Optimizing for Stories (vs Optimizing Reality)

Epistemic sta­tus: highly con­fi­dent in the ba­sic dis­tinct, though it’s not at all profound. Fur­ther de­tails, mod­els, and ad­vice are some­what spec­u­la­tive de­spite be­ing drawn from vary­ing amounts of ob­ser­va­tion. Es­say a lit­tle rushed since oth­er­wise I’d be un­likely to pub­lish.

When set­ting out on a ven­ture, one faces some choices.

  • Op­ti­mize re­al­ity: How much do you op­ti­mize for the suc­cess of your stated goal? I define goal to be some de­sired state of re­al­ity such that suc­cess or failure is as­sessed with re­spect to re­al­ity.

  • Op­ti­mize sto­ries: How much do you op­ti­mize for the ap­pear­ance of suc­cess at your stated goal, i.e. sto­ries? I define a story as a col­lec­tion of facts about a re­al­ity as­serted to be true and rele­vant, usu­ally pre­sented to oth­ers but some­times only to one­self.

Ideally, op­ti­miz­ing for one would be the same as op­ti­miz­ing for an­other. Very of­ten, how­ever, op­ti­miz­ing one is not the same as op­ti­miz­ing for other. What’s worse, the two ends com­pete for the same set of limited re­sources.

A con­crete ex­am­ple might be some­one who sets out to de­velop and sell a medici­nal tea which re­lieves hang­over symp­toms. Over­all suc­cess is sel­l­ing your tea and mak­ing money. In­vest­ing in op­ti­miz­ing re­al­ity would mean in­vest­ing in ex­per­i­ments to de­velop and im­prove the tea. In­tro­duc­ing var­i­ants, ran­dom­ized con­trol­led tri­als, etc., etc. Op­ti­miz­ing story means hav­ing a re­ally good ex­pla­na­tion for why your tea is able to do what you claim it does, plus hav­ing a great web­site with good copy, pub­lish­ing tes­ti­mo­ni­als, pub­lish­ing your method­ol­ogy and ex­per­i­men­tal re­sults, etc.

Suc­cess can be at­tained by both, in com­bi­na­tion, but also pos­si­bly each on its own. If you have a tea which works well, word will spread and you’ll end up with many cus­tomers seek­ing your gen­uinely pal­li­a­tive tea. Alter­na­tively, if your mar­ket­ing ma­te­ri­als are per­sua­sive, you might ac­crue many cus­tomers too, even if your tea is no bet­ter than placebo. Of course, truly effec­tive tea plus a well-con­veyed story about its great prop­er­ties will gen­er­ate more sales than effec­tive tea or a good story alone.

Op­ti­miz­ing Reality

The fol­low­ing points are worth not­ing:

  • When op­ti­miz­ing re­al­ity, suc­cess de­pends on re­al­ity at large be­ing a cer­tain way.

  • There is only one re­al­ity when it comes to what you’re op­ti­miz­ing for.

  • You can’t fool re­al­ity. It is what it is no mat­ter what you say or do.

  • Op­ti­miz­ing for re­al­ity is eas­iest to do when out­comes are easy to mea­sure and feed­back is quick.

  • Eng­ineers deal­ing with con­crete, mea­surable phe­nom­ena are likely to be op­ti­miz­ing re­al­ity.

  • Gen­uinely cre­at­ing value, though of­ten ex­pen­sive, is a good strat­egy for cap­tur­ing value to your­self. Likely it’s the best long-term strat­egy in many do­mains.

    • I would claim that ev­ery ma­jor suc­cess­ful com­pany is pro­duc­ing clear value to some­one. Face­book, Google, AirBnB, ExxonMo­bil, United Air­lines, Star­bucks. At the end of the day, there’s some­thing real there.

Op­ti­miz­ing Stories

  • For any goal, there might be many pos­si­ble sto­ries which will lead to suc­cess.

  • Sto­ries don’t have to be true or ac­cu­rate to be effec­tive.

  • It can be ra­tio­nal, or out­right nec­es­sary, for one’s suc­cess to op­ti­mize for sto­ries with­out re­gard to re­al­ity.

    • If your suc­cess ul­ti­mately de­pends on some­one else be­ing con­vinced, then re­al­ity at large doesn’t mat­ter.

  • Sales­peo­ple, mar­keters, poli­ti­ci­ans, startup founders and gen­er­ally those whose suc­cess de­pends on per­suad­ing oth­ers will spend much of their effort op­ti­miz­ing sto­ries.

