Press "Enter" to skip to content

The Death of Agile (Allen Holub)


00:00:14so I’m going to change tone a little bit

00:00:18here because I’m not going to talk about

00:00:18code at all I’m going to talk about how

00:00:20to produce code and in particular I’m

00:00:22going to talk about agile not because

00:00:26well let’s put it this way I think agile

00:00:29is by far the best way that we know to

00:00:31produce software and at the same time I

00:00:33think the notion of agile software

00:00:35development is being completely

00:00:36destroyed by most of the agile community

00:00:39so when I talk about the death of agile

00:00:42what I’m really talking about is the

00:00:43death of what I think of as real agility

00:00:45and a replacement of real agility with a

00:00:48kind of fake agility that is suitable

00:00:51for corporations but doesn’t really work

00:00:52in the real world so let me explain

00:00:54exactly what I’m talking about so I

00:00:57would explain it if my clicker was

00:00:58working there it is good so the reason

00:01:00I’m doing this talk is because as a

00:01:02consultant I often come into companies

00:01:04and give them advice about agile process

00:01:06and I am just getting sick to death of

00:01:09dealing with the aftermath of the

00:01:11certified scrum guys and right is that

00:01:15they come in they do their thing and

00:01:17they leave this behind them and the the

00:01:21real there are a lot of issues

00:01:23surrounding that but scrum is really a

00:01:26problem is the problem it is not the

00:01:28solution and the basic problem is the

00:01:30notion of a cargo called now the cargo

00:01:32cults happened in the Pacific Islands

00:01:35during World War two is they’ve

00:01:37Islanders are living happily on their

00:01:38islands without any cares in the world

00:01:41eating fruit as a drops off of trees but

00:01:43they were living in the Stone Age a few

00:01:45minutes later it seems a bunch of

00:01:47soldiers land on the island they make a

00:01:49runway and then airplanes start landing

00:01:51and the airplanes bring with them the

00:01:5420th century so all of a sudden these

00:01:56guys have gone from the Stone Age to the

00:01:5820th century because of these air bases

00:02:00and five years later all of those

00:02:02soldiers get back onto their airplanes

00:02:04and they fly away and the people who

00:02:06lived on the island are back in the

00:02:08stone Age’s and what they saw were the

00:02:10airplanes is what they said if we could

00:02:12only bring the airplane

00:02:13they’re going to bring back the cargo so

00:02:16what they started doing is building

00:02:18things that look like airplanes in the

00:02:20hopes that if they build it they will

00:02:22come

00:02:22if the airplanes flying overhead see

00:02:25their brethren on the ground they’re

00:02:26going to land and bring the cargo with

00:02:28them which of course didn’t happen now

00:02:32the problem with a cargo cult then is

00:02:34what they’re doing is they’re going

00:02:35through the motions as the guys in the

00:02:36cargo cults would build fake conning

00:02:39towers and they would wear coconuts over

00:02:41their ears pretending to be head phones

00:02:42and they would act as if they were

00:02:44working Airport in the hopes that the

00:02:46airport airplanes would land and the

00:02:49problem is that we do this all the time

00:02:51as we see cargo cults everywhere and the

00:02:54biggest cargo cult of all in fact is the

00:02:57modern corporation is that corporations

00:03:00are cargo cults because they believe

00:03:02that there are certain ways to do things

00:03:03and they cannot conceive that there is

00:03:06another way of doing things I

00:03:07participate a lot in one of the LinkedIn

00:03:09agile groups and one of the

00:03:12characteristics of this group is that

00:03:13there’s a constant stream of wrong

00:03:15questions there’s a constant stream of

00:03:16questions that have that are so wrong

00:03:18that you don’t even understand how

00:03:19someone can ask them in the context of

00:03:21agile somebody will always say why ever

00:03:23every couple weeks somebody posts a

00:03:25question what’s the role of a project

00:03:27manager on the agile team and the answer

00:03:30is there is no project manager on the

00:03:32agile team but people don’t understand

00:03:34that is that they’re so convinced

00:03:36because that’s the way the corporation’s

00:03:38think that a project manager is a

00:03:40necessary part of work that they cannot

00:03:42imagine a world in which a project

00:03:44manager does not exist so like the

00:03:46proverbial camel here is that a

00:03:49corporation cannot fit through an eye of

00:03:51the needle is that it can’t make the

00:03:53changes that it needs to make because

00:03:54it’s a cargo call because it does things

00:03:56over and over again by rote without

00:03:58really understanding why it’s doing any

00:04:00of the things that it does in particular

00:04:03what they’re looking for is a very

00:04:05specific process that they can follow

00:04:07the corporation’s are really into

00:04:08process is that what they want to do is

00:04:11they want to do something easy they want

00:04:12to bring in the trainer’s they want to

00:04:13train a few people in the engineering

00:04:15department to follow some process and

