People & Blogs 

Jeff Patton – 5 things you’ll need to fix Agile product ownership

$23.20

Agile Software Development with Scrum (Series in Agile Software Development)

From Amazon

Sales Rank 50542 Schwaber, Ken/ Beedle, Mike

eXtreme Programming is an ideal many software shops would love…
Buy now

Language:
[Applause]all right you can hear me in the backokay the application is picking up okaythis is okay happen to be here anywayand off the screen my talks are supernot formal and in fact I'm going tospend time drawing pictures and I willshow a couple slides things like thathow many that this is this is an opencharity event and that means membershipare supposed to be people working withnonprofit organizations mostly how manypeople here work with nonprofitorganizations now hands up okay how manypeople snuck in okay good all right letme see if I can contextualize a littlebit I grab something I've already talkedabout but I want to make this relevantto nonprofits all right for people thatdon't know me I've been around a longtime I've been in suffered alone for 25years and I started at a very smallproduct company in 2000 I startedworking with actual development actuallyI didn't start working without adevelopment because the term was noteven coined until 2001 I started with aprocess called extreme programming andin 2001 that term agile was coinedcertainly about I was doing a naturalprocess extreme programming where thingslike cosmic like stories and terms likevelocity come from scrum where termslike product owner and sprints come fromand these days when we say at least it'sscrub has won the agile brand more sonow people talk about scrum or agilescrum and we're starting to see a commonAverell practice therenow I warned people that I startedearlier with this agile stuff and Iusually warn people that I hated itsince I started with it now but that'shyperbole look I don't hate it Iwouldn't still be doing it if I hate itit's there is a lot of super valuableuseful stuff but I started in agiledevelopment I started learning theprocess but my job title on my firstabsolute process was product manager andI had come from a product background andI've spent the last honestly 17 18 yearsof my career trying to fix this stufffirst as a product manager of my owncompany and then as a consultant and nowI get to talk to other people about it Idon't know why people listen becausealways been playing look goodanother moment is fix a lot of problemsin software development but we canthings are better than they used to bewe deliver more software morepredictably than we ever have before wenowthere's a lot more on collaboration butthe stators scrum the straight of agilethat's a little nutty and scary latelysuch that if you follow common advilpractice or scrum by the book okaynow it's actually even worse I think ifyou're often if you're doing scrum oragile by the book you end up withsomething worse than if you were look ifyou're a product owner how many peoplehave had to have the job of productowners sometime in their life okaythat's you know that if you're a productnumber this job absolutely sucks yeahit's terrible we used to use terms forproduct owner of like a single ringapple neck or one throat as a joke weheard those termswe built a process where it's not ourfaultas the team that guy's father ifsomething goes wrong that's now we talkabout meeting a really good productowner and that's absolutely true butagile development is set up this almostand tagging istic perspective of theproduct owner now I think we've got todo some work to fix how agile talksabout product ownership and products ingeneral and I think it's sort ofimportant that we do I took this picturea little over a year ago figure out whenI took it exactly at the Barbican Centrenot too far from here this was at aconference called mind of productsanybody heard of a Congress part of theproduct okay great you might have beenthere look at this conference this is aconference for people who are productmanagers and mostlytheir companies and the product managersome he loves people and some engineersit's a these are product leaders andit's the conference is usually around1500 people they fill up the BarbicanCentre big talks and there was oneSpontini the year this picture there wasone spontaneous applause line in themiddle of the talk and it's when thespeaker said look I hate a July it's allpoints and velocity and no one evertalks about the customer now there andthen a roomful of product people applyagile development has has a real problemin the product world and I spend a lotof my life trying to fix that and what Iwant to talk about are the are you knowfive specific things we need you need tostart doing to fix this people who arehere in the nonprofit world let me seehow much I can context you're like thisa little bit just talk about the firstone we first stopped treating productowners like customer you see if I canexplain what I mean by that I have toshift my preferred presentation motherthat's sharpie marker there is a commonway that we work with people when we arebuying a service I'm going to call thisa client vendor pattern now when we workthis way someone takes on the role ofthe clients and someone takes on therole of a vendor we do this all the timetell the story the reason in my mind isour car got hit by somebody and weneeded to get our car repair the personwho hit us agreed to pay for it and thenmy wife took it to it a body shop thatto get an Show long it would take so look if mywife is the client in this model andwhen your decline this model is your jobto know what you want and did you findsomebody who acts as a vendor in thismodel look it's there your talk to knowwhat you want to get to have to explainit and it's the vendors job well theyhave to