what i learned in 40 years of doing intelligence
TRANSCRIPT
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
1/8
Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)13
Al l stat ement s of fact, opin ion, or an alysis expr essed i n th is ar ti cle ar e th ose of the
person i nterviewed. N othin g in the art ic le should be constr ued as assert in g or i mpl y-
in g U S govern ment endorsement of i ts factual statements and in ter pretations.
Every intelligenceproduct must be rootedin a strong
understanding of the
audience it is written for.
M ar ti n Petersen i s a reti r ed seni or CI A D ir ectora te of I nt ell igence (DI )
officer a nd t he aut hor of a num ber of ar ti cles on i nt ell igence and in tell i-
gence ana lysis. I n l ate 2009 h e was asked by t hendi r ector of
i nt ell i gence M i chael M orell to cr eate a cour se for m anagers on review-
in g analyt ic produ cts and t eachi ng tr adecraft, w hi ch became th e Ar t of
Review Sem in ar. Th is art icle is adapt ed fr om remar ks deli ver ed t o DI
managers in September 2010.
An a dva nt ag e of getting older is increa sed perspective. I ha vebeen doing, thinking an d writ ing about int elligence a nd intelli-
gence ana lysis for a lmost 40 yea rs now. The business we a re in ha s
cha nged a great deal in tha t t ime, but more in i ts form th an in i ts
fundamentals .
I wa nt to focus on th ree broad topics: understa nding t he cus-
tomer, the importa nce of a service ment a lity, an d th e six th ings I
lear ned in doing a nd st udying intell igence an alysis during my
career in t he DI . While these experiences ar e draw n from w ork in
th e CIA, I believe the principles a pply across the I nt elligence Com-
mun ity (IC).
In the First Person
What I Learned in 40 Years of Doing IntelligenceAnalysis for US Foreign Policymakers
Martin Petersen
Un der st a n d i n g t h e Con sumer : F i v e Fun dam en t a l T r u t h s
I believe every intelligence
product m ust be rooted in a
strong understand ing of the
a udience it is writ ten for, an d I
believe there ar e five funda -
menta l t ru ths about the ana ly t-
ical products and their
consumers.
T r u t h n um ber o n e: t h ep r o d u ct i s op t i o n a l
equ i pm en t for ma ny key
consumers .
The m ost precious comm odity
in Wash ingt on is not informa -
tionthere is an overabun-
dan ce of informa tion, data ,
opinion, an d secretsbut t ime.
The fut ure in Wa shingt on is
four years at its longest point
an d every da y i t is one day
shorter. It is not sur prising
th en th at consumers of our ser-
v ices are in a hurry and t ha t
th ey ar e very busy people; th epresidents day is actually
plan ned in five minute incre-
ment s. These people ha ve
ma ny, ma ny sources of informa-
tion, an d ma ny of the people w e
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
2/8
Lessons in Serving Policymakers
14Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)
serve believe they a re better plugged int o the
world tha n we a re. And in man y cases, they are.
Our customers in the policyma king realm often
do not underst an d our mission, our values, or our
sta nda rds. They t end t o be skeptica l of int elli-
gence, especia lly if they a re new to th e policyma k-
ing w orld. They formed their views a bout w ho we
are, wha t w e do, and h ow w e do i t from th e same
sources other America ns do: popula r media , th e
press, an d congressiona l reportsnot a lwa ys th e
most a ccura te or sophisticat ed of sources and
generally not th e most flat tering. Our consumers
ha ve strong world views a nd clea r policy a gen-
das, a nd t hey often a ssume we have a policy
agenda , t oo.
I t is not surpris ing t hen t ha t policymakers donot a lwa ys see how w e ca n help them: After a ll,
I, the policy-
maker, am smart
an d ha ve excel-
lent sources of
informat ion (including all t he ones you have), a nd
I a m very busy, so wh y should I spend some of my
most pr ecious commodity on you? The rea lity for
intelligence officers is tha t w e must woo them,
sell them on t he need for our services, an d dem-
onstra te th e value of our ma terial da ily through
its t imeliness a nd i ts sophisticat ion. I f you are a nint elligence officer, th e tit le will often get you in
th e door, especia lly th e first t ime, but it will not
keep you there. Newcomers to th e IC ma y not
realize tha t t he CIA presence in th e Oval Office
during the G eorge W. B ush a dministrat ion w a s
th e exception, not th e rule.