00:04:17they think that by following that

00:04:18process that there will be some sort of

00:04:20good result but in fact you don’t get a

00:04:22good result by following a specific

00:04:24process it just doesn’t work in other

00:04:27is what the corporation’s want is this

00:04:29kind of agile right they want a canned

00:04:32thing right something that you can just

00:04:35roll open up the can and do it and it’s

00:04:38somehow going to magically change the

00:04:40way that things work in a positive way

00:04:42but often when they bring in the agile

00:04:44this kind of agile they they don’t

00:04:47really want to change the way things

00:04:48work they just read about agile

00:04:50someplace and they think it’s a good

00:04:51thing and it’s think that there’s going

00:04:52to be some kind of benefit but they

00:04:54don’t really know what the benefits are

00:04:55so this is just really really a bad idea

00:04:57now what is agile really agile is

00:05:01defined by the agile manifesto and the

00:05:05agile manifesto as you can see here is

00:05:06very simple there’s only four points on

00:05:08it and the first one is the most

00:05:09interesting one is that the way that

00:05:12people work with one another is more

00:05:14important than processes now notice that

00:05:16it doesn’t say instead of process is

00:05:19important

00:05:20agile has a lot of process associated

00:05:22with it and the processes help but the

00:05:26processes are not canned things are

00:05:27things that the individuals come up with

00:05:29as they interact with each other in

00:05:32other words the word agility has to

00:05:33apply to the process as much as it

00:05:35applies to anything else if you’re not

00:05:37agile in the way that you build software

00:05:40then you’re not going to be agile so the

00:05:43whole notion of individuals and

00:05:45interactions over processes and tools is

00:05:46kind of fallen by the wayside as scrum

00:05:48is a process and a lot of people believe

00:05:51that scrum and agile are the same thing

00:05:52and that’s just wrong scrum is a process

00:05:56so by saying we’re doing scrum what

00:05:58you’re saying is that the process is

00:05:59more important than the people that are

00:06:00doing the process and that just is not

00:06:02going to work in any kind of situation

00:06:04now my rule of thumb here then is to

00:06:07make sure to see if something makes

00:06:08sense in any sentence where you see the

00:06:10word agile if you can replace that word

00:06:12with the word flexible then you’ve got a

00:06:14proper sentence if you say we things we

00:06:17do things in an agile way that works for

00:06:19me if you say agile says that we should

00:06:23do X that doesn’t work for me at all

00:06:25what we want is flexibility the whole

00:06:28point of a of agile is to get agility in

00:06:31other words what we want to be able to

00:06:34do is we want to quickly adapt to

00:06:35changing requirements be those changes

00:06:38coming from inside or outside either the

00:06:40customer is mandating them

00:06:41or we are mandating them because there’s

00:06:43some process internal process things of

00:06:45technology thing that changes and we

00:06:46need to change internally but what we

00:06:48need is agility we need flexibility we

00:06:50need the ability to change in a

00:06:51corporation at least many corporations

00:06:53are incapable of doing that so I suppose

00:06:56the one thing that I will agree with the

00:06:59agile skeptics about is that I think the

00:07:01vast majority of large companies cannot

00:07:03ever be agile because it is the very

00:07:05nature of those large companies to not

00:07:07be flexible about things in flexibility

00:07:12is just it’s just built into things

00:07:14right there’s our rule of thumb here

00:07:16where we have to work with then is how

00:07:17do you how do you define that it’s agile

00:07:18if there is a significant requirements

00:07:21change and it could be significant right

00:07:23I suddenly have to make this a cell

00:07:24phone app can you do it that’s the

00:07:28question

00:07:29again can you do it quickly you know

00:07:31think about what’s involved with making

00:07:33this change there’s a lot of training

00:07:36that has to go on there’s a lot of

00:07:38learning that has to go on you might

00:07:40have to bring new people into the team

00:07:42in order to help you make the change

00:07:43right our goal is to get the app into

00:07:45users hands in a few weeks

00:07:46well again speaking as a consultant I

00:07:49can tell you that when I’m hired to do

00:07:50training often as long as four months

00:07:54elapses from the time that I’m first

00:07:55contacted by somebody who needs the

00:07:57training until the time that I can come

00:07:59in and actually give the training

00:08:00believe me if it takes four months to

00:08:02get training you’re not agile so we have

00:08:06a corporate process that’s there for

00:08:08some reason

00:08:08right somebody has decided that

00:08:10corporate governance mandates that we

00:08:12have controls over our spending and what

00:08:15that’s doing is it’s making it

00:08:16impossible for the organization to be

00:08:18agile friction Neil talked a moment ago

00:08:20about friction is this is what we’re

00:08:22talking about here is friction is that

00:08:24what agility is really about is

