Don’t scale agile – descale your organization!

Agile orga­ni­za­tions are lean, flex­i­ble and adapt­able. This is accom­plished less by “agiliz­ing” exist­ing struc­tures and process­es but rather by con­sis­tent­ly ques­tion­ing these. The solu­tion, of course, is not to fall vic­tim to the cog­ni­tive bias of Maslow’s ham­mer and to apply the shiny new agile meth­ods to all exist­ing cal­ci­fied orga­ni­za­tion­al struc­tures and their waste­ful pro­ce­dures. Rather the mot­to is “Don’t scale agile – descale your organization!”

Form Without Function

At the lev­el of the indi­vid­ual team or a few teams work­ing on a sin­gle prod­uct, agili­ty is still quite straight­for­ward. Scrum as the best known and most used frame­work is well described in the Scrum Guide and has proven its val­ue prac­ti­cal­ly over two decades. It starts to get excit­ing when we go beyond that lev­el and when tay­loris­tic orga­ni­za­tions that have grown and become encrust­ed over decades try to become more agile. When viewed from a dis­tance it often hap­pens what Richard Feyn­man described as a “car­go cult”: form with­out func­tion.

In the South Seas there is a car­go cult of peo­ple. Dur­ing the war they saw air­planes land with lots of good mate­ri­als, and they want the same thing to hap­pen now. So they’ve arranged to imi­tate things like run­ways, to put fires along the sides of the run­ways, to make a wood­en hut for a man to sit in, with two wood­en pieces on his head like head­phones and bars of bam­boo stick­ing out like anten­nas — he’s the con­troller — and they wait for the air­planes to land. They’re doing every­thing right. The form is per­fect. It looks exact­ly the way it looked before. But it doesn’t work. No air­planes land. So I call these things car­go cult sci­ence, because they fol­low all the appar­ent pre­cepts and forms of sci­en­tif­ic inves­ti­ga­tion, but they’re miss­ing some­thing essen­tial, because the planes don’t land.

Richard Feyn­man, 1974

Law of the Instrument

So some­one at the top end of the orga­ni­za­tion’s hier­ar­chy has decid­ed that from now on they want to be agile. Against the back­ground of VUCA and dig­i­tal­iza­tion, this sounds quite plau­si­ble and nobody can object to a bit more self-orga­ni­za­tion and self-respon­si­bil­i­ty amidst all the bureau­cra­cy that has devel­oped over time. So they embark on their jour­ney, read books, attend train­ing cours­es, obtain cer­ti­fi­ca­tions, employ hordes of coach­es and con­sul­tants, vis­it Spo­ti­fy, trav­el to Sil­i­con Val­ley and so on.

Grad­u­al­ly every­thing is turned upside down and changed to agile. Of course, in most cas­es with­out sub­stan­tial­ly going to the struc­ture of the orga­ni­za­tion and the accus­tomed pro­ce­dures. Agile projects do not lead to agile orga­ni­za­tions if they are embed­ded in encrust­ed struc­tures, from the approval of the project to the oblig­a­tory steer­ing committees.

In order to breathe a lit­tle more agili­ty into this par­a­lyzed cir­cus, every­thing gets “agilized” – no mat­ter how well it is suit­ed. Com­mit­tees then sud­den­ly become agile com­mit­tees that work in sprints or with Kan­ban. And the process mod­el also gets a prod­uct own­er and teams. Done.

I sup­pose it is tempt­ing, if the only tool you have is a ham­mer, to treat every­thing as if it were a nail.

Abra­ham Maslow

This insid­i­ous form of car­go cult is based on a cog­ni­tive bias called Law of the Instru­ment or Maslow’s Ham­mer. The shiny new agile meth­ods are applied to all exist­ing struc­tures and process­es with­out ques­tion­ing them and rethink­ing them in terms of customer ori­en­ta­tion and the opti­mal val­ue stream for it. Because then the prop­er con­clu­sion would be that these com­mit­tees are per­haps not need­ed at all and that the process mod­el will be fur­ther devel­oped in a suit­able(!) way in a com­mu­ni­ty approach by those who work with­in this model.

Some­thing is wrong if work­ers do not look around each day, find things that are tedious or bor­ing, and then rewrite the pro­ce­dures. Even last month’s man­u­al should be out of date.

Tai­ichi Ōno

---

Pho­to: Russ Hen­dricks