If th e IC is going t o be part of th e regular rou-
tine in t he White H ouse, not only must w e have
something t o say t ha t people there can not get
somewhere elsewh ich ha s to be more th a n ha v-
ing secretsbut w e ha ve to be mindful of how w e
deliver it. We a re not only optional equipment; w e
ar e also guests a t t heir dinner part y. If we spill
the w ine, insult th e host , a nd oversta y our wel-
come, we w ill not be invited ba ck.
Speaking t rut h to power first req uires access to
power. My personal experience is tha t our con-
sumers will take frequent ba d news an d unha ppy
assessments as long as they are well-reasoned,
supported by data an d argu ment, and presented
wit hout ra ncor, value judgment s, or a rrogan ce.
T r u t h n um ber t wo : t h e w r i t t en p r o d u c t i sfo r ever .
A collea gue wh o spent h a lf his ca reer in the DI
an d ha lf in th e Nat iona l Clandestine Service
(NC S) once said only ha lf jokingly, You know
wh a t t he DI s problem is? You guys wr ite th ings
down. In t he NCS we believe in the oral tr adi-
tion. He was r ight in t he sense tha t t he written
word is for ever. Once it is printed, th ere is no ta k-
ing it ba ck or modifying it.
B r ief ings and background notes a re importa nt
par ts of doing t he mission, but th ey leave no per-man ent record. One can f ight over wha t w a s said
in a briefing,
bu t the wr i t ten
word is in bla ck
an d white. I t is
the WorldIntel-
ligenceReview (WIR e) a rt icle, the seria l flyer, th e
intelligence assessment, a nd the na tional intell i-
gence estima te (NIE ) th a t end up in th e archives,
an d i t is the paper product tha t gets held up a t a
congressional hea ring or eviscerat ed on a n edito-
r ial page.
And w hen I sa y forever, I mean for ever. Rela-
tively few people ha ve read t he now infa mous
NIE done in 2002 on Ir a qs w eapons of mass
destruction (WMD), but everyone knows w ha t it
say s. And everyone yea rs a nd yea rs from now will
know wha t i t sa id , because i t is viewedrightly
or wronglyas fat a lly flaw ed and r esponsible for
the second Iraq wa r. I t w il l never go awa y, and i t
joins th e pant heon of other real a nd ima gined
CI A fa ilures. Every t ime we publish, we go on
th e record an d t he record is there for ever,for the
second guessers, the hindsight experts, an d an y-one wit h a n a genda . Thus, our judgments need to
be as precise as w e ca n ma ke them, supported by
evidence a nd argument , a nd a ccura tely ref lect
our level of confidence every ti me.
Policymakers do not always see how we can help them.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
3/8
Lessons in Serving Policymakers
Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)15
T r u t h nu mber t h r ee: t he pub l i c does no t
segr ega t e success an d f a i l u r e.
Crit ics of int elligence, our customers, an d t he
general public do not sa y th at th e products of a
certa in office in CIA or DI A a re really grea t, buttha t the products of another office in tha t agency
are a wful . Nor do they say t ha t one type of ana ly-
sis, say politica l, can be trust ed, but t ha t our
work on something else, say S&T is unreliable.
Nor wil l they say tha t a l though they were wrong
last t ime, we ca n trust them this t ime.
No, customers remember, a nd t hey q uestion.
Sometimes th ey question fa irly, but often th ey do
not, especia lly those customers w ho find wha t is
being said to be inconvenient or unhelpful in
ad van cing a policy position th ey favor. From th eCI A a lone, I ca n produce a list of wha t I call
everybody
knows: every-
body know s
th e CIA failed to
predict t he fall
of the Sh a h of Ira n in 1979 or t he collapse of the
Soviet Un ion in 1991 or the India n nuclear t est in
1998 or this or th at . The facts a re often fa r more
complex, but t hey ha ve entered the popula r
mythology. And the consumers of intelligence say
out loud Why sh ould I tru st y ou on t his issue
wh en you were w rong on tha t one? Weak perfor-
man ce in one DI a rea immediat ely calls into
quest ion a ll w ork in th e CIA.