listen and they have tounderstand and ask questions andespecially when we are performing aservice it's their job they have tounderstand it well enough that they canfigure out how that they're going to doit so that they can figure out so theycan figure out how long it's going totake because they're responsible forthat time thing we need to understandwell enough that we can agree on whatthe scope is and we figure out how longand how we can figure out the cost andwell we when this person describes whatit wants for getting getting formalabout it we call it their requirementsmy wife didn't call it that and say hereare my requirements but these guys didgive back that estimate for time now mywife didn't like the estimate she wasit's super expensive she asked why Iasked him to explain it and and then shedecide she liked this vendor so much shecould have had holder pushed back on theprice she actually went to some otherplace and got a better price and talkedwith him she ended up taking the car toa different place now in the real worldwe might push back against this estimatewe might negotiate and try and agree ona price because if we can do that I willget a better price if I can paywhat I want and I then I get more valuethat's the way it works I pay less I getmorethat's awesome if I'm on this side ofthe model I I need to protect myinterests I need to make sure that I'mgetting paid enough to do my servicethat's what I need to do so younegotiate back and forth but look oncewe agree we get back order and that'swhen things get tricky for the vendorthat's when they learn if they reallydid understand if they really did us tomake how long if they really did workhard they really did work out how to dothis thing they need to do the work theyneed to focus on being on on time theyneed to focus on the quality they needto protect their cost and theyeventually that well they deliver theygive back the word make sense now thispowder makes sense and it works well butyour client works out of you're a vendorbut it gets really stupid if you bothwork for the same company because if Istart haggling let's say for a minute wecall these people the business and wecall these people IT if the businesssays I don't like that estimate and doit for cheaper we may try and have withan agreed we eventually agreed and Ihave this false sense of paid I wantthis negotiation I'm gonna do betterbut because we both work for the samecompany just because you've got a betterprice doesn't mean you got better god itstill cost the same amount to do it itstill comes out of the same pocket weall work for the same company and nobodygot a deal here at all now we reallypush on a time estimate and these peoplehave to worktime and they click that to get it yeahone of the things that happens is thequality leaks out and in the clientbetter athlete powder if you finish iton time it reduces the quality that theclient loses but when we both work forthe same company we all lose it's justdumb and what's annoying is this patternis preserved hardened inside of thestraw it's probably call these thisperson the product owner and we callthis person this group of the team andthe product owners responsible forknowing what they want and building abacklog and explain the infant givingacceptance criteria the team isresponsible for constant time and thisfine content still contract is perfectlyhardened into scrum every two weeks wefix the time two weeks we fixed this thecosts from the team size and then wehave a planning meeting to agree on thescope and at the end of that two weekswe see if you met your time start let'sgo it cost contract it's kind of nice toshort of the two weeks because we startto really get some information on howgood we are at that that's the way thatthis works familiar with extremeprogramming is a process so does anybodyever do extreme programming process whatwould the does anybody remember what theexperiment even call this person orproduct donor do you remember what theycalled this person a customer nowusually a customer means somebody whobuys something that again when they workweird that doesn't make sense that in myfirst job in in a development in XP Iwas called the customer my businesscards and product manager and actually Iactually said in a different room thanthe team and the way it work now I havebeen building software for 10 years partof that this is not the way that I workI build software in a productorganization and my organization's soldthis product other people and we usedthe pattern I described a lot that I'mgoing to use some language from a guynamed Marty Kagan Kagan is written abook on product management calledinspired and he's famous for things werelots of things but basically he used torent a product at eBay started eBay whenhe was less than a hundred peoplehealthy Baker of the size that it wasthe lucky baby 2008 he was responsiblefor all the product management that alluser experience on any day now hedescribes the job of a product manageras a product manager it is your job isyour job to build a successful productand a successful product is acombinational it's just well there'sthree things we worry about we worryabout whether this product is valuablevaluable means there's been around a lotof things that's valuable to yourorganization it's ato understand what's valuable to yourorganization you need to understand howyour organization's business model workshow it makes money for a non-profit whatits mission is and how it achieves itsmission the other sentence vision theyneed to understand its strategy becauseyou need to carry it into yourorganization you also need to understandyour customers and users market inproduct terms customers don't they alsoneed to not confuse customers and usersusers use software customers