetized and it was also the domain
50:17
name was available and it was it worked
50:19
internationally and it was short and
50:21
easy to spell so there were a whole
50:23
bunch of good features of that