  • Very of­ten your “cus­tomer” only cares that you have a “good story” and not at all that your story matches re­al­ity.

    • Con­sider a re­sel­ler of your medici­nal tea. If they’re un­scrupu­lous, it doesn’t mat­ter to them whether or not the tea works. It only mat­ters to them that your story is good enough to per­suade end-con­sumers.

    • Con­sider a hos­pi­tal IT em­ployee whose job is to pur­chase soft­ware which helps doc­tors. The sup­posed goal of im­proved doc­tor effi­ciency is not mea­sured in any way. The hos­pi­tal IT em­ployee’s suc­cess will be judged by the im­pres­sive-seem­ing­ness of the soft­ware they se­lect—how good the story it comes with—not any ac­tual re­al­ity. In other words, the IT em­ployee, as a cus­tomer, is in­cen­tivized to be shop­ping for good sto­ries and only good sto­ries.

      • But you needn’t as­sume peo­ple only pur­chase sto­ries for the sake of oth­ers! Many peo­ple, hav­ing a loose re­la­tion­ship, are happy to pur­chase a story which makes them feel good. “Oh, those crys­tals from the mys­tery moun­tains have the right fre­quen­cies to bring me good for­tune? Yes please!”

  • In cases where those hear­ing the sto­ries lack the ex­per­tise to as­sess re­al­ity, e.g. non-ex­perts try­ing as­sess an ex­pert, “dumbed-down” sto­ries com­pre­hen­si­ble to the non-ex­perts will com­pletely out­weigh re­al­ity it­self. Cf. Over­con­fi­dent talk­ing down, hum­ble or hos­tile talk­ing up.

  • If it’s only sto­ries which mat­ter, yet you split your efforts be­tween sto­ries and re­al­ity, then you will likely be out­com­peted by some­one who spent all of their re­sources on craft­ing good sto­ries. Cf. Moloch.

  • Even those who care a lot about re­al­ity it­self can slide into a fo­cus on sto­ries if em­piri­cal feed­back is ab­sent or very slow.

  • Notwith­stand­ing all of the above, sto­ries will some­times reach the end of their tether and the lack of a good re­al­ity to sup­port your sto­ries will catch up with you.

  • We don’t just tell sto­ries to oth­ers, we also tell them­selves to our­selves.

    • The dan­ger of be­ing too much in the habit of tel­ling sto­ries is that we don’t merely risk­ing fool­ing oth­ers, but also our­selves.

“Story Economies”

I con­jec­ture that what arises in our mod­ern world are “economies” of sto­ries whereby peo­ple buy and sell sto­ries of­ten with­out re­gard for re­al­ity.

Another ex­am­ple: imag­ine an an­a­lyst work­ing at a startup crafts a re­port which high­lights all the ways in which the com­pany is rapidly im­prov­ing. The an­a­lyst’s man­ager isn’t too wor­ried about whether the re­port is a bit bi­ased to­wards the pos­i­tive—they know the CEO will be pleased. The CEO doesn’t mind if the re­port is a bit bi­ased to­wards the pos­i­tive—they know the board will be pleased. The board doesn’t mind if the re­port is a bit bi­ased—they know that the next round of in­vestors won’t re­ally be able to tell the differ­ence, it will just make the com­pany look good.

Here you have a whole chain of peo­ple who only care about the story. At the very end there’s some­one who cares about the re­al­ity, but they’re very of­ten not in a great po­si­tion to eval­u­ate it them­selves. They prob­a­bly don’t know even the right ques­tions to ask. All they’ve got is the story which has been placed be­fore them.

Some peo­ple grav­i­tate more to­wards sto­ries than oth­ers, e.g. sales­peo­ple and poli­ti­ci­ans. Some of them might read­ily ad­mit that they chiefly deal in sto­ries some­what ten­u­ously linked to re­al­ity, yet I wa­ger that many, if not most, won’t. The most per­sua­sive sto­ries are those you de­voutly be­lieve your­self. Hence the vast over­con­fi­dence of startup founders. And, in the im­mor­tal words of Ge­orge Costanza: it’s not a lie… if you be­lieve it.

Sto­ries about your­self …to your­self and others

If there’s one do­main where we’re end­lessly craft­ing and broad­cast­ing sto­ries, it’s the sto­ries we tell about our­selves. I’m this kind of per­son. I might de­cide that the story I want to tell is that I’m a “sci­ence nerd”. So I read sci­ence books and sci­ence mag­a­z­ines. I have my an­swer ready when peo­ple ask me what I do for fun and I know ex­actly what to post on so­cial me­dia. My In­sta­gram is full of home­made vol­ca­noes and pho­tos from my per­sonal back­yard telescope.