P resident K ennedy famously sa id of the CIA
th a t it s successes will be secret a nd its fa ilures
will be tru mpeted. To which I a dd my own corol-
lar y: in t he intelligence business success is tra nsi-
tory, a nd fa i lure is permanent.
T r u t h f ou r ( cl o sel y r e l a t ed t o t r u t h t h r ee) :
ou r i n d i v i d u a l a n d c ol l ec t i ve c r ed i b i l i t y
an d t h us ou r ab i l i t y t o do t hem i ssi onr i des on ever y p i ece of f i n i sh ed
in t e l l i gence th a t goes ou t t he doo r .
Sa d to say, no one cares wh at I th ink about a
part icular issuea nd no one car es wha t you per-
sonallyt hink eith er. They do car e tr emendously
about wha t the CIA orD IAorname the IC orga-
nizationthinks. The finished intelligence prod-
ucts tha t go out the door a re not persona l
products but corporat e ones.
IC products ha ve bra nd na mes, and t hey are
importa nt a nd powerful ones. They can open
doors , but they w il l not keep a ny a na lyst inside
circles of pow er if tha t bra nd na me is devalued by
shoddy w ork. Our customers rea d our products
for many r easons: to lear n, to make better deci-
sions, to know wha t the P residents Da ily B rief-
ing t ells th e president, t o look for a mmun ition in
a policy f ight, or to discredit wha t the IC says .
E very poorly-rea soned piece of finished int elli-
gence tar nishes a bra nd na me a bit a nd over t ime
can produce cracks in t he trus t t hey place in us to
live up to our t radecraf t . When tha t h appens
there is nothing one ca n say an d eventua lly t hebroader t rust is
lost . Ask BP an d
Toyota . One bad
oil well and a
few sticky accel-
erat ors undid
year s of excellent performa nce, an d shouting but
our record is still bett er th a n t ha t of [someone
else] m a kes no d ifference. We do not drill oil
w ells or build cars. We do the missiont he mis-
sion of protecting the U nited St at es. Our a bili ty
to ra ise th e level of the deba te or t o help policy-ma kers ma ke th e best decisions possible or t o
speak t rut h t o powerhowever one defines the
missionrests on one thing a nd one thing only:
our reputa tions for a na lytic rigor, objectivity, and
tota l integrity. Lose that a nd w e lose everyth ing.
T r u t h f i v e: ou r c ust omer s a r e sma r t er an d
mor e soph i st i c a t ed th an we g i ve th em c r ed i t
f o r ; t h ey have the i r own in depend en t sou r ces
o f i n f orm a t i o n a n d a n a l y si s w i t h wh i ch we
a r e comp et i ng .
And t hese customers a re continua lly cha nging.
We ha ve to esta blish our credibility a ndusefu l -
nessindividual by individual , a dministrat ion by
a dministra tion. There is no down time wh en it
comes t o qua lity.
Every poorly reasoned piece of finished intelligence tarnishesa brand name.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
4/8
Lessons in Serving Policymakers
16Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)
These five trut hs dema nd t ra decra ft excellence,
they deman d exacting s tan dards. (see the D I
Quality Framework above for an example), and
th ey dema nd t he pursuit of perfection. They
demand th at we lear n from our past , an d they
demand t ha t we a sk the best of ourselves every
tim e. To do the m ission; t o serve th e policy-
ma ker; to protect the na tionrequires nothing
less.
Th e Im po r t a nce o f a Ser v i c e M en t a l i t y
Excellence requires more than a standard of
qua lity. I believe it a lso dema nds a specific
approach t o the cra ft of int elligence ana lysis: it
requires a service menta lity. A service menta lity
is th e opposite of a product menta lity, w hich often
seems to drive the w ork of intelligence an a lysis,
an d th e difference is easiest t o explain by compar-
ing th e tw o. In a product men ta lity, the focus is
on th e producer, w ho thin ks of a product a s his or
hers . I t is a lso about pa ckaging t ha t product a nd
disseminat ing it w idely. Success is measur ed in
num bershow ma ny unit s were produced or how
ma ny r eceived each unit . It is a bout filling a book
or producing a product t o demonstrat e tha t a n
an alyst is ready for the next big s tep in a career.
