lor decades, large companles have been wary of corporaLe venLurlng.
Some have seen Lhelr venLure
lnlLlaLlves fall ouLrlghL, and many more have glven up Loo qulckly: 1he medlan llfe span of corporaLe venLurlng programs has LradlLlonally hovered around one year. Lven flrms wlLh successful funds have someLlmes sLruggled Lo make use of Lhe knowledge galned from sLarL-up lnvesLmenLs. 1o be sure, runnlng successful corporaLe vC programs lsn'L easy: Companles' processes and rules can make Lhem slow-fooLed and unfocused. 8uL as dlsappolnLmenL wlLh 8&u resulLs grows, Lhere are lndlcaLlons LhaL corporaLe venLurlng may be galnlng ground-and respecL. Start-ups 8acked by Corporat|ons Lxce| SLarL-ups LhaL wenL publlc afLer belng funded by aL leasL one corporaLe venLure-caplLal lnvesLor ouLperformed Lhose funded excluslvely by lndependenL vCs. 1hese flgures, drawn from 1980-2004 daLa, are averages for Lhe Lhree years followlng Lhe lC.
Source: 1homas !. Chemmanur and Llena LouLsklna, 2008 Companles hoplng Lo acqulre knowledge and aglllLy from corporaLe venLurlng can beneflL from followlng slx sLeps, lncludlng allgnlng goals, provldlng Lhe rlghL lncenLlves, and creaLlng sysLems Lo Lransfer knowledge. l'll go Lhrough Lhese sLeps one by one, showlng how flrms can esLabllsh venLure funds LhaL are as savvy and nlmble as Lhe besL prlvaLe vCs. 8uL flrsL, leL's conslder Lhe poLenLlal beneflLs of corporaLe venLurlng. 1he Case for Ventur|ng ln Lhe pasL, corporaLe lnLeresL ln creaLlng venLure funds Lended Lo wax and wane ln sync wlLh Lhe general vC cllmaLe. Waves of corporaLe venLure acLlvlLy-ln Lhe laLe 1960s, Lhe mld-1980s, and Lhe laLe 1990s-corresponded wlLh booms ln vC lnvesLmenLs and venLure-backed lCs. 8uL now we're seelng a corporaLe-venLurlng surge even durlng lacklusLer days for LradlLlonal venLure caplLal. ln Lhe flrsL half of 2011, when lndependenL funds were sLruggllng Lo ralse caplLal ln Lhe wake of Lhe global flnanclal crlsls, more Lhan 11 of Lhe vC dollars lnvesLed came from corporaLe venLure funds, a level noL seen slnce Lhe doL-com bubble. 1hls new acLlvlLy may lndlcaLe LhaL as research funcLlons face severe pressure Lo reln ln cosLs and produce resulLs, companles are looklng for alLernaLlve means Lo learn and lnnovaLe. Companles as dlverse as Coogle, 8MW, and Ceneral Mllls are complemenLlng LradlLlonal 8&u by [olnlng wlLh oLher lnvesLors Lo puL money lnLo promlslng sLarL-ups. 1he loglc ls lndeed compelllng. A faster response. 8y provldlng boLh an lnslde look aL new Lechnologlcal flelds and a paLh Lo posslble ownershlp or use of new ldeas, corporaLe venLurlng can allow a flrm Lo respond qulckly Lo markeL LransformaLlons. Lllly venLures was [usL one of several corporaLe venLure lnlLlaLlves ln Lhe 1990s and 2000s LhaL helped pharmaceuLlcal companles caLch up wlLh Lhe rapld advances ln blosclence LhaL were LhreaLenlng Lo render Lhelr chemlsLry-based experLlse lrrelevanL. ln a sLudy of 71 venLure lnlLlaLlves by blopharmaceuLlcal flrms from 1983 Lo 2003, Pyunsung uanlel kang and vlkram k. nanda of Ceorgla 1ech found LhaL companles LhaL made flnanclally successful lnvesLmenLs also experlenced greaLer success ln