buysoftware so the users and choosers is abetter way to think of it for especiallyfor bigger or enterprise-class productsthe person who decides to buy thesoftware it's not the same one who usesit I just came from sa P in Waldorfworking with them but look if you use saP in your business I'm pretty sure youweren't the one that decided to adopt saB in your business the one who decidednot they say P is usually on a golfcourse somewhere elseValley that won't means you need tounderstand those people because it needsto be you're investing in its value it'syour organization and it's has a valueproposition to your customers it alsoneeds to be usable because look at myvalue proposition and people buy it butthose users actually have to use thisstuff so you do usable you need tounderstand how users work where theproblems are how they work you need tofigure out where the solutions are howdo we change the way they work and buildsoftware support the way they work ifyou can figure out solutions it's notenough just to name what they argueabout the design what they look like andit's not enough to just a designer drawpictures of what they look like youusually need to prototype it and kind ofprove that people can actually use it sothey usually got to do that and thenfinally because any idiot could come upwith super cool ideas we can't afford tobuild it has to be feasible to buildgiven theand tools that I've got it's a newfeasible right you have to know how tocode you have to because everybodysitting on a pile of crap legacy codeyou have to understand your code howmuch it's gonna cost to change your codeyou have to understand things they'reless tangible in code that is alwaysunderstood by business concerns or usersconcerned things like scale and securityand performance things like that andthen finally because you're talkingabout a technology products or code codeages in dog yearsand so the five-year-old Cody's old ifyou want to build something thatactually last five or ten years you'vegot to be mindful of technology trendsor where they're going if you're justnow getting up to just now thinkingabout adopting cloud based architectureyou're sort of behind five ten years agoeverybody said mobile first and if youjust now said gosh I wonder if we shoulddo a mobile app you're behind if youhaven't even heard the aboutimplications of AI or Internet of Thingsyou're behind if you're buildingtechnology you've ever mindful of what'scoming so I didn't write down all thestuff as I was talking about I was readbecause it's a lot words I guess set sowhat kidding it is saying is you have tounderstand all our core business a lotabout users be able to do that all thatabout technology and that's what ittakes to build a successful product andwhen I say all that stuff there justdoesn't say that it's freakingimpossible for any one person to do andit is so a kagan'sapproach is to say it's a look you can'tyou're accountable for a successfulproduct so you better be a strongcollaborator you better surroundyourself with a product team that doesand that means you need people thatunderstand youthe way they were it's usually a UXperson these days or product designerand you better understand yourself withsomebody who is a lead or a senior insenior engineer that understands thosethings the product team is composed ofthese people and in fact what we lookfor is not a product owner making thedecisions when we talk about thedecision-making body we look for whatcake and we call a core product team nowif I talk about that concept actuallyshow this guy's name is Sharif Sharifworks at this company called Atlassiananybody who's Atlassian software likeyour confluence the first time I walkedinto at lassen now years ago he wasshowing me around showing we wearproduct teams sit and product he meansI've got some developers I've got a UXperson I've got a product manager andthen they all sit together at the sameplace you show this product keeps heresince you're this product team sits hereand pointing to a group of guests thatwere together and I said this is wherethe triaxistry out is it license jargon for coreproduct team I can ask you to try out onany team and they'll tell me who thelead engineer is the UX person is andthey are the decision-making bodythey're the ones that do this because Idon't want somebody doesn't understandtechnology making technology decisions Idon't want somebody who doesn'tunderstand user experience makingdecisions on what the he why should looklike or what should do we do wantsomething ultimately accountable butthey better be leaning heavily on theseotherpeople to make help them make gooddecisions I asked Sharif to send me apicture of his core product team he sentme this picture of them working togetheron stuff deadly and actually that theyhad an architect that a designer and Isaid well where are you in the picturesas well he's asked me to send you apicture and I took the picture I saidokay these are the people who workedclosest with it that's not everybody onhis team that's their core now I talkabout this core team concept becausethese people work closely together andtalk about who builds the backlog whodecides they decide together and we lookfor that product of manager to be astrong leader now I I started with thisidea this minor product company and Ilook I just drew this picture for youand it should look familiar because thefreakin logo for buying the product thisconcept of what's the important from abusiness perspective of a user'sperspective and what they're advocatingis a strong approach to collaborativeproduct ownership and what feels reallyat odds to these people of mind theproduct is this notion that there's aproduct owner that makes a decision anda team that does what they're toldthat's not the way I work and