In a service menta lity, the focus is on the cus-
tomerthe consumer of our servicesand specifi-
cally on how best t o meet th e customers needs. It
is not a bout t he a uth or or t he producing compo-nent ; it is a bout the recipient . It is a bout h elping
tha t customer understand a n issue. I t is about
being timely, releva nt , expert, a nd corporatein
our approaches to providing service, intelligence
a na lysis. Su ccess is measur ed not by the number
of units produced, but by how w ell the product
a ddresses and a nsw ers the specific concerns of an
identified and targeted audience.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
5/8
Lessons in Serving Policymakers
Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)17
P roduct a nd service are not mu tu ally exclusive.
Idea lly every product we produce should be
infused w ith a service menta li tya lthough we
often a ct like we a re in th e product business.
What difference does it make? When the productis more import a nt t ha n t he service it provides, we
relax our s ta nda rds to get t he productan other
unit of productionoff th e assembly line a nd out
th e door. Cl ose enough becomes good enough , an d
th e br and name suffers.
To infuse every product w ith a service menta l-
ity requires tw o th ings of intelligence an a lysts:
One is a set of s ta ndar dsthe DI Quali ty Fra me-
work in C IAs case; the other is ma stery of a sim-
ple techniqu ea sking t wo qu estions before
wr iting or briefing: wh o is the primaryau di ence
for t hi s pi ece and wh at i s th e speci fi c i nt el l i gence
questi on t hey
need h elp wi th?
I t is very hard
for th e aut hor of
a piece to ha ve a
service menta lity w hen he or she is focused on a
broad int elligence topic rat her t ha n a specific
intelligence quest ion. It is th e difference betw een
w e need a piece on t he demonstra tions in Tuni-
sia a nd w e need a piece on t he options t he Tuni-
sian government has for addressing the cause ofth e demonstra tions. A good int elligence question
ha s th e followin g propert ies: it bounds or na r-
rows th e subject ma tt er to be a ddressed; it gener-
ally conta ins a w ha t, w ho, why, or w here is i t
going element ; it is specific as t o the t opic or
event being addressed; an d i t is a question a nd
generally not a yes or no quest ion.
It is possible to ha ve man y different intelli-
gence quest ions for the same event. Curr ent int el-
ligence pieces genera lly work best w hen t hey a re
organized around one central question, although
th ey ma y touch on oth ers. Which question to
focus on is determ ined by w ho is selected as th e
primary a udience an d wha t tha t a udience is most
interested in or most needs to understa nd.
Forty years of experience have ta ught me tha t
failing t o identify a specific a udience a nd a n int el-
ligence question up front is often a t t he root of th e
w eakest a na lytic efforts. In t he Art of Review
Seminar we t a lk about The Road t o Ruin, the
first st ep on w hich is not clearly defining th e
issue to be addressed. This in t urn ea sily lea ds to
other, too common, failings in an aly tical w riting:
A fa i l u r e to p r esen t a c lea r ba si s fo r
j u d gm en t s.
A wea k piece typica lly specula tes on wha t h a p-
pens next but seldom provides the reason a n a na -
lyst believes the speculation is correct. The most
underused w ord in CIA DI a na lysis is beca use.
Every ma y a nd l ikely to an d could requires a
beca use s tat ement or i ts equivalentthe rea -
son we believe wh a t we believe. Absent t he
beca use, or its equiva lent, th a t a rt icle is just
a nother opinion in a tow n full of opinions.
Th e u se of
imp rec ise
l a n g u age .
It is not so
much tha t lan-
guage in a work of ana lysis is opaque but t ha t t he
point it is try ing t o make does not come through.
I t is s tat ing tha t X benefi ts from Y with out pro-
viding a s ta ndard by which to measure th e bene-
fit or spelling out precisely how a nd w hy X
benefits. Words like limits, ben efits, sug -gests, and a ll adverbs need a because or why
or how t o convey precise meaning. In tern a l
inconsistencies, not surprisingly, are often rooted
in imprecise langua ge.