drug developmenL. lL's llkely LhaL developlng capablllLles such as Lhls on Lhelr own would have Laken Lllly and oLher pharma companles far longer and been far more expenslve. Clven Lhe Llme and resources needed Lo updaLe research faclllLles and recrulL sclenLlsLs wlLh Lhe rlghL experLlse, Lhe growLh of knowledge ln lnLernal laboraLorles can be palnfully slow. A better v|ew of threats. A venLure fund can serve as an lnLelllgence-gaLherlng lnlLlaLlve, helplng a company proLecL lLself from emerglng compeLlLlve LhreaLs. uurlng Lhe 1980s, for example, when lnLegraLed-clrculL makers were searchlng for alLernaLlves Lo slllcon (Lhe basls of Lhe domlnanL chlp Lechnology), Lhe slllcon-chlp speclallsL Analog uevlces creaLed a venLure program Lo lnvesL ln compeLlng Lechnologles. lLs goal was Lo gaLher sLraLeglc lnformaLlon aL relaLlvely low cosL. Analog's porLfollo dldn'L do very well. !usL one of lLs 13 companles wenL publlc, and only afLer so many flnanclng rounds LhaL Analog's sLake was heavlly dlluLed. 8uL Lhe reason for Lhe lacklusLer performance was slgnlflcanL: Maklng chlps ouL of anyLhlng oLher Lhan slllcon Lurned ouL Lo be sLubbornly dlfflculL and expenslve. Cnce Lhls reallLy hlL Lhe markeLs, makers of slllcon chlps saw Lhelr valuaLlons splke, Analog's lncreased sevenfold from 1979 Lo 1983. 8uL Lhe corporaLe venLurlng program had provlded lnsurance: lf Lhe alLernaLlves had been vlable, Analog would have been covered. 1radlLlonal 8&u doesn'L do a good [ob of snlfflng ouL compeLlLlve LhreaLs. More and more, corporaLe 8&u unlLs Lend Lo focus on a narrow range of pro[ecLs, Lhus poLenLlally neglecLlng dlsrupLlve advances LhaL occur ouLslde Lhe company. lenLy of execuLlves ln companles wlLh robusL 8&u funcLlons lle awake wonderlng wheLher Lhelr flrms are abouL Lo be bllndslded by Lechnologles Lhey've never heard of. Las|er d|sengagement. AnoLher beneflL of venLurlng, one LhaL's closely relaLed Lo acceleraLlng Lhe company's response Lo change and LhreaLs, ls LhaL lL glves execuLlves a fasLer way Lo dlsengage from lnvesLmenLs LhaL seem Lo be golng nowhere. As ls well-known, many companles flnd lL dlfflculL Lo abandon Lhe noL-qulLe-good-enough lnnovaLlons LhaL someLlmes come ouL of lnLernal labs. 1hese pro[ecLs can llnger ln producL developmenL for years, reslsLlng LermlnaLlon (desplLe much Lalk abouL 8&u porLfollo managemenL). nokla's lnslsLence on developlng lLs phones uslng Lhe Symblan operaLlng sysLem, even as lLs compeLlLlve poslLlon wenL lnLo free fall, ls a classlc lllusLraLlon. 1he arm's-lengLh relaLlonshlp beLween companles and Lhelr venLure funds offers advanLages ln Lhls regard: 1he besL funds Lend Lo be qulcker on Lhe Lrlgger Lhan Lhelr corporaLe parenLs. Lven lf a corporaLlon ls unwllllng Lo LermlnaLe an unpromlslng lnlLlaLlve, Lhe presence of co-lnvesLors may force Lhe declslon. A b|gger bang. 8y comblnlng lLs own caplLal wlLh LhaL of oLher vCs, a corporaLe venLure can magnlfy Lhe lmpacL of lLs lnvesLmenLs. 1hls ls parLlcularly beneflclal when Lechnologlcal uncerLalnLy ls hlgh. A dramaLlc example ls Lhe llund, supporLed by Apple and launched ln 2008 by Lhe venerable vC flrm klelner erklns Caufleld & 8yers on Lhe day when ouLslde developers were flrsL allowed Lo begln worklng on apps for Lhe lhone. 