it's notthe way they like to work then the otherweird thing is this whole definition ofa product owner look product managerslead product teams product even ownproducts that's a big differencethere's the punchline so again that'swhere it's a dissonance comes from andthat's what bothers us and mefirst thing is you build product teamsyou don't separate them don't applyclient vendor anti-pattern to this Iknow it's fiddly add one more thingthese are the big things up front andI'm going to have to get shorter there'sa foundation now it's not about how fastyou deliver about how much you deliverlook in the client vendor anti-patternit is about how fast and how much youdeliver in fact when you apply thismodel this person wants stuff and theirgoal is to get as much stuff as they canor if cheapest they can it's not the wayit actually works in a product companynow let me draw this model if anybody'sever seen me given and give a talkbefore I apologize because I draw thismouth every single time and I will keepdrawing it until I don't have to makemost product design processes start atthe beginning with ideas those ideas canbe for whole new products they can addfor features or capabilities they'd liketo describe how we go about doing anddesign work to figure out exactly whatit is we decide how we specify thosethings and they might describe thingslike use cases or tools or specifyingstuff but these are the things that mostprocesses refer to as requirements thenwe're going to turn those ideas intoactual stuff that we can ship and sincewe're spending money to do this we worryabout the cops then cost in software isusually a function of that team theworking to build those things we alsoworry about the time because we're thosethings are sort of connected that'sgoing to take and we call all this stuffthe scope and we worry about agreeing onthat and everybody's heard of this timecost scope the dependency it call thisan Iron Triangle or triple constraintand the idea is it's a if we fix timeits cost that the scope is going to havethe Flex if we fix the time and scopealthough we may have to have our peopleand that's how that goes and if we fixall of them then the quality is going toleak out so there's a four thing inthere everything can't call it a tripleconstraint there's one in animaldevelopment we don't say scope we saythey'll backlog and an agile developmentlike I was joking about before we dothings time constant scope but all thein short cycles we take this long cycleit takes to actually get something wewant and break it into small timeperiods and we fix time constants gointo small and we measure how much scopewe delivered in that time period we callthat measurement of velocity this isn'tthis is not a product developmentprocess is a project process it's a wayto get more stuff faster or they havesome predictability of this stuff you'regoing to get and that's cool thatpredictability and the focus on somequality here but a product developmentprocess starts way upstream a productdevelopment process starts by payingattention to the product you have todayand paying attention to people who areusing it when he's pay attention peoplewho are using it that's what he fearthese people are ticked off and unhappyand sometimes angry or sometimes just uhfrustrated and struggling and it's thepaying attention to them that's wherethose ideas come from now the clientvendor vacuum powder the ideas come fromthe client if you develop a product thathas a lot of customers you pay attentionto a lot of people and they make toughdecisions on which people to focus onand when we we have to do that now orbefore we start and when we ship thisstuff later what we hope is that thosepeople are happy and unfortunatelylittle a lot of times especially forcommercial product is a lot of peoplesome people are less happy than somepeople just aren't if you can't pleaseeverybody now the terms are a lot ofpoint out here is it's everythingbetween there and there that is outputthat's what we make is not what mattersit's not why we did it it's not wherethe value comes from the value comesfrom what happens when things come outthat we measure what we want is whathappened after we call that outcomebecause of that and we measure output interms of time constant scope of velocityand story points or however we do thathere we measure outcome completelydifferent we measure outcome and howthese people change their behavior thesepeople need to see what we build theyneed to try what we built they need touse what we built that they need to keepusing what we build and ideally theythey tell others they say good thingsabout it because we deliver somethingthey didn't do any of that that's a failshe probably already detect if you'reusing something like scrum and you showyour product in a sprint review and youtrying to determine how value there isthere the value only comes out afterthisthat's the annoying thing about valueif there were no outcomes as they onlyhappen when things come out not itsready now let's track back track thislooks like a nifty process for makingpeople happy but companies are in thebusiness of making people happy theproblem really starts upstream back herein your organization your organizationhas a completely differentresponsibility it has a responsibilityto sustain itself now I realize that alot of nonprofit organizations heretruth is the nonprofit organizationswish they did run a little bit more likebusinesses because a nonprofitorganization has to sustain itself alsonow it might sustain itself by workingwith donors but the way it effectivelyworks with donors these are by showinghow much profit delivered but by showingthe outcomes that it got from what itdid you cannot get more donors byshowing how much software you shouldstate myself that's