Th e S i x Th i n g s I L ea r n ed
We all learn th e cra ft of intelligence ana lysis by
doing. The lessons are itera tive an d frequently
opaque, a nd t hey genera lly come slowly. Often
th ey a re only clear in looking ba ck. Now looking
back over nearly 40 yea rs, I think I ha ve lear nedth e followin g six thing s.
F i r st , how one th i n k s abou t t h e m i ssi on
a f fec ts deep l y how on e does th e m i ssi on .
I t hink t he intelligence an aly sts mission is less
a bout connecting t he dots (alt hough sometimes
it is) or predicting t he futu re (alt hough some-
times it is) or spea king tru th to power (alth ough
Excellence requires a service mentality approach to the craft ofintelligence analysis.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
6/8
Lessons in Serving Policymakers
18Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)
we often do) tha n i t is a bout un derstanding t he
world. Dots a nd prediction and t ruth ca n cause us
to na rrow our focus in a w orld of intelligence
cha llenges tha t ar e cha ra cter ized by t heir com-
plexity a nd most importa nt, by t heir dynamicna ture. In 40 years I learned th at quite often th e
most importa nt piece of the puzzle, a nd often t he
ha rdest one to get a ha ndle on, is wha t t he
U nited St a tes is doing in a given situa tionor, in
milita ry intelligence terms, understan ding t he
B lue component of a sit ua tion.
I a lwa ys th ought of my job a s bounding u ncer-
ta inty a nd by doing so helping ma ke my guy
smart er tha n their guy, whether i t wa s across a
conference table or across a ba tt lefield and
ena bling our policymakers t o make t he best deci-
sions possible given th e time a nd informa tion
available. Some-
t imes tha t
involved conn ect-
ing dots or pre-
dicting cours es of
action or provid-
ing warning , bu t i t a lways meant unders tanding
the forces at work in an y s i tua tionthe key var i-
a bles a nd drivers an d our adversa rys perspec-
tive. It is th e difference betw een stra tegic
understanding a nd ta ctical comma nd of an issue.
Second , i n t el l i gence fa i l u r es com e f r om
f a i l i n g t o st ep ba ck t o t h i n k a bou t
un der l y i n g t r end s, f or ces, and assum p t i o n s
no t f r om fa i l i n g to conn ec t do t s o r t o
p r e d i c t t h e f u t u r e.
When our focus becomes too tactical we fail to
see the str a tegic. We must learn to step ba ck from
time to t ime and a sk ourselves: wha t a re we not
seeing th a t we w ould expect to see if our line of
a na lysis w ere correct . The IC s 24-hour pr oduc-
tion cycle often m akes t his ha rd t o do, but
because i t is ha rd to do, it is essential th at we doit .
An underst a nding of history and cultur e is key
to coming to gr ips with t he assumptions t ha t
underpin much of our an a lysis. And I a m not
ta lking a bout our history and culture, but t he his-
tory a nd culture of the count ries we work on as
th e people and l ead er s of those coun tr i es und er-
stand t hem. Every a na lystregar dless of disci-
pline or roleneeds a deep apprecia tion of how a
people see themselves, their historical a mbitions,
a nd t heir grievan ces. For an a lysts focused on for-
eign lea ders, or politics, or economics, it is essen -tial th at they understa nd how power is acquired,
th e preferred w ay of exercising power, and th e
a ccepta ble and una ccepta ble uses of power, as
well a s t he defining life experiences of the key
a ctors in the count ries they specialize in.
Th i r d , good an a l ysi s ma kes th e comp le x
com pr ehensi b l e, wh i ch is no t th e sam e as
s imp le .
The key t o makin g t he complex compreh ensible
is having in mind a specific a udience a nd a very
precise intelligence question for t he a na lysis tota ckle. Dat a dumps and murky a na lysis a lmost
a lways are
rooted in tr y-
ing to write
about a devel-
opment w ith-
out f irs t asking, Who is my a udience an d w ha t
specifi c questi on does it n eed a nswered? I t is tha t
difference betw een we n eed a piece on the r iot-
ing in Athens a nd w e need a piece on t he gov-
ernment s options for a ddressing th e underlying
cause of th e rioting.