1he $100 mllllon fund-whlch subsequenLly doubled ln slze-lnvesLed ln companles developlng games and Lools. ln Lhls way, Apple rapldly bullL a crlLlcal mass of appllcaLlons for lLs new phone whlle spendlng very llLLle. (1he conLrasL wlLh Apple's rlval nokla, whlch eschewed such an approach when promoLlng lLs Symblan sysLem, ls sLrlklng.) Clven Lhe success of Lhe llund, lL ls noL surprlslng LhaL slmllar efforLs have been launched by, among oLhers, 8esearch ln MoLlon (Lo encourage Lhe developmenL of Lhlrd-parLy appllcaLlons for Lhe 8lack8erry) and lacebook (whlch Leamed wlLh klelner, Amazon, Zynga, and oLher Lech lumlnarles Lo esLabllsh Lhe slund, devoLed Lo promoLlng companles LhaL work wlLh soclal medla slLes). Increased demand. 1he llund also serves as an example of a dlfferenL klnd of leveraglng: 8y encouraglng Lhe developmenL of Lechnologles LhaL rely on Lhe parenL corporaLlon's plaLform, venLure lnvesLmenLs can help lncrease demand for Lhe corporaLlon's own producLs. lnLel CaplLal Look Lhls approach ln laLe 1998, when lL esLabllshed a fund LhaL would help speed Lhe enLry of lnLel's nexL- generaLlon semlconducLor chlp lnLo Lhe markeL. lund managers lnvesLed ln many sofLware and hardware makers (ofLen lnLel compeLlLors) whose producLs caplLallzed on Lhe new chlp's power. 1hose lnvesLmenLs acceleraLed Lhe chlp's adopLlon by several monLhs, accordlng Lo lnLel. A k|s|ng 1|de Large companles have been wary of creaLlng corporaLe vC funds, Lhe medlan llfe span of Lhese funds has been abouL one year. 8uL as dlsappolnLmenL wlLh 8&u grows, Lhere are lndlcaLlons LhaL corporaLe venLurlng ls galnlng ground-even ln a lacklusLer envlronmenL for LradlLlonal venLure caplLal.
Source: naLlonal venLure CaplLal AssoclaLlon lnLel CaplLal also played a role ln seedlng companles developlng wlreless lnLerneL producLs around Lhe 802.11 neLwork sLandards, whlch had been champloned by lnLel: ln Lhe flve monLhs before Lhe 2003 lnLroducLlon of Lhe wlreless-enabled CenLrlno chlp seL, Lhe fund revealed lLs lnLenLlon Lo lnvesL $130 mllllon ln ComeLa neLworks and oLher companles LhaL were promoLlng Lhe adopLlon of Wl-ll neLworks. 1he rapld upLake ln lnLel's wlreless producLs ln subsequenL years reflecLs Lhe company's success ln uslng corporaLe venLurlng Lo creaLe an ecosysLem of wlreless players. n|gher returns. llnally, Lhere's Lhe purely flnanclal aspecL of venLurlng. lor lndependenL vCs, maklng money for Lhe llmlLed parLners ls Lhe prlmary lf noL Lhe sole ob[ecL. lor corporaLe venLure funds, galnlng sLraLeglc beneflLs ls usually Lhe maln goal, proflLs from venLurlng Lyplcally aren'L slgnlflcanL enough Lo maLLer Lo Lhe parenL company's boLLom llne. SLlll, proflLs are always nlce Lo have. Companles brlng a loL of value Lo Lhe sLarL-ups Lhey fund, ln Lhe form of repuLaLlon, skllls, and, of course, resources-from research sclenLlsLs Lo sophlsLlcaLed laboraLorles Lo armles of salespeople. 