so sustainingyourself with the organization isactually a lot about proving outcomesyour organization should have somevision for where it wants to go itshould have some strategy for I wasgoing to get here and it should havesome metrics and those metrics are a lotabout how well the organization issustaining itself the those metrics areflat people inside the building areunhappy they're not unhappy becauseusing the software sucks they're unhappybecause the board can't sustain itselfnow if you're a consumer commercialorganization the flat numbers arerevenuemaybe costs are high if you're notprofit it's it's donor contributionsthings like that you don't have a goodcontributions that's back then hopefullywe get some guidance on where we need tofocus we move forward here you deliversomething a few people use it that'sawesome but hundreds to thousands dothis stuff happens we start to see goodthings or things change in theorganization those metrics start to goup people inside the organization arehappening again they're not happybecause this software is awesome they'rehappy because the organization couldsustain itself I'm going to call I'mgoing to give a different term thislonger term behavior and its impact andwe measure impact in things like returnfor commercial organizations it's likereturn on investment or brand awarenessor market share let me talk to customersabout this bugging them they're notlaying awake at night worrying aboutyour market share your brand awarenesstheir problems are not your problems andyour organization's has been most kindof people understanding that if you'retalking to the business and they'reconcerned about these things they needto realize the business you don't getwhat you want unless these people getwhat they want there's a nastydependency here if you thought yourturtle constraint was time constantscope it's not exactly the same kind oftriple constraint you've got to worryabout these things has anybody heard ofthis language system of outcome outputoutcome and impact before for me talkingabout it could be heard from elsewherebut isn't this was stolen from thenonprofit world has anybody heard thisand social change your nonprofit world Istole it because businesspeople suck somuch because they use words like goaland objective that could be a bowling ifthis delivered on time it can be a goalto increase customer satisfaction can beable to get return on investment and theprecision that I heard from someone outof the nonprofit world using thislanguage was super useful so I startedpushing it into business context andusing a lot more now today in commercialsoftware development all right here'sthe the problem is so that you'refollowing this model in the clientvendor model look these people areresponsible for output they'reresponsible for scope and time and costmanaging scope so that they can do thetime and cost but these people areresponsible for outcome and impact andthe tragedy here is the peopleresponsible for outcome and impact oftendon't have the skills it takes tounderstand what is usable or what isvaluable or feasible they may understandwhat's valuable with it they aren'toperating with a full with the peoplethey need to make decisions and that'swhy we don't use this pattern in aproduct company that's why if we want tofocus on outcome and impact we pullthese people together on a singleproduct team look let's not leave withthis when you ship something you need toknow if you've got value and you do thatteach people to be good[Music]now most organizations are prefilled ifyou're non-profit you're probably goodat measuring how much donorcontributions you have and how muchmoney to the bank but we build softwareall the time to helping you be able tomeasure anyone actually used it or notand if you ship a piece of software andyou cannot tell me if somebody used itdo you really care people here familiarwith agile development in this conceptof user stories curious is anybody havethe right acceptance criteria forstories but give you the quick threequestions to ask to get acceptancecriteria right the first thing to ask isif we build this how will we confirmit's really done that's good one thingthat I like adding to that is how willwe demonstrate this functionality whenwe review it because that drives us tothink a little bit deeper about how itwill show it and how people will use itand the question the last question ishow will we measure outcomes after thisthing is in use that drives us to sayhow how will I actually measure someoneused it it drives us to put analytics inour product so that we can actuallycheck if someone used it plan to measureoutcomes and look if you said this ifyou ship a feature you can't measure itthat's not so good don't mistakevelocity for value but without valuecomes from that other stuff now your jobI hit this hardthe annoying most annoying problem aboutthis whole stupid model is this ideathingthe problem with ideas versus thateverybody has them that's know anythingand because everybody has them there isalways too many this causes a lot ofback pressure on this systemit causes us to feel like we should weneed to serve or satisfy all thesepeople by building everything now here'sthe other rub is that most of theseideas suck look I told you I had a greatidea for a new tech product the new techstartupokay what's the failure rate of techstartups weird thing that happens isonce you get a successful product outthere and you start adding features orcapabilities suddenly everybody believesall of our ideas are awesome from now onhow many of you have ever bought aproduct and you don't actually useeverybody people like I use Atlassianand things like you're in confluence forexample they suffer from a problem a lotof product companies do where lots ofcustomers