We do very well a s a rule in responding to qu es-
tions from policymakers. We come up short when
weha ve to supply the a udience a nd t he question
ourselves andwe star t t o wr i te before we have
done al l the thi nk i ng. If we think in t erms of
a nsw ering well defined quest ions, w e ca n ma ke
complex situa tions comprehensible, a nd w e a lso
sta nd a better chance of ma king clea r wh at we
know a nd do not know a ccura tely, conveying our
level of confidence, and presenting a convincing
basis for our judgments.
Fou r t h , t her e i s no subs t i t u t e fo r k now i ng
wh a t o n e i s t a l k i n g a bou t , wh i c h i s n ot t h e
sam e as kn ow i ng th e fa c t s .
Former C IA Director Michael H ay den once
fam ously said, If it is a fa ct, it a int int elligence.
The business of int elligence ana lysts is more
about putt ing facts in perspective tha n i t is ha v-
Absent the because, or its equivalent, an article is just anoth-er opinion in a town full of opinions.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
7/8
Lessons in Serving Policymakers
Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)19
ing comma nd of th e facts. We a re paid n ot for
wha t w e know, but for our a bili ty to th ink about
wha t w e knowor think w e know. I t is about
knowing w ha t is importa nt. I t gets back to those
as sumptions, drivers an d varia bles I dwell on.
Sourcesclandestine, open source, technical,
diploma tic, etc.a re not t he sa me as kn owledge.
Sources a re not the equiva lent of, or a subst itut e
for, expertise, the type of knowledge I talked
about in t he second th ing I learn ed. All sources
ar e best t hought of as opinions, some more
aut horita tive tha n others , but a l l should be sub-
ject to careful reflection a nd compa rison to wh a t
we know a nd believe. The da ngers in sources are
three-fold:
We tend to give greater credence to those tha tsuppor t wha t
we a l ready
believe.
S ou r ces a r e not
a scientific
sample but a small s l ice of a much larger a nd
more complex information picture.
They never answer the cr i t ical question of wha t
a re we not seeing but should see if our a na lysis
were correct.
Dur ing one of the most cha llenging times in my
an alyt ical career, I w orked for t he f inest a na lyst I
ever knew. In th e middle of th e Tian a nmen C ri-
sis in 1989w hen every ones ha ir wa s on fireI
found h im lat e one aft ernoon going through a
sta ck of musty old report s. I asked him wha t he
wa s doing. He said , I a m looking for t hings tha t
did not ma ke sense th en, but do now. He found
some, an d it profoundly a ffected our line of a na ly-
sis.
F i f t h , i n t el l i g ence an a l y si s sta r t s wh en w estop repor t i n g on even t s an d sta r t exp l a i n i n g
t h em .
Our production cycle puts a premium on being
ag ile, quick, an d sma rt. I t is often 24 hours or
less. The DI is one place where a consumer ca n
ask a question and get an a nswera thoughtful
an d considered oneovernight . It is one of t he
DIs great est s trengths. I t is a lso one of i ts great
vulnerabilities. It makes it harder to step back
an d t hink a bout underlying causes, dr ivers, a nd
var iables, especially in a crisis situa tion. My
Tian a nmen st ory is the exception. My career a san a na lyst ta ught me tha t lesson one (how we
th ink a bout th e mission) a nd lesson t w o (under-
sta nding forces at w ork) a re the key to operat ion-
a lizing lesson fiveth e need t o explain events.
S ix t h , man ager s of i n t el l i g ence ana l y st s get
th e beha v io r th ey r ewa r d , so they had bet t er
k n ow wh a t t h ey a r e r ewa r d i n g .
This is a message for a l l mana gers and a l l who
aspire to man agement. I t is my experience that i f
you have clea r s ta ndar ds an d a re seen as consis-
tent a nd fair in a pplying them, your unit w ill l iveup to the s ta ndard. And , you m ust also hold your-
sel f t o th e same
standards. If
you va lue ana -
lytic trade-
craf t , ta lk
a bout it a nd pra ctice it. If you wa nt open commu-
nicat ion wh ere different int erpreta tions are con-
sidered, invite it. I f you wa nt honesty, be honest.
And rew a rd t he behavior you profess to value.