00:08:26eliminating friction within the

00:08:27organization and if you have an

00:08:30organization in which you cannot

00:08:31eliminate friction the the chances of

00:08:34you being agile are basically zero which

00:08:38brings me to my second kind of major

00:08:39point here is that the corporation’s

00:08:42believe that agile just happens inside

00:08:44engineering and that all that you do is

00:08:49bring in the trainers and teach them

00:08:50scrum and you’re done but in fact agile

00:08:53is an organizational thing

00:08:55there are no silos in an organizational

00:08:57organization the organization as a whole

00:09:00is what’s agile not the team’s not the

00:09:04people on the teams not the departments

00:09:06but the entire organization because if

00:09:09I’ve need my training I need my training

00:09:11now so if there’s a separate finance

00:09:14department that’s providing friction

00:09:15it’s getting in the way of me getting my

00:09:17training there is no way that there’s

00:09:18going to be any agility inside that

00:09:20organization the next issue is equally

00:09:27important is the organization is

00:09:29structured the way it is in order to

00:09:31support a process it has a process and

00:09:35it must support if if you change the

00:09:37process you must also change the

00:09:39organization because if the organization

00:09:42is all built around supporting a

00:09:44waterfall process for example then you

00:09:47can’t just suddenly become agile without

00:09:49changing the support because support for

00:09:51waterfall will not work if you’re trying

00:09:53to support agile and there are lots of

00:09:56things that are part of agile that

00:09:59people think of as things that cannot be

00:10:00changed which have to be changed right

00:10:03the QA department this is one of the

00:10:04reasons it’s so difficult for agile to

00:10:06move into big corporations in every

00:10:08agile process that works testing is

00:10:12integrated into the development process

00:10:13tightly integrated you’re testing every

00:10:1620 minutes on the outside I test every

00:10:18two or three minutes as I’m programming

00:10:20by the time the so called sprint the

00:10:24iteration is finished the code is fully

00:10:26tested and ready to deploy which is to

00:10:29say there’s nothing for a QA department

00:10:30to do which means that the director of

00:10:33QA is not going to be very happy about

00:10:34this which means that he’s going to be

00:10:37fighting really hard to prevent agile

00:10:39from happening inside the organization

00:10:41because he’s going to lose his job now

00:10:43if it’s really an agile organization he

00:10:45wouldn’t lose his job is what he would

00:10:46do is change his job in a flexible way

00:10:48so that he could actually do real QA

00:10:51which is to say it would be his job to

00:10:53look at everybody’s processes and make

00:10:55suggestions about whom how to improve

00:10:57them but a lot of people are unable to

00:10:58make those changes because again they do

00:11:00not have the flexibility product

00:11:02development groups they don’t exist in

00:11:04an agile organization the product that

00:11:06is developed by the customers

00:11:08who are talking to you on a day-to-day

00:11:09basis if your customer cannot walk in

00:11:13the front door of the corporation you

00:11:15cannot be agile and it’s very difficult

00:11:18for somebody who’s not an employee of

00:11:19the corporation to walk in the front

00:11:20door believe me the governance practices

00:11:24we were talking about finance related

00:11:26practices a moment ago is this basic

00:11:28idea of if I need the training I need it

00:11:30now not in six weeks

00:11:31if I need a new piece of software I need

00:11:33it now an agile team has a budget and it

00:11:36spends it however they feel like it the

00:11:38entire team decides to go off to a

00:11:39conference they do that they don’t ask

00:11:42anybody they just do it they need to buy

00:11:44a piece of software they do it right now

00:11:47they’ve got to manage their budget the

00:11:48budget is not infinite but they’re the

00:11:50ones making the choice not some separate

00:11:53governance related organization

00:11:57deadlines there is no such thing in the

00:11:59agile world all of the planning and

00:12:02agile is based on priorities all of it

00:12:05100% of it we’ve all seen the Iron

00:12:09Triangle picture right scope versus time

00:12:12versus money well in agile the

00:12:15assumption is what’s going to change

00:12:16most often is scope requirements always

00:12:19change and if scope changes in a way

00:12:23that some event that needs to happen

00:12:26isn’t going to happen I have to release

00:12:28at some tradeshow the only options you

00:12:31have are to reduce the scope a little

00:12:32bit or to increase spending which is to

00:12:34say bring on another team so if you’re

00:12:37managing in an agile world and you have

00:12:39a time deadline if you will do you’re

00:12:41looking at like a trade show what you’re

00:12:43doing is actually observing the process

00:12:46is that agile is a very transparent way

00:12:48of doing development and doing a

00:12:50projection based on the observed

00:12:52performance of the teams that are doing

00:12:54the work and if your projection is not

00:12:56right then you bring a new team on or

