Why some companies break through with radical innovation (and others don’t)

Revolu­tion­ary break­throughs in bio­medi­cine and tech­no­logy emerge at break­neck speed: from Regeneron’s AI-driv­en anti­body ther­apies that can identi­fy prom­ising drug can­did­ates in weeks rather than years, to cut­ting-edge CRISPR gene ther­apies that lit­er­ally rewrite genet­ic code. These rad­ic­al advances don’t just improve exist­ing solu­tions, they com­pletely redefine what’s pos­sible. Each rep­res­ents a fun­da­ment­al shift that cre­ates entirely new mar­kets and renders pre­vi­ous approaches obsolete.

Yet for every break­through that makes head­lines, count­less ground­break­ing ideas nev­er see the light of day, bur­ied with­in organ­isa­tions that inad­vert­ently sup­press their own poten­tial. While some com­pan­ies like Mich­elin embrace open innov­a­tion through hun­dreds of extern­al R&D part­ner­ships that tap into glob­al expert­ise, and oth­ers like Microsoft under­take dra­mat­ic cul­tur­al trans­form­a­tions from “know-it-all” to “learn-it-all” envir­on­ments, many organ­isa­tions still struggle to nur­ture truly trans­form­at­ive ideas. What sep­ar­ates the break­through innov­at­ors from the incre­ment­al improvers?

The answer lies not just in R&D budgets or bril­liant sci­ent­ists, but in some­thing more subtle yet power­ful: the social net­works that con­nect people with­in organ­isa­tions. Net­work theory—the study of rela­tion­ships and connections—reveals that the way people col­lab­or­ate fun­da­ment­ally shapes wheth­er an organ­isa­tion pro­duces revolu­tion­ary ideas or settles for mod­est improvements.

Two com­pre­hens­ive stud­ies of US phar­ma­ceut­ic­al and bio­tech­no­logy firms ana­lysed over 19,000 pat­ent records from 93 lead­ing com­pan­ies between 2001 and 2013. This massive data­set allowed research­ers to trace the social con­nec­tions behind each innov­a­tion, map­ping who col­lab­or­ated with whom and identi­fy­ing the net­work pat­terns that pre­ceded both break­through and incre­ment­al innov­a­tions. Their find­ings chal­lenge con­ven­tion­al wis­dom about what drives innov­a­tion and offer sur­pris­ing insights about collaboration’s role in creativity.

The research uncovered a fas­cin­at­ing para­dox: the same tight-knit rela­tion­ships that excel at incre­ment­al innov­a­tion can actu­ally sup­press rad­ic­al break­throughs. This coun­ter­in­tu­it­ive find­ing has pro­found implic­a­tions for how we think about team design and col­lab­or­a­tion. In short, net­works that drive incre­ment­al suc­cess, some­times stifle rad­ic­al think­ing at the same time.

Dense, inter­con­nec­ted teams built on trust and fre­quent com­mu­nic­a­tion excel at refin­ing exist­ing ideas and imple­ment­ing improve­ments effi­ciently. These close-knit groups share tacit know­ledge seam­lessly, coordin­ate com­plex efforts with min­im­al fric­tion, and can rap­idly iter­ate on proven con­cepts. Think of how Apple’s hard­ware and soft­ware teams work in tight coordin­a­tion to deliv­er incre­ment­al improve­ments in each iPhone generation—better cam­er­as, faster pro­cessors, refined designs that build on estab­lished foundations.

How­ever, these same net­works tend to main­tain estab­lished ways of think­ing and res­ist ideas that chal­lenge exist­ing norms or threaten cur­rent suc­cess for­mu­las. Group mem­bers devel­op shared men­tal mod­els and sim­il­ar per­spect­ives, cre­at­ing what research­ers call “cog­nit­ive lock-in.” When every­one thinks alike, truly nov­el ideas can seem for­eign or risky. Con­sider Kodak’s engin­eers, who developed the first digit­al cam­era in 1975 only to see it dis­missed by lead­er­ship as merely “cute”, a clas­sic example of how entrenched think­ing with­in tight-knit net­works can blind organ­iz­a­tions to revolu­tion­ary potential.

The data backs this up with strik­ing math­em­at­ic­al pre­ci­sion: regres­sion ana­lyses show that stronger ties and dens­er net­works sig­ni­fic­antly reduce rad­ic­al innov­a­tion, with coef­fi­cients of –2.29 and –1.77 respect­ively. In prac­tic­al terms, this means that every increase in net­work dens­ity cor­res­pond­ingly decreases the like­li­hood of pro­du­cing break­through innovations.

The most rad­ic­al innov­a­tions emerge from unex­pec­ted places: the edges of organ­isa­tion­al net­works. These peri­pher­al positions—less con­nec­ted to the organ­isa­tion­al core but more open to out­side influences—prove sig­ni­fic­antly more likely to gen­er­ate trans­form­at­ive ideas. Without the pres­sure to con­form to estab­lished group think­ing, people at these net­work edges can pur­sue truly nov­el approaches and chal­lenge fun­da­ment­al assumptions.