ask for things that theythought they wanted they actually put itin there and people saw it they tried itthey maybe use it but they just didn'tkeep using it and their product isfilled with a lot of things theircustomers thought they wanted but thendid not use and that's why that productthat's why Microsoft Word that's whyMicrosoft Excel and a lot of productsare so polluted with features we don'tuse there are studies that show that 60to 70 percent of features in productsfit into the rarely or never usedcategory the companies are super good atthis are eBay's andand Amazon's because they constantlymeasure whether people use and theyconstantly pull things out that don'tget used as super valuable look most ofour ideas suck now when I give that lastthird problem here is anybody evervisited a startup accelerator orincubator before the co-working spaceslots of startups working in the sameplace I think you should be a little bitof training and health it's a financialsupport to do this if you walk intowhere these places they've gotta besuper depressing right are they yespeople driving around moping feeling badworking length complaining about itsaying this sucks I hate this place nousually they're pretty upbeat placesperiod people are pretty fired up withthe one I told you 80% or 90% of peoplein your company would be laid up in thenext year I promise you they would notbe pretty fired up it's not because thefounders of startups are stupid notbecause they think the failure rate ofstartups is 10% of the success rate is90% if you ask any one of these startupsthey will say yes most startups failthey can look around at all the otherstartups in the room and say most ofthese companies fail most with theseideas suck but but not ours we everyidea we have would come in with a hugeamount of bias we have a future perfectimaginefor outcome in fact we have a hard timeimagining that something will suck thismakes the job of product managerssupercritical your job is not to buildmore of the crap that people ask you foryour job is to build less because youknow that most of these things suck andwe don't know which you have to buildthe process that helped you figure thatout your job isn't to build more yourjob is to build less to minimize thatmountain at the same time that youmaximize that outcome and impact now Istarted this by saying stop focusing onhow much you're building they startedfocusing on that you've got to minimizeoutput maximize and measure outcome andimpact now three four and five let's seeif we can go look you need to assumeyou're not building something worthwhileI sort of just explained why most thingsaren't gonna be worthwhile it's anybodyever written a user story before andthey've used this format the alley userI want so that I can get some sort ofbenefit that's cool the cool thing aboutthat is it forces us to not just talkabout the softwareforces it actually start talking aboutthese people who and other what theywant and well the why and the web so hastwo sides the problem that sucks and whymeans how we'll use it you don't thinkabout that user store for story formatlike your business and why it wants whatit wants and it means we're hidden deedsin things like that now the other thingis there's no humility in that there'sno look I want this and here's why nowyou just have been gonna build it that'sthe way it works but that's not the wayproduct companies actually work you knowI'm showing this picture and I'm gonnasome people see me give a talk beforeand I'm not going to explain why thesepeople have such I can say what elsethat sour expressions on their face butthe thing I want to point out thesepeople work for a US company calledCarmack's for the they are officiallythe world's largest used car dealershiptheir fortune 100 company went what 14billion in revenue in that slide I was acouple years ago they were across 16billion this year they had 170 locationsI think they're 190 locations somethinglike that these they're pretty big theyare again Earth's biggest used cardealership these people work there atCarmack's this is a conversation there'sa product manager there there's a UXperson I'm standing next to thedevelopment team and the people in thispicture are the CIO a director ofproduct the Director of Operations andwhat they're talking about is this thingon the wall some of you may have seenthese before if you've looked at leanuser experience and lean startup youmight have seen something calledhypothesis statement and it does notexplain what people wantit explains well our hypothesis now lookI'm going to use this word hypothesisbut hypothesis is a sexy word for ourbeliefs or our best it's it says thisone says we believe that by providingMauri Mauri is a name of a personaMaureen is somebody this is the financegroup of Carmack's and they believe thatproviding somebody who needs findingfinancing in a car she needs to know nothow much the car cost but what herpayment is to make a buying decision theproblem today with car buying is youlearn what the payment is after youdecide this is the car I want and thendo the financing paperwork then youfinally get that number and it's afunction of your credit to function theprice of the car the rate especially forused cars a function into the car themileage on the car and Maureen gets thisring and then she says Oh payment ismore than I can afford I need to go backand look at a different car and thenthey have to do this whole thing againit takes her a lot of time so theybelieve that they provider decisions orthese are really lender approved loansexactly what your favorite will ban onall cars she's gonna be able to reallychoose to make a better buying decisionand buy faster now this is not a trivialthing what it means is they have tochange the way lenders