00:12:58you reduce scope those are your options

00:12:59but there are no deadlines there are no

00:13:01milestones there’s no estimation in the

00:13:03traditional sense it just doesn’t happen

00:13:06that’s a hard one for people to give up

00:13:08but there we go so there aren’t any

00:13:10timesheets because you don’t need them

00:13:11you’re not doing time-based stuff

00:13:13there’s no kind of organizational

00:13:15friction is that there’s a basic lean

00:13:17concept that any effort or time or

00:13:20energy being spent

00:13:22doing something that is not directly

00:13:24supporting putting valuable software

00:13:27into the hands of the users

00:13:29that’s all waste filling out an expense

00:13:33report or a timesheet is not providing

00:13:35customer value so one of the things that

00:13:40an agile organization has to do is

00:13:41reduce friction by eliminating

00:13:43everything that is in the way of putting

00:13:46value into the hands of the customers

00:13:48including all of this junk that we’ve

00:13:49just been talking about there is no

00:13:52project management and agile you trust

00:13:54people to do the work the teams manage

00:13:57themselves there’s no middle managers at

00:13:59all in fact in an agile world so all of

00:14:02this stuff has got to go and big

00:14:05corporations are not going to do it is

00:14:07they are cargo cults they believe that

00:14:10all this stuff is necessary for a

00:14:12company to function and that a company

00:14:14who doesn’t do this cannot function and

00:14:16that’s nonsense of course there’s some

00:14:18very large companies that function quite

00:14:19well in an agile way but most companies

00:14:22don’t believe that that’s the way things

00:14:23are going to work they have their

00:14:26managers and they expect them to work

00:14:28like managers and more to the point when

00:14:29some things become agile they expect

00:14:31that these same people are just going to

00:14:32change their job titles and suddenly

00:14:33everything’s going to be fine right and

00:14:35it’s not what it is right is that the

00:14:38job titles is not just a change in title

00:14:39is it’s functional it’s a basic change

00:14:41in what you do agile companies are

00:14:43organized very differently this is

00:14:44Spotify people talk a lot about scaling

00:14:47agile Spotify has about 500 programmers

00:14:49in it right now which is about which is

00:14:51big enough for me to say this model will

00:14:54work regardless of the size of the

00:14:55company I would argue that if you have a

00:14:57project that is requires more than 500

00:14:59programmers that what you need to do is

00:15:01break up that project into smaller

00:15:02projects there is no project on the

00:15:05planet that requires more than 500

00:15:06programmers I’m sorry

00:15:08in fact most most projects with 500

00:15:10programmers on them could be done quite

00:15:12well by 5 programmers so if it’s what if

00:15:17I can do this we can scale I’m not going

00:15:18to spend a lot of time looking at this

00:15:20picture but the main thing that I want

00:15:21to point out here is that we do not have

00:15:23a hierarchy we do not have a org chart

00:15:28that’s in a tree form with managers and

00:15:30controls and that kind of stuff we have

00:15:33a kind of matrix though this isn’t a

00:15:35classic

00:15:35matrix organization but there’s no boss

00:15:38the individual tribes which are groups

00:15:40of groups of groups if you will do have

00:15:43one guy a system architect who’s

00:15:45effectively a member of all of the teams

00:15:47whose job is to make sure there’s some

00:15:48architectural coherence across the

00:15:50system so the people are building but

00:15:52he’s not ordering people around either

00:15:54he’s saying well you know we’re doing

00:15:56this over here already

00:15:57so if we do it this way instead it’ll

00:16:00all fit better and this is a process for

00:16:02grownups so the people on Team B say ok

00:16:05we’ll do it that way then because

00:16:07they’re being flexible so nobody’s

00:16:10ordering anybody around this is a very

00:16:12collaborative kind of process which

00:16:17brings us to scrum is that scrum scrum

00:16:25is none of the things that I was just

00:16:27talking about it’s a fixed process it

00:16:31has two characteristics that a lot of

00:16:33people don’t know about is first of all

00:16:35if you look at all the processes in the

00:16:36world all the agile processes in the

00:16:38world scrum is one of them and people

00:16:43equate scrum to agile I don’t know how

00:16:45many times I’ve heard people say well

00:16:46agile says X when they mean scrum says X

00:16:50the second problem is that scrum is a

00:16:53relatively small thing and agile is a

00:16:56relatively big thing there’s a lot

00:16:59associated with agile in fact agile is

00:17:01not really even just agile is that lean

00:17:03techniques have a lot to do with running

00:17:05an effective company in an agile way so

00:17:09it’s even bigger than one guy here scrum

00:17:12is a very small thing scrum is defined

00:17:13by a 16 page document and frankly you

00:17:17can’t define any process in a 16 page

00:17:20document it just can’t be done so it’s

00:17:23very small more than the point

00:17:26the scrum guys imagine that scrum can be

00:17:28done in a kind of bubble where the the