1hey also change Lhe way ouLslde lnvesLors vlew Lhe young flrms' prospecLs. rlvaLe and publlc equlLy lnvesLors ofLen anLlclpaLe LhaL a corporaLlon-backed sLarL-up wlll ulLlmaLely be boughL by Lhe company LhaL lnvesLed ln lL-and aL an aLLracLlve valuaLlon, reflecLlng Lhe sLraLeglc beneflLs Lhe sLarL-up can offer lLs new owner. 1hus lL's perhaps noL surprlslng, as 1homas !. Chemmanur, of 8osLon College, and Llena LouLsklna, of Lhe unlverslLy of vlrglnla's uarden School of 8uslness, have shown, LhaL sLarL-ups backed by corporaLlons are more llkely Lhan Lyplcal vC-backed flrms Lo aLLracL Lhe aLLenLlon of hlgh-quallLy markeL players-from lnvesLmenL banks Lo equlLy analysLs Lo lnsLlLuLlonal lnvesLors-when Lhey go publlc. uurlng Lhelr flrsL Lhree years as publlc companles, Lhe researchers found, flrms backed by corporaLe venLure funds show beLLer sLock prlce performance, on average, Lhan Lhose backed by LradlLlonal venLure groups. Mak|ng It Work uesplLe corporaLe venLurlng's compelllng loglc, venLure funds someLlmes run lnLo Lrouble. 8llllons of dollars have gone down Lhe draln as corporaLlons have sLruggled Lo deploy Lhelr venLure caplLal groups effecLlvely. MosL of Lhe problems are rooLed ln lncompaLlblllLles beLween Lwo mlnd-seLs: LhaL of Lhe rlsk-lovlng, someLlmes ruLhless venLure caplLallsL, and LhaL of Lhe process-bound corporaLe execuLlve. lf companles aren'L careful, Lhelr venLure caplLallsLs can become ensnared ln Lhe agendas of myrlad corporaLe sLakeholders or demoLlvaLed by lnadequaLe or poorly deslgned flnanclal lncenLlves. And Lhe parenL company can mlss ouL on valuable knowledge. 1hese slx sLeps can help companles avold Lhe plLfalls. A||gn goa|s w|th corporate ob[ect|ves. AllgnmenL of goals across Lhe venLure fund, Lhe sLarL-ups, and Lhe parenL company enables a corporaLe venLure group Lo draw on Lhe parenL's experLlse. WlLhouL LhaL allgnmenL, corporaLe venLurers are less llkely Lo make good lnvesLmenL declslons and aLLracL hlgh- callber enLrepreneurs-and useful knowledge ls less llkely Lo flow from Lhe sLarL-ups Lo Lhe corporaLe parenL. ln a sLudy of flnanclal reLurns from more Lhan 30,000 lnvesLmenLs ln enLrepreneurlal flrms, aul A. Compers, of Parvard 8uslness School, and l found LhaL corporaLe venLure funds are more successful lf Lhe sLaLed focus of Lhe corporaLe parenL and Lhe buslness of Lhe porLfollo flrm overlap. ln comparlson wlLh sLarL-ups LhaL aren'L llnked wlLh Lhe company's goals, well-allgned sLarL-ups are less llkely Lo be LermlnaLed and more llkely Lo go publlc, produce hlgher numbers of paLenLs wlLhln four years of golng publlc, and have beLLer sLock prlce performance. Stream||ne approva|s. A venLure fund's goals should be noL only allgned wlLh Lhe parenL company's, buL also few ln number. A sLreamllned approval process can help. ln many companles, varlous lnLernal consLlLuencles musL approve Lhese funds' goals-a slLuaLlon LhaL can lead Lo absurd resulLs. When l8M abandoned lLs llreworks arLners afLer Lwo years, Lhe fund's lnlLlal proposed lnvesLmenLs were sLlll Lled up ln lnLernal revlew wlLh numerous dlvlslonal vlce presldenLs. uelays llke Lhls noL only drlve corporaLe venLure professlonals crazy buL also slgnal Lo exLernal lnvesLors and sLarL-ups LhaL Lhe fund ls lneffecLlve. A LorLuous approval process lnevlLably burdens Lhe fund wlLh Loo many goals. 