You’ve re­cently joined a ma­jor or­ga­ni­za­tion in a se­nior man­age­ment role. How can you or­ga­nize your plans?

One sim­ple way to think about them is with what can be called the “Sta­bi­lize-Reflect-Ex­e­cute” cy­cle.


You first check if there are any ur­gent is­sues and ad­dress them im­me­di­ately. Are there burn­ing prob­lems or op­por­tu­ni­ties that need to be dealt with? Se­cond, you do any­thing you need to do to best pre­pare your­self for re­flec­tion. If there are peo­ple you need to talk to in or­der to get nec­es­sary in­for­ma­tion, you set that up up­front.


Once ur­gent is­sues are dealt with and you are able to prop­erly ac­cess the situ­a­tion, you work to do so. For ex­ec­u­tives this can mean a lengthy pe­riod of dis­cus­sions with all of the rele­vant peo­ple and thoughts on strat­egy be­fore mak­ing for­mal an­nounce­ments. This could take a few weeks or months.


Now is the time to be­gin work­ing on non-ur­gent im­por­tant prob­lems, which should be the main ones. You fol­low through with your re­flec­tion. Ex­e­cu­tion may in­volve de­cid­ing on pur­su­ing fu­ture larger sta­bi­lize-re­flect-ex­e­cute loops.

Let’s sum­ma­rize. “Sta­bi­lize” refers to han­dling ur­gent is­sues and prepar­ing for re­flec­tion. This is similar to the no­tion of get­ting one’s “house in or­der.” “Reflect” refers to de­cid­ing how to best deal with the im­por­tant non-ur­gent is­sues. “Ex­e­cute” refers to work­ing on the im­por­tant is­sues. This is ba­si­cally a sub­set of the Eisen­hower Method for situ­a­tions where these three steps make up the ma­jor­ity of the work.


I think this cy­cle plays out in many im­por­tant situ­a­tions, so may be worth some in­de­pen­dent study. Some ex­am­ples of these cy­cles in­clude:

Ne­c­es­sary Conditions

The sta­bi­lize-re­flect-ex­e­cute cy­cle is good for some spe­cific situ­a­tions. I think it may re­quire the fol­low­ing:

1. There are some tasks that are both ur­gent and im­por­tant.

If this is not true, the “sta­bi­lize” step isn’t nec­es­sary.

2. There are some tasks that are both non-ur­gent and im­por­tant.

If this is not true, the “re­flect” and “ex­e­cute” steps aren’t nec­es­sary.

3. There is sig­nifi­cant ex­pected value for spend­ing time figur­ing out how to do the non-ur­gent and im­por­tant tasks.

If this is not true, the “re­flect” step isn’t nec­es­sary.

4. Figur­ing out how to do the im­por­tant tasks can be done in one batch per full cy­cle.

If this is not true, then the work may be a bit of a mess of sta­bi­liza­tions, re­flec­tions, and ex­e­cu­tions, as op­posed to one se­rial pro­cess.


We could use this model to com­pare use cases

I think that differ­ent use cases (like the ex­am­ples above) share a good amount of similar­i­ties. I hope that they could be seen as such, and used to help un­der­stand each other. One may think that we don’t have may similar refer­ences classes to things like AI take­offs and global gov­er­nance. While this may be gen­er­ally true, I hope that this similar­ity could provide a bit of a counter.

We should ac­knowl­edge un­cer­tainty of the “Ex­e­cute” stage

One try­ing to un­der­stand the sta­bi­lize-re­flect-ex­e­cute of an ac­tor may try to pre­dict the ac­tions of the ex­e­cu­tion stage, but I this should be re­garded with skep­ti­cism. The point of the re­flect stage is to bet­ter un­der­stand how to en­act the ex­e­cute stage; its’ pres­ence sug­gests that the ex­e­cu­tion stage could go in differ­ent ways. This means that a lot of at­ten­tion on sta­bi­lize-re­flect-ex­e­cute pro­cesses should be on the first two parts, rather than the third.

Com­par­i­son to the Eisen­hower Method

Image from James Clear

The ur­gent /​ im­por­tant dis­tinc­tion comes from the Eisen­hower Ma­trix. That ma­trix is sup­posed to ap­ply to more situ­a­tions, so is more generic. Work around the ma­trix typ­i­cally recom­mends “quickly do­ing” the “im­por­tant & ur­gent” tasks, then “plan­ning” the “im­por­tant & non ur­gent” tasks.

Com­par­i­son to De­ci­sion Cycles

Wikipe­dia lists sev­eral ex­am­ples of in­ter­est­ing de­ci­sion cy­cles. Th­ese are se­quences with spe­cific steps for de­ci­sion mak­ing. For in­stance, in qual­ity con­trol, “PDCA (Plan-Do-Check-Act) is used.” I be­lieve the de­ci­sion cy­cles typ­i­cally in­clude some kind of learn­ing com­po­nent, which is ab­sent in the sta­bi­lize-re­flect-ex­e­cute cy­cle.

Many thanks to Toby Ord for dis­cus­sion & in­spira­tion, and to Owen Cot­ton-Bar­ratt and Max Daniel for feed­back.