00:17:32nobody outside has to care about what’s

00:17:35going on inside that bubble but we were

00:17:36just seeing people do have to care is

00:17:39that the purchasing procedures of the

00:17:42organization affects what goes on inside

00:17:43the bubble so it’s not a separate bubble

00:17:46it’s not something that can be isolated

00:17:48the scrum guys also have no respect at

00:17:51all for the existing processes of the

00:17:53organization when they come in is they

00:17:55imagine they can just bulldoze those

00:17:56processes out of the way and replace

00:17:57them with scrum but what they’re really

00:17:59replacing them with is nothing because

00:18:02the scrum is not going to work in a

00:18:04situation where you’re not changing the

00:18:05entire organization and if scrum is not

00:18:09going to work people are not going to do

00:18:10it a lot of places a lot of times when

00:18:12you see the organizations that have

00:18:14brought in the scrum guys a year later

00:18:15they’re not doing anything even close to

00:18:17scrum anymore because they’re not

00:18:19getting the support they need to make it

00:18:20work so they don’t do it in terms of the

00:18:23smallness you look at extreme

00:18:25programming and extreme programming is

00:18:27made up of a largest set of practices

00:18:31the interesting thing about this is the

00:18:33practices are all interwoven and

00:18:35interlocked in complex ways for example

00:18:40one of the key practices in extreme

00:18:43programming is that you need to be

00:18:44testing all the time you’re testing

00:18:46every two or three minutes as you’re

00:18:48writing the code another important

00:18:51practice in extreme programming is

00:18:53constant aggressive refactoring is that

00:18:56every time you look at a piece of code

00:18:58if it’s not right you fix it but the

00:19:01fact is you can’t do that if you don’t

00:19:03have your tests in place you can’t

00:19:06cherry pick these they’re all meshed

00:19:08together to make one unified process now

00:19:11you can modify this process and in fact

00:19:13the the XP teams do that all the time

00:19:15but they know what they’re doing they

00:19:17understand well I have these benefits

00:19:19and this has this interaction with those

00:19:21things and if I change this then I have

00:19:23to change this other stuff too in order

00:19:25to compensate you look at scrum though

00:19:28and what have they done first of all

00:19:30they have cherry picked two things they

00:19:32have stolen the XP planning game

00:19:35directly as scrum planning is nothing

00:19:37but extreme programming planning and

00:19:38then they have stolen a bunch of the

00:19:40little stuff stand-up meetings and that

00:19:42kind of stuff and then they’ve taken to

00:19:45other things that transmogrified them to

00:19:46the point where they’re unrecognizable

00:19:48the whole team among other things says

00:19:51you should have an actual customer in

00:19:53the room with you people think that when

00:19:57you’re doing extreme programming you

00:19:58don’t do design that’s nonsense what you

00:20:01are doing is you’re doing the design

00:20:02incremental e as the program evolves

00:20:03instead of spending three months talking

00:20:06to a user before you start coding you

00:20:08put the user in the room with you so you

00:20:10can ask them the question immediately

00:20:11before you start coding the user

00:20:14interaction is still there you’re still

00:20:16talking to your users you’re still doing

00:20:18design but you’re doing it as you’re

00:20:19working well they’ve gotten rid of the

00:20:22user and they’ve replaced it with this

00:20:23management Lackey called the the project

00:20:25manager where the product owner edit

00:20:29write the product owner is not a

00:20:30customer the product owner is usually a

00:20:32representative of the marketing

00:20:33department so all of the guarantees that

00:20:37you’ll get about billion valuable

00:20:38software they’re gone and then they put

00:20:42in the rigid time box which they call a

00:20:44sprint which is also a huge waste of

00:20:46time

00:20:46think how much waste there is in trying

00:20:49to size a story so you can do it in a

00:20:52two-week sprint and think of all of the

00:20:54sturm and drawing that happens when

00:20:57you’re one day off right we’re at the

00:21:02end of the sprint but there’s still one

00:21:03more day’s work to do and that’s such a

00:21:06hassle that often people don’t do that

00:21:08day’s work and you end up with buggy

00:21:11software and that’s just not okay so

00:21:15this is flying in the face of agility

00:21:18also this stuff just doesn’t work the

00:21:23next problem has to do with the

00:21:25certification mills which produce

00:21:29certificates which have no value at all

00:21:34you take a 16-page scrum guide you read

00:21:37through it so I’ve spent an afternoon

00:21:38curled in front of the fire with my

00:21:40scrum guide I go online I take a

00:21:43completely automated test which is

00:21:45graded by a computer comprised of 35

00:21:48multiple-choice questions

00:21:49I get a grade of 60% I don’t know about

00:21:53you but when I when I was in elementary

00:21:54school 60% was considered a failing

00:21:56grade but if I get a failing grade

00:21:58that’s good enough I get my certificate

00:22:00I am a master now the problem is is that