There is a Chinese proverb: If your vision
extends one year, grow wh eat ; if it extends 10
year s, pla nt tr ees; if it extends 10,000 years, grow
a nd develop men. Ma na gers, your job is t o grow
men a nd w omen who can do the mission. The
standard of success, I believe, is uncompromis-
ingly simple: Did I lea ve the unit I led stronger
tha n I found i t?
Why I t A l l Ma t t er s
I f there is an underlying reali ty to al l tha t I
ha ve learn ed, it is the obvious: we a re in a very
difficult business. It is more life a nd dea th nowth a n it wa s in my h eyday. The consequences of
getting a na lysis wrong a re much great er now.
Int elligence is a lso more political now in t he
sense tha t w ha t is done today is more open th an
it ha s ever been a nd a s a result m ore subject to
part isan sniping.
If we think in terms of answering well-defined questions, wecan make complex situations comprehensible.
-
8/11/2019 What I Learned in 40 Years of Doing Intelligence
8/8
Lessons in Serving Policymakers
20Studies in Intelligence Vol. 55, No. 1 (Extracts, March 2011)
There are some who say th e United St at es is a
declining power or tha t it is the source of ma ny of
th e worlds problems. Time w ill tell on th e first
question, but I believe the U nited St at es is a force
for good in the w orld, and h ow powerful a forcedepends as mu ch on our knowledge a s on our mili-
ta ry a nd economic might. I tell intelligence ana -
lysts I teach that more often tha n not they a re the
source of tha t know ledge. It is their professiona l-
ism and tra decraft tha t provide checks on t he sys-
tem, light t he wa y, an d levera ge US pow er. All the
dolla rs spent on int elligenceth e collectors in th e
field, the t echnical systems, and the lives at
riskar e for
naught, unless
tha t knowledge
comes t ogeth er in
wha t ana lysts do
every da y.
As th e deputy executive director at th e CIA, I
a ddressed each class of just -promoted CI A Senior
Int elligence Service officers, an d ea ch t ime I
asked for a show of ha nds of those who believed
th ey would never see WMD u sed on U S soil in
th eir lifetimes. The question alw a ys sta rtled
them, an d I never saw a s ingle hand raised. We
can not af ford to accept a nyth ing less tha n th e
pursuit of perfection. We ca nnot a ccept a nyt hing
less th a n holding ourselves to the highest st a n-da rds. We ca nnot a ccept an yt hing less th a n our
best effort every time, every day. The potential
consequences ar e too great .
And I know it is damn h ar d. Intell igence ana ly-
sis is less fun tha n a policy rota tion or an over-
seas a ssignment. I t is less honored and
romanticized tha n other aspects of the G reat
G a me. I t is frustrat ing. I t is exha usting. And even
th e best effort s w ill be picked at . The a na lysts
work w ill be criticized by the knowledgeable an d
th e ignora nt a like. It w ill even be demonized attimesindependent of its qua lityand it w ill
a lwa ys be hostag e to the politics of th e moment .
B utand I say t his with my four decades of per-
spectivewhat intelligence analysts do has
impact. It m a tt ers. I ha ve seen the quiet victories
of intell igence an d t he mista kes averted, and I
have seen critics become advocates because of
w ha t ana ly s t s
do every da y.
We a ll chose
careers in intel-ligence for the
sa me reason: t o make a difference, to do the mis-
sion. The colleague who teaches the Kent Schools
Art of Review Semina r w ith me t ells a s tory
a bout Abra ha m Lincoln, who in one of the da rk-
est hours of the C ivil War at tended a Sunda y ser-
vice in tha t l i t t le church tha t s t i ll s ta nds a cross
from the White H ouse. On his wa y ba ck, he wa s
asked by a fellow parishioner wh at he thought of
th e young reverend. Lincoln replied tha t he ha d a
strong voice and clea r messag e, but t ha t he fa iled
to do one t hing; he fa iled to ask u s to do some-thing grea t .
I a m asking every an alyst w ho reads t his to do
something great . Do wha t brought you here. Do
th e mission every da y t o the best of your a bility.
And, ma y G od bless you for doing it .
What intelligence analysts do matters. I have seen the quietvictoriesmistakes avertedand critics become advocates.