1o please 8&u, Lhe fund mlghL alm Lo galn knowledge abouL emerglng Lechnologles. 1o please Lhe buslness developmenL group, lL mlghL look for sLarL-ups LhaL could become acqulslLlon LargeLs. 1o saLlsfy Lhe ClC, lL mlghL alm for a cerLaln Lhreshold of flnanclal reLurns. Managers' energles are spread Loo Lhln, and Lhe fund wanders from goal Lo goal wlLh no clear ob[ecLlve. 1hls problem conLrlbuLed Lo Lhe specLacular fallure of Lxxon LnLerprlses' venLure-caplLal efforL. 1he program began ln 1964 wlLh a mandaLe Lo explolL Lechnologles ln Lxxon's corporaLe laboraLorles. lL Lhen shlfLed Lo maklng mlnorlLy lnvesLmenLs ln lndusLrles from advanced maLerlals Lo alr-polluLlon conLrol Lo medlcal devlces. lL laLer changed course agaln, focuslng on compuLlng sysLems for offlce use. 8efore Lhe lnlLlaLlve was abandoned, ln 1983, Lhe compuLlng-sysLems lnvesLmenLs alone had generaLed an esLlmaLed $2 bllllon ln losses. A compllcaLed corporaLe declslon process can also lead Lo lneffecLlve lnvesLlng paLLerns. lf geLLlng approval ls arduous, lnvesLmenLs are made only when Lop execuLlves are flred up and moLlvaLed Lo acL qulckly-usually because Lhe medla are hyplng a parLlcular Lechnology or markeL segmenL. 8uL Lhese are usually Lhe worsL Llmes Lo lnvesL, wlLh valuaLlons hlgh and probable reLurns low. A sLreamllned approval process allows a venLure fund Lo acL qulckly on promlslng buL unheralded lnvesLmenLs, Lhus enabllng a conLrarlan approach LhaL mlghL lead Lo Lhe ldenLlflcaLlon of neglecLed opporLunlLles. rov|de powerfu| |ncent|ves. CorporaLe venLure professlonals ofLen expecL Lhe level of compensaLlon and Lhe lncenLlves LhaL lndependenL vCs en[oy. 8uL corporaLe leaders are Lyplcally Lroubled by Lhe dlsparlLy beLween whaL venLure managers expecL Lo earn and Lhe compensaLlon of execuLlves wlLh comparable senlorlLy ln oLher parLs of Lhe company. And Lhey prefer Lo provlde lncenLlves LhaL are Lled Lo Lhe performance of Lhe company, noL of parLlcular lnvesLmenLs. We can'L have people ln separaLe rowboaLs," Ceneral LlecLrlc's !ack Welch once sald, ln reference Lo a venLure Leam's lncenLlves. We don'L wanL anybody ln our company golng Lo a meeLlng wlLh a dlfferenL lnLeresL from everybody else." SLarL-ups LhaL are allgned wlLh Lhe parenL company's goals are more llkely Lo go publlc and have beLLer sLock prlce performance. 8uL LreaLlng venLure lnvesLors llke oLher managers can lead Lo a loss of LalenL and moLlvaLlon on venLure Leams, and a lack of focus on long-Lerm corporaLe goals. CorporaLlons LhaL fall Lo provlde adequaLe lncenLlves face a sLeady sLream of defecLlons once [unlor lnvesLors masLer Lhe venLure process. AfLer Loo many board meeLlngs for whlch Lhe corporaLe lnvesLor parks hls llesLa nexL Lo Lhe lndependenL venLure caplLallsL's lerrarl, Lhe LempLaLlon Lo go elsewhere becomes overwhelmlng. 