00:22:06those corporations think of this as some

00:22:07kind of diploma they think of the

00:22:09certificate as being meaningful and why

00:22:13because they do that what you get is

00:22:15something that is at

00:22:16destructive because if you imagine that

00:22:18the guy that has a certificate is

00:22:20competent and he’s not and things don’t

00:22:22work you don’t blame the guy you blame

00:22:25the thing that he’s supposed to be

00:22:27competent in doing you say agile doesn’t

00:22:30work you don’t say that scrum certified

00:22:33scrum guy didn’t actually know what he

00:22:35was talking about so certification is

00:22:38not only meaningless it is actively

00:22:39destructive especially when companies

00:22:42use certification as a criteria for

00:22:44hiring because it doesn’t tell them

00:22:46anything useful

00:22:47they bring in people that don’t know

00:22:49anything about agility but they’re

00:22:50certified and then they imagine that

00:22:52they’re somehow agile it just doesn’t

00:22:54work now the worst case of this is safe

00:22:58the problem with safe is that safe is

00:23:00designed to be acceptable to a

00:23:05corporation but it’s not safe that is is

00:23:10that in other words if you look at safe

00:23:13it’s just the same old same old

00:23:15at the bottom you’ve got scrum which is

00:23:17barely agile in the middle you’ve got

00:23:20the same old stuff that you always used

00:23:21to do well we haven’t accomplished

00:23:24anything real here we look at those flat

00:23:26hierarchies a moment ago you don’t have

00:23:27a flat hierarchy and safe you just have

00:23:29the same old corporate hierarchy that

00:23:30you ever had all the governance stuff is

00:23:32still in place it’s not real and it’s

00:23:35not necessary agile scale is just fine

00:23:38we saw that with Spotify why do we need

00:23:41to do this special thing and it’s

00:23:42because again the corporation is a cargo

00:23:44call they can’t imagine taking a big

00:23:46project and breaking it up into smaller

00:23:47projects so this is fake agile this

00:23:52isn’t real agile now mixed into this is

00:23:55a whole set of tool vendors who are

00:23:57supporting this dysfunction is that all

00:24:01of these tools force you into working in

00:24:04an old-style way is that what they’re

00:24:06doing is taking waterfall style

00:24:08management and trying to impose that on

00:24:10top of agile process by doing ridiculous

00:24:12things like burn down charts there are a

00:24:15complete waste of time as you spend more

00:24:16time entering data into this thing then

00:24:19you get useful information out of it and

00:24:20that time is waste it’s time not spent

00:24:24providing direct customer value what do

00:24:27you need when you do agile in the way of

00:24:29tools right

00:24:29is agile exactly and the thing is is

00:24:33that agile is basically very very simple

00:24:36what agile is first of all is that it’s

00:24:38a culture and the culture is the most

00:24:43important thing not the process and

00:24:45again I’m not saying process isn’t

00:24:47important but a specific process is no

00:24:50better than any other specific process

00:24:52provided that is working out within an

00:24:54altar of agility so for agile to work

00:24:57you have to put the proper culture in

00:24:59place which is to say that agile

00:25:00infusion in order to be functional has

00:25:02to come from the top down it is the CEOs

00:25:05job to manage the culture of the

00:25:07organization so if the CEO does not

00:25:10understand what agility is about there

00:25:12is no way that he or she is going to set

00:25:14up a culture that supports it so culture

00:25:18is central the second issue is trust

00:25:22everything in agile depends on trusting

00:25:24people to do their job they’re one of

00:25:28the REA in other words you don’t have

00:25:29managers why don’t you have managers and

00:25:31it’s because I trust people to do work I

00:25:33don’t need to sit there with a bullwhip

00:25:34and crack it over their heads in order

00:25:36to get them to work I trust them and you

00:25:40get trussed up and down the chain I

00:25:41don’t need to make them get sign offs in

00:25:44order to go to a conference because I

00:25:46trust them to assess their abilities

00:25:49understand that if they go to the

00:25:51conference they can increase those

00:25:52abilities and then let them go it’s all

00:25:55about trust the other issue is that the

00:26:00things on the left here communication

00:26:02simplicity courage all these values and

00:26:04principles inform the things on the

00:26:06right the practices is that the

00:26:09practices exist because the principles

00:26:11exist and the principles exist because

00:26:13the values exist so unless you have

00:26:17those principles and values in place the

00:26:20practices that you come up with they’re

00:26:22not going to be functional so we get

00:26:25back to the principles to the basic

00:26:26principles is that everything depends on

00:26:30ultimately communication and simplicity

00:26:33and courage and feedback and if you

00:26:36don’t have that inside the organization

00:26:38as a whole then you can’t really do

00:26:40functional work

00:26:43everyone following what I’m getting at

00:26:46here this is probably the most important

00:26:48thing that I can say is that the culture