The above fic­tional ex­am­ple might have the re­deem­ing fea­ture that at least this fic­tional per­son is cre­at­ing a gen­uine re­al­ity to match the story. They are learn­ing a tonne of sci­ence genre facts. Still, I wa­ger there’s a trade­off. Do­ing sci­ence-y things which are eas­ily com­mu­ni­ca­ble and demon­stra­ble in­tro­duces a con­straint. Pos­si­bly lev­el­ing up as a sci­en­tist would mean read­ing text­books with facts that are in­com­pre­hen­si­ble and bor­ing to those not at that level. By try­ing to have the best story to tell, they’ve hand­i­capped their own ex­cel­lence. (How­ever, if the story is pri­mar­ily for one­self, this con­straint is avoided. “I know just how sci­ence-y I am!)

If you want to know, I tell my­self the story that I’m a per­son who’s afraid of los­ing my­self to try­ing craft my­self into some­one op­ti­mized for im­press­ing oth­ers. Though I do it. I’m do­ing it right now. There may be no es­cape.

Cf. Elephant in the Brain.

You can’t es­cape stories

At this point you might be think­ing, “gee, sto­ries are awfully de­ceit­ful and non-co­op­er­ate, I want to co­op­er­a­tive and hon­est and I’m just go­ing to provide di­rect facts!” and “I re­ally, re­ally don’t want to de­ceive my­self with sto­ries!”

I don’t think you can es­cape sto­ries en­tirely. I would claim that as soon as you sum­ma­rize your facts or data, the mere se­lec­tion of which facts to pre­sent or sum­ma­rize is the craft­ing of a story. Even dump­ing all your data and ev­ery ob­ser­va­tion is likely to be bi­ased by which data you col­lected and what you paid at­ten­tion to. What you thought were the rele­vant things to re­port to an­other per­son.

That said, I think there’s sto­ry­tel­ling which at­tempts to be hon­est effort to share re­al­ity as is so that some­one else can make an in­formed judg­ment. It’s challeng­ing if one’s suc­cess is threat­ened by less scrupu­lous com­peti­tors, but it’s pos­si­ble to choose do­mains where mea­surable feed­back fa­vors those who’ve op­ti­mized ac­tual re­al­ity.

You’re not always do­ing oth­ers a fa­vor if you try to give them raw facts with no bi­ased con­clu­sions. The world is large and messy and con­fus­ing such that peo­ple usu­ally like to be handed a story about who you are and how you will be­have. They want you to be a nerdy, book­ish type, or an out­doorsy type, or a foodie. If you give me a story and promise to act in ac­cor­dance with it, that makes sim­ple. It’s clear what to talk about, what to get you for your birth­day, etc., etc.

At least for those spend­ing much time out in main­stream cul­ture, it helps to have one or two sto­ries pre­pared about your­self. “Masks.” They func­tion a bit like APIs, re­ally. Peo­ple of­ten protest that they don’t like be­ing put in boxes, but those boxes help you re­late to peo­ple be­fore you’ve spent the many, many hours to have ab­sorbed the messy re­al­ity that any given hu­man is.

What to do, what to do

Real­ity on the ground is com­plex, in­cen­tives are messy, things which work in the short run don’t nec­es­sar­ily work in the long run. I can’t say “here’s my one sim­ple recipe to de­ter­mine the right al­lo­ca­tion of re­sources be­tween op­ti­miz­ing story vs op­ti­miz­ing re­al­ity.”

I proffer the ob­vi­ous ad­vice:

  • No­tice the in­cen­tives for each do­main you deal in—how much does your suc­cess de­pend on sto­ries vs di­rect re­al­ity?

  • By judg­ing the in­cen­tives in the do­mains you deal, as­sess how much you can judge the sto­ries you are pre­sented with.

  • Ac­cept that the trade­offs are hard. Per­son­ally, I wish I could deal only in truth and provide only open and trans­par­ent facts. Un­for­tu­nately, I might need ei­ther to com­pro­mise or to find my­self severely dis­ad­van­taged in cer­tain are­nas, e.g. poli­tics.

    • More­over, the in­cen­tives mean de­spite my­self, self-in­ter­est bias means I’m likely to pre­sent things to oth­ers in ways which fa­vor my­self.

  • Ac­cept that you prob­a­bly need a story about your­self. If you like, keep the story sep­a­rate from your­self so that you might let your­self be more. Cf. Keep your iden­tity small.