1he corporaLlon, havlng borne Lhe cosL of Lralnlng Lhe lnvesLor, doesn'L reap Lhe beneflL of hls or her experLlse. CL lLself pald Lhe prlce: ln 1998 and 1999, CL LqulLy losL 18 lnvesLors, a number of whom wenL on Lo leadlng vC flrms. Lll Lllly's venLure lnlLlaLlve was aL flrsL an organlc parL of Lhe company-lLs venLure caplLallsLs were corporaLe employees wlLhouL any proflL share. AfLer a slew of defecLlons, Lllly analyzed compensaLlon levels and found LhaL only Lhe mosL [unlor sLaffers aL Lllly venLures were belng rewarded aL anyLhlng llke a markeL level. SLlll, Lhe company's senlor managemenL and P8 professlonals reslsLed changlng Lhe pay scheme. lL wasn'L unLll 2009 LhaL Lllly's managemenL agreed Lo Lurn Lhe venLure group lnLo a freesLandlng organlzaLlon. AfLer a slew of defecLlons, Lllly dlscovered LhaL only Lhe mosL [unlor sLaffers aL Lllly venLures were belng compensaLed aL anyLhlng llke a markeL level. lor recrulLmenL and reLenLlon, compensaLlon levels ln a corporaLe venLure lnlLlaLlve should maLch Lhose offered by lndependenL venLure groups. AL Lhe same Llme, pay should be llnked Lo corporaLe goals as well as sLarL-ups' long-Lerm performance. ln a sLudy of corporaLe venLure funds, Cary uushnlLsky, of London 8uslness School, and Zur Shaplra, of new ?ork unlverslLy, found LhaL Lhose LhaL closely llnked pay Lo demonsLraLed lnvesLmenL success (ofLen, boLh flnanclal and sLraLeglc reLurns Lo Lhe corporaLe parenL) were more llkely Lhan oLhers Lo make successful lnvesLmenLs and Lo lnvesL ln earller-sLage companles-evldence LhaL Lhey were nlmbler and more aggresslve. lndeed, many of Lhe programs wlLh Lhe greaLesL sLablllLy-ln Lerms of boLh managemenL Leam and mlsslon-have been characLerlzed by hlgh-powered lncenLlves. An example ls ClaxoSmlLhkllne's S8 Cne, whlch operaLed under a slngle head, eLer Sears, from 1983 Lo 1999. uurlng mosL of LhaL perlod, Lhe corporaLe vCs recelved 13 of Lhe proflLs Lhey generaLed and bonuses, based on less Langlble beneflLs Lo Lhe corporaLlon, LhaL could represenL as much as 3 of Lhe fund's caplLal galns. 1hls approach kepL venLure lnvesLors senslLlve Lo boLh Lhelr flnanclal ob[ecLlves and Lhe parenL company's sLraLeglc needs. Create an exper|menta|, fa||ure-to|erant m|nd-set. 8lsk averslon can be a serlous problem for a corporaLe venLure-caplLal fund. SomeLlmes LhaL aLLlLude sLems from Lhe corporaLe parenL's culLure. When a venLurlng Leam boasLs LhaL no flrms ln lLs porLfollo have been shuLLered, corporaLe execuLlves may lnLerpreL Lhe announcemenL as a slgn of success. 8uL glven Lhe naLure of Lhe enLrepreneurlal process, and Lhe facL LhaL a slgnlflcanL fracLlon of lndependenL venLure lnvesLors' LransacLlons end ln fallure, Lhe perfecL record may be a slgnal LhaL Lhe Leam ls playlng lL Loo safe, lnvesLlng ln companles wlLh an eye Lo avoldlng fallure. Well-sLrucLured lncenLlves can help: 1hey can focus corporaLe venLurers on maxlmlzlng lnvesLmenL success, wheLher sLraLeglc or flnanclal, and mlnlmlze Lhelr worrles abouL geLLlng Lhelr knuckles rapped for shuLLerlng lnvesLmenLs or selllng sLarL-ups aL a loss. St|ck to your comm|tments. Whlle lL's lmporLanL Lo LermlnaLe morlbund pro[ecLs, lL's also lmporLanL noL Lo walk away from promlslng ones. A low level of corporaLe commlLmenL Lo good pro[ecLs can be hlghly damaglng Lo a fund and lLs lnvesLmenLs. SomeLlmes merely a change ln Lop personnel can prompL a company Lo reLhlnk lLs commlLmenL Lo venLurlng ln general and Lo varlous lnvesLmenLs ln parLlcular. ln some organlzaLlons, lL's a rlLual for new execuLlves Lo dlscard Lhelr predecessors' pro[ecLs. 8uL lf a parenL company ls seen as a flckle lnvesLor, professlonals wlll be wary of [olnlng lLs venLure unlL, enLrepreneurs wlll be relucLanL Lo accepL lLs funds, and lndependenL vCs wlll be heslLanL Lo [oln ln, seLLlng off a deaLh splral. 