00:26:50is more important than anything and that

00:26:52the principles are more important than

00:26:53the practices and that any practice that

00:26:56works in the framework of the principles

00:26:58is a good practice at least from an

00:27:00agile point of view the process itself

00:27:05is very simple you start off by talking

00:27:08to your customers and find out what they

00:27:10need and then you break off a small

00:27:13chunk of that and you build it and you

00:27:17give it to them that’s it everything

00:27:21else is icing and you do this on a small

00:27:25loop write that loop is a two-week loop

00:27:28and the reason it’s a two-week loop is

00:27:30because the general assumption of all

00:27:32the agile processes is that people don’t

00:27:34know what they need until they have

00:27:36something in their hands so upfront

00:27:39requirements gathering doesn’t work

00:27:41because the requirements that you gather

00:27:42are never correct so the process itself

00:27:49is very simple the tools for the process

00:27:52are very simple you need a whiteboard

00:27:54you need some post-it notes you need

00:27:57some index cards and some thumbtacks and

00:27:58that’s it if you’re working remotely

00:28:03maybe you’ll mix a couple other things

00:28:04into this you need some good

00:28:05communication software you probably need

00:28:07to take that that whiteboard and put it

00:28:09into a computer someplace but that’s

00:28:11inferior actually this agile is all

00:28:13based on a face-to-face communication

00:28:16two people are physically standing in

00:28:17front of a physical board will be able

00:28:19to do better work than two people

00:28:20working electronically the real problem

00:28:22with the electronic version is it’s

00:28:24hidden inside the computer instead of

00:28:25being sitting up on the wall where you

00:28:26can see it every day when you’re walking

00:28:28by it so when it’s hidden you tend to

00:28:31not think about it if it’s a physical

00:28:33thing it’s out there where you can think

00:28:35about it so those are all the tools you

00:28:37need and again we go back to our

00:28:39corporations and they’re not thinking in

00:28:41these terms they’re thinking in terms of

00:28:42big expensive tools because they’ve

00:28:44decided that that’s what they need in

00:28:45order to handle this a complexity that’s

00:28:47completely unnecessary and we don’t need

00:28:50any of that we need very simple things

00:28:54so we come back to the basic agile

00:28:57principles we finished the first one

00:28:59let’s look at the other ones our main

00:29:02goal here is to produce working software

00:29:03now over comprehensive documentation

00:29:05there are still still some corporations

00:29:07that have a documentation fetish but

00:29:09they’re getting to be less and less as

00:29:11time progresses general rule of thumb is

00:29:13that if nobody is ever going to read

00:29:14anything don’t bother to write it the

00:29:17vast majority of documentation is never

00:29:19read by anybody so it’s not it’s not

00:29:23valuable it’s waste it’s another example

00:29:25of waste but by documentation they don’t

00:29:29just mean software documentation they

00:29:31mean all paperwork inside the

00:29:32organization any kind of paperwork

00:29:35that’s getting in the way of producing

00:29:36the working software is not is not

00:29:38desirable you want to collaborate with

00:29:42your customers not negotiate with them

00:29:46customers don’t know what’s hard so they

00:29:49often ask for things that you can’t

00:29:51build a negotiation leads to dysfunction

00:29:54I want you to do such-and-such and I’m

00:29:58saying sure I’ll have it next week

00:30:00there’s an xkcd comic there was the –

00:30:03who was to this effect a couple weeks

00:30:04ago and then somebody comes in a few

00:30:06minutes later and says I need this other

00:30:07thing and you go sure I need 15 research

00:30:11assistants in a 7 million dollar budget

00:30:13and I’ll have it for you in eight years

00:30:14and the person the person who’s actually

00:30:17the question is often the same person

00:30:19they don’t know that one thing is art

00:30:20and another thing is not hard well if

00:30:23you could if you negotiate the next

00:30:26thing that person is going to say how

00:30:28about three years with eight research

00:30:31assistants and you go no well maybe four

00:30:33right and this is not solving the

00:30:35problem for one thing even the 15

00:30:38research assistants in five years or

00:30:40eight years as a guess so negotiating is

00:30:43not going to help any if you collaborate

00:30:45though you sit down and you go that’s a

00:30:47really hard problem but if we make this

00:30:50little change it turns into a really

00:30:52easy problem do you think that’ll work

00:30:53and then the guy who comes in with the

00:30:55idea says well yeah maybe but if we do

00:30:57this other little thing maybe it’ll work

00:30:58a little bit better and you collaborate

00:31:00so this goes way beyond contracts this

00:31:03goes goes into the way that you work on

00:31:05a day to day basis

00:31:08you need to be able to respond to

00:31:10changes at all levels I just said a

00:31:12moment ago that if the word agile

00:31:14applies to the process itself if the

00:31:15process isn’t working change it