This phe­nomen­on explains why some of his­tory’s most revolu­tion­ary innov­a­tions came from out­siders or bound­ary-span­ners. The per­son­al com­puter emerged not from IBM’s cent­ral research labs but from gar­age-based entre­pren­eurs on the peri­phery of the tech­no­logy estab­lish­ment. Sim­il­arly, many bio­tech­no­logy break­throughs come from aca­dem­ic-industry col­lab­or­a­tions where research­ers oper­ate at the inter­sec­tion of dif­fer­ent know­ledge domains.

The under­ly­ing mech­an­ism makes intu­it­ive sense when we con­sider inform­a­tion flows. Strong ties, while bene­fi­cial for trust and com­mu­nic­a­tion, often con­nect sim­il­ar people who share redund­ant inform­a­tion and sim­il­ar per­spect­ives. Weak ties, des­pite their rela­tion­al chal­lenges like reduced trust and poten­tial mis­un­der­stand­ings, provide access to fresh, non-redund­ant inform­a­tion that can spark break­through think­ing. These con­nec­tions act as bridges between dif­fer­ent worlds of know­ledge, bring­ing togeth­er insights that would nev­er meet with­in dense, homo­gen­eous networks.

These insights offer tech­no­logy lead­ers a prac­tic­al roadmap for man­aging innov­a­tion, sug­gest­ing that dif­fer­ent types of innov­a­tion require fun­da­ment­ally dif­fer­ent organ­isa­tion­al approaches:

For incre­ment­al innov­a­tion: Foster dense, col­lab­or­at­ive net­works with strong intern­al bonds. Encour­age fre­quent face-to-face inter­ac­tions, build deep trust between team mem­bers, and cre­ate envir­on­ments where exist­ing know­ledge can be sys­tem­at­ic­ally refined and effi­ciently imple­men­ted. Estab­lish clear com­mu­nic­a­tion chan­nels, reg­u­lar review pro­cesses, and shared suc­cess met­rics. This approach works excep­tion­ally well for product improve­ments, pro­cess optim­isa­tion, and per­form­ance enhancements.

For rad­ic­al innov­a­tion: Delib­er­ately cul­tiv­ate net­work diversity and struc­tur­al gaps. Avoid overly tight-knit groups that rein­force con­ven­tion­al think­ing. Instead, cre­ate “struc­tur­al holes” (stra­tegic gaps in social net­works that force dif­fer­ent groups to con­nect and share fresh per­spect­ives). Encour­age col­lab­or­a­tions across depart­ments, indus­tries, and even com­pet­it­ors. Sup­port employ­ees who serve as “know­ledge brokers”, con­nect­ing pre­vi­ously unlinked groups and trans­lat­ing insights across domains.

The crit­ic­al insight is that organ­isa­tions need both approaches oper­at­ing sim­ul­tan­eously. While loose con­nec­tions provide access to nov­el ideas and diverse per­spect­ives, teams still need enough strong ties to effect­ively eval­u­ate, devel­op, and imple­ment new con­cepts. This means act­ively sup­port­ing cross-depart­ment­al col­lab­or­a­tions, extern­al part­ner­ships, sab­bat­ic­als in dif­fer­ent indus­tries, and hir­ing people from out­side the sec­tor who bring fresh viewpoints.

Prac­tic­ally, this might involve cre­at­ing ded­ic­ated “innov­a­tion out­posts” in dif­fer­ent geo­graph­ic regions or sec­tors, estab­lish­ing form­al rota­tion pro­grams that move people between depart­ments, or insti­tut­ing “innov­a­tion tour­na­ments” that bring togeth­er diverse teams to tackle spe­cif­ic challenges.

Lead­ing organ­isa­tions must go bey­ond net­work design to address cul­tur­al bar­ri­ers. This means incentiv­ising exper­i­ment­a­tion, tol­er­at­ing ambi­gu­ity and fail­ure, and reward­ing cal­cu­lated risk-taking—all essen­tial for over­com­ing the nat­ur­al con­ser­vat­ism of close-knit net­works. By under­stand­ing the dual dynam­ics of net­work cohe­sion and broker­age, man­agers can archi­tect organ­iz­a­tion­al struc­tures that nur­ture both steady improve­ment and trans­form­at­ive breakthroughs.

The research offers some­thing rare in busi­ness strategy: a clear, evid­ence-based roadmap for one of man­age­ment’s greatest chal­lenges. In an era where revolu­tion­ary ideas can reshape entire indus­tries overnight, under­stand­ing the social forces behind innov­a­tion isn’t just aca­dem­ic, it’s essen­tial for survival.

For more info

• Jia Zhang, Jian Wang, Jos Win­nink, Simcha Jong (2024). Turn­ing cre­at­ive ideas into suc­cess­ful innov­a­tions: dif­fer­en­tial effects of net­work struc­ture for rad­ic­al and incre­ment­al innov­a­tion. Journ­al of Tech­no­logy Transfer.
• Jia Zhang, Jian Wang, Jos Win­nink, Simcha Jong (2025). Col­lab­or­a­tion net­works and rad­ic­al innov­a­tion: Two faces of tie strength and struc­tur­al holes. Journ­al of Informetrics.