1o aLLracL hlgh-callber ouLslde lnvesLors Lo Lhelr venLurlng efforLs, companles should adopL Lhe aLLlLude of lndependenL vCs: As long as a sLarL-up ls healLhy, commlLmenLs are blndlng. lf a llmlLed parLner conLrlbuLes even a small amounL of Lhe LoLal caplLal promlsed aL Lhe Llme of closlng, Lhere ls an expecLaLlon LhaL Lhe LoLal amounL promlsed wlll be provlded. Lven durlng Lhe depLhs of Lhe flnanclal crlsls, lL was rare for lnvesLors Lo walk away from Lhose commlLmenLs. narvest va|uab|e |nformat|on. knowledge doesn'L auLomaLlcally flow from sLarL-ups Lo Lhe large organlzaLlons LhaL have lnvesLed ln Lhem-aL leasL noL ln a Llmely manner. 1he barrlers Lo knowledge Lransfer are many: 1he corporaLe venLurlng and buslness developmenL groups may be locaLed far from Lhe flrm's cenLral operaLlons. Lveryone ls busy wlLh day-Lo-day Lasks. 1here's a culLural gap beLween Lhe young M8As who domlnaLe mosL venLure Leams and Lhe flrm's senlor execuLlves. And, of course, Lhe fledgllng Lechnologles belng developed by porLfollo companles may noL seem appllcable wlLhln Lhe corporaLlon. 8uL a fallure Lo glve Lhe corporaLe parenL access Lo Lhe knowledge generaLed ln lLs lnvesLmenLs defeaLs a large parL of Lhe lnLelllgence-gaLherlng loglc of corporaLe venLurlng. Companles cannoL leave knowledge splllovers Lo chance. nor can Lhey slmply puL an operaLlng manager on Lhe board of each porLfollo flrm Lo be Lhe parenL company's eyes and ears, as CL and oLhers have done. A manager runnlng a 2,000-person refrlgeraLor assembly planL ls unllkely Lo have much Llme Lo worry abouL a 10-person sLarL-up LhaL doesn'L seem Lo be worklng on problems of lmmedlaLe relevance Lo Lhe corporaLlon. Cne of Lhe mosL successful meLhods l've seen for Lransferrlng knowledge from sLarL-ups Lo corporaLe parenLs ls Lhe creaLlon of llnked unlLs dedlcaLed Lo Lhls Lask. 1hls was Lhe approach Laken by Lhe u.S. CenLral lnLelllgence Agency's venLure-caplLal program, ln-C-1el. lounded ln 1999 Lo acqulre novel Lechnologles, Lhe fund prlmarlly made equlLy lnvesLmenLs ln young flrms, many of whlch had developed producLs for Lhe prlvaLe secLor-for lnsLance, Lechnologles for deLecLlng card counLers ln caslnos. lL was dlfflculL for people ln Lhese young companles Lo ldenLlfy who ln Lhe lnLelllgence communlLy mlghL be lnLeresLed ln Lhelr Lechnologles, and lL was hard for lnLelllgence professlonals Lo lmaglne how consumer- orlenLed Lechnologles mlghL be adapLed Lo Lhelr needs-Lo see, for example, how sofLware for ldenLlfylng Ml1 sLudenLs aL Lhe Caesars alace black[ack Lables could be used Lo ldenLlfy Al Caeda members. Moreover, communlcaLlon beLween Lhe sLarL-ups' execuLlves and Lhe Agency's producL developers was severely consLralned by llmlLs on sharlng classlfled lnformaLlon.