worklenders have to give them algorithms orgive them approvals and give them theability to calculate the car payment onevery single car on their lock that'swhat they mean by decisions it means ifshe fills out a quick application thenif she starts looking online or lookingon the lot or browsing she doesn't justsee the price of the car she sees herexact payment and she could more or lessthe quiet now and bypass that all thepaperwork has already been done all thefinance has been done and Carmack sellsenough cars you know for the same reasonthat Apple could change the wayvoicemail worked in 2007 they can changethe way lendersin 2017 this is Phil that's cool so lookI like replacing this with a weed baththis is what they back they bet that ifthey provide more read decisions thanall vehicles that she's going to feelmore empowered than to buy her car justgoing to say for time and that result isgoing to result in a higher it's goingto higher in a higher rate of conversionMorgan does not care about harm axisconversion rate they do in fact thatthey increase the conversion rate bySiegel different percentage it's goingto make hundreds it's going to make tensof millions to hundreds of millions ofdollars a year in benefit this is a betand we start to ask people well theywould that's things start to change nowlook I've got two pictures I can drawhere actually let's do this I am in drawthis picture look then we ask peopleactually what would you bet I've got ahypothesis here and let's say top tobottom access that is my confidence andmy confidence is high would be closercopy analysisthere's also it means my risk is low nowif you were to ask phrasing is a reallybet that this is the benefit we're goingto get you ask people how common are younow because people are biased and Ibelieve their suggestions are good thatwe asked kind of in a tricky way and sayok great you you buy that let's say weread it down you knowyou believes gonna do this what wouldyou bet me would you bet me lunch you'rerightwould you bet your next vacation thereyou're right you bet me your car allright would you bet me your house you'reright would you bet me your retirementsavings that you're right therecompany's about to spend the equivalentof your retirement savings one more tobuild this thing does he better bewilling to bet more than lunch on thisin organizations where there's not anyaccountability it's where no one thing'sbail no one's to blame asking thisquestion sort of helps at least makingsomebody a friendly back that's much forme to the next year as a useful sort ofdiscussion now once you get the hangthat okay this is a bet and we aren'tgoing to start to build some software westart to think about what do we do -well to test our bet now there's a axishere at the bottom that ranges from waysto test our backward to increase ourconfidence or reduce risk a range ofthings that range from cheap and fast toexpensive and time-consuming now if webet that if we're going to build somesoftware that it's going to we're goingto get this benefit I will submit you tothe most expensive way to test your betis to build scalable shippable softwarethe stuff that is potentially shippablethe stuff you want every single sprintthat's the most expensive way to proveyou're right now if we're not so sure ifwe're right if we start by understandthat we have hypothesis we say okayI'm not so sure I'm right one of thethings we do is build this other weirdbacklog I call it a learning backlog andthis learning backlog isn't built out ofstuff to build it's built out ofassumptions that scaris or risks or justflat-out holes in what we know questionsthat we need to learn and we prioritizethis backlog by what scares us the mostwhat's the thing keeping me from sayingyes I would bet my house that this isright what information could we learnthat would do this and it's put that atthe top and then we say okay great ifthis is the big thing I have thequestion big thing I'm going to learnwhat's the least I could possibly do inorder to learn that and well then wewell we call this thing a test plan andthen we well we create creative testslook if we're not sure people want thesoftware we might create a prototype orsomething to show we might do somethingthat that's some good test value butultimately a lot of times it's thebiggest worry is it can be build this ontime the biggest worry is what peopletry this and use it and keep using itwell enough of them do that that wecould get return on investment we comeup with ways to test that and all thetimes it means getting out and puttingsomething in front of people andwatching how they use it or looking atanalytics and what we get back from thatis in software it's this data and thatchanges our hypothesis and we go aroundthis again how many people have heard ofthings like Lean Startup beforejust read sorry same thing look you knowthis as a build measure and learningcycle and call us a validated learningloop and the goal of this isn't to buildfaster than cholermus is to learn fasterand the way we learn faster is by doinglast I narrowly focusing on specificthings to learn and by doing less andlook at our biggest concern is thatpeople really want our product reallyeven have the problem we're solving wecan do that by just interviewingcustomers the biggest concern is whetherpeople can DS we're building we could doan interview and we could build a simplepaper prototype we could do a higherfidelity prototype with some things weeventually need to actually buildsomething it's a working prototype causea live data prototype consumer companiesuse things like a B tests well sort oflive data things like that there's a lotof the ways we used to simulate thingsat some point time we actually may be aworking software but only kind ofworking for a small