00:31:17notion of retrospectives comes out of

00:31:19this but the most functional agile teams

00:31:21do not do retrospectives because our

00:31:24highly functional agile team will

00:31:26instantly recognize that something’s not

00:31:28working every better will huddle

00:31:29together they’ll come up with a solution

00:31:31they’ll put the solution in place and

00:31:32then they’ll move on a retrospective is

00:31:36kind of a transitional measure it’s

00:31:37something that you need before you get

00:31:38good enough to be able to make those

00:31:40changes quickly then there are the

00:31:45principles let’s look at just a few of

00:31:48them the welcoming changes of

00:31:50requirement even late in development

00:31:52right and the basic notion of harnessing

00:31:55to change for the composite for the

00:31:56customers competitive advantage is

00:31:58important here is that an agile team is

00:32:01in a service role it’s their job to

00:32:03solve other people’s problems not to do

00:32:06stuff because it’s cool and not to do

00:32:07stuff because it’s techie and it makes

00:32:09them feel good but to do stuff because

00:32:11it increases the but improves the

00:32:12bottom-line your customer

00:32:13whoever that customer should be the

00:32:17change the requirements are going to

00:32:18change they always change so setting up

00:32:20things where they can’t change is not

00:32:21going to work another flaw in scrum is

00:32:23they say once the scrum starts there

00:32:26will be no changes well what’s what’s

00:32:28that about I’m five minutes into the

00:32:31scrum and I realize that we’ve got a

00:32:32requirement wrong and we can’t make any

00:32:33changes it’s not going to work as you

00:32:37have to accept them at any point in the

00:32:39process the developers of the business

00:32:42people are working together on a daily

00:32:44basis that means they’re in the same

00:32:45room with each other that does not mean

00:32:47that they send memos back and forth that

00:32:51does not mean that the business people

00:32:52ask nicely before they do something

00:32:55arbitrary anyway it means that they’re

00:32:57on the team literally they’re sitting in

00:32:59the same room you’re talking on a day to

00:33:02day basis the individuals are more

00:33:06important than anything and that it is

00:33:09the job of the company to support the

00:33:10individuals if you have a situation

00:33:13where somebody has to argue in order to

00:33:15get the support they need to do their

00:33:16job there’s no way that you’re going to

00:33:18have any agility

00:33:23the constant pace issues have to do with

00:33:27the way again most corporations work

00:33:28deadlines milestones all of this stuff

00:33:31what this is a way to manipulate people

00:33:33into working overtime company meals

00:33:36high-quality cafeterias inside the

00:33:38building that’s all manipulation that’s

00:33:41all ways to get you to work overtime

00:33:44agile companies will not do that because

00:33:46they know that tired people can’t do

00:33:48good work you have to be able to come in

00:33:52every day rested and ready to go and if

00:33:54you can’t do that you can’t be agile the

00:33:59team’s themselves have to organize

00:34:01themselves right I did a class yesterday

00:34:04the end of the class we did an exercise

00:34:06I said we’re going to go on a break when

00:34:08I get back I want you guys broken into

00:34:10broken up into four groups of five or

00:34:11six I want you to pick a project that

00:34:13you’re going to work on and then I left

00:34:15them alone and I came back and they did

00:34:18it well the organization has to work

00:34:21that way you don’t assign people to

00:34:23teams you don’t break up teams teams

00:34:26form they do the job and then maybe they

00:34:28reform but usually not they usually stay

00:34:30together as a team and that’s best but

00:34:32it’s up to the teams to make those

00:34:33decisions not some sort of some sort of

00:34:35management layer so I’m going to finish

00:34:38this by thinking about Charles Darwin is

00:34:40that agile is by far the most effective

00:34:44way to develop software and the

00:34:46companies that do things in an agile way

00:34:48are going to have a huge evolutionary

00:34:50competitive advantage over the companies

00:34:53that don’t those big corporations that I

00:34:56was just talking about they’re the Dodos

00:34:58is that they’ll keep doing what they’re

00:35:02doing but their competitors who are

00:35:04working in an agile way are not limited

00:35:06in the same way that the corporation’s

00:35:08are limiting themselves so they’re going

00:35:11to just wither away maybe there should

00:35:13be a picture of Marx instead of a

00:35:14picture of Darwin but the point in other

00:35:17words the point that I’m making is that

00:35:18eventually evolution will help us here

00:35:19eventually the companies that can’t do

00:35:21things in an agile way will cease to

00:35:22exist but most companies don’t have that

00:35:26as a goal it’s not part of their their

00:35:28corporate mission statement that they

00:35:29will cease to exist in five or ten years

00:35:31because the competitor can move faster

00:35:32so really when it comes to Jill

00:35:35we have no choice we have to embrace

00:35:36change and just do it so that’s that’s

00:35:41what I have to say