number of people wemay deliver something to the I builtsoftware for brick and mortar retailersfor years we always built something torelease it to one store iterated it andthen we would scale it out we usuallywork with chains stores no smaller than100 stores and usually up into the twoand three thousand but I did not buildsomething that would scale to threethousand stores on a shoot that would benuts it would take a long time so I canbuild something I can to ship to a smallsubset of early adopters people thatwill try this thing now here's thequestion to ask I go back to that betthing is if you are building softwareright nowwe form a design hypothesis statementit's a little what would you bet if yourthe process of building potentiallyshippable software but most I can giveyou up to is lunch put you right here inthis chart this is the stupid zonethis is stupid and expensive somecompanies actually do invest in a lot ofdesign sprints for everything with it'sobvious or notthey do lots of interviews lots ofprototypes but when people arehemorrhaging the solution is obvioussometimes they're still up here but lookI beg you my car is going to work outthis is all so stupid just wasting timeyou can see when I draw this thing thereis a safe zone right up the middle wherewhat we're doing in here this is safethis is where we scale our in includinginvestment with our confidence that'sthe way we actually do this and to racedirectly to potentially shippablesoftware if you are not confident enoughto that's at least the value of thatsoftware is just complete and expensiveso don't look don't started with youdon't assume you're right herecommunicate your best actually discussthe risk not the risk that you candeliver it on time because that's a risktoo but use small bets experiments testlearning activities to removeuncertainty from your betshey for people familiar with agiledevelopment anybody heard the term spikebefore estimators always understood thisvalidated learning processdescribe it's just that out ofdevelopment sprung from a type thanthere a pattern where the only thingwe're responsible for his time tossed inscope or feasibility risks when we startto expose value risks and usability riskit makes sense to do time box testsagainst these things too it's when wework with its client vendor actingpowder that we can consider outcome andimpact somebody else's problem when it'sour problem you can bet we're going tofigure out a way to test and theevidence that agile hasn't had to worryabout in the past is the fact that wehad to wait for lean startup and lean UXto add it back in now let's see if I canbring a couple of these thingsthe next one is all slice look good thatcan go fast with this look at stop saidspending so much time in the officesomebody said I asked and somebody gotout what my role was an XP and thatsomebody actually repeated back thensomebody hears say the customer on-sitething but I in XP I was called acustomer and there was a principle of XPthat was customer on site and we wantsomebody there with the team the wholetime now look imagine you're JaneGoodall if you know who Jane Goodall isas she studies primates and she spent alot of time in the field talking aboutshe spends a lot of time in jungleswaiting for chimpanzees to do somethingthat takes a long time into the wet it'sawful and look if you remember to sellit's a flip Jane we have I've had thisagile process called chimpanzee on siteit will say use about a copy of yourprimate study all you have to do isbring the chimpanzee to you if you havea primate question just turn in anabsolute the chimp just really thatdoesn't work when I started in the 1990suntil software for brick-and-mortarretailers I have lots of boring photosof hanging around carpet and homedecorating stores watching as peoplework to understand how they built thingsthey don't really client of ours that wedid work for wasn't gonna call Kinkos Igot my FedEx is now FedEx office I haveold pictures you can tell they're oldbecause look that's what monitors usedto look like I didn't know what that isI we set with them and watch them workand do this this used to be my boss backthen is now the CIO a local universityand he said gave me this weirdbackhanded comment introducing me to hisexecutives he said look this is Jeff andthe years ago I work with him he saidlook you did almost taking your team outand actually talking to customers thingthey said look right we know what you'redoing what does stop you but after weafter you started doing it it seemed tomake such a big difference we justdecided not to say anything that's whatI didn't know any better I didn't knowwhat the bill I just knew I wasaccountable for the success of myproduct I could not ask any one of mycustomers not the right thing to buildwas we very quickly that they did notcare about each other and theydefinitely did not care about mycompany's ROI and they did my bestinterest or they're each other's bestinterests at heart but the end of theday I knew we had to help them but don'twait for permission go to where they aretalk with themthe stories here sit with them whilethey work let them explain to you wherethe tough parts are or how things arework so them I guess where they arelook these the guy on the right is aproduct manager for a line of portableprinters the guy in the left is fromChina he designs portable printers theguy in the middle take we'll take apicture of you and your family and forthe Gateway Monument in Mumbai for 30rupees these guys spend time watchingthese people work that printer does notlook very portable to me by the waythese guys make friends there's someamount of empathy that we build whiledoing this and I got to tell you if youactually go out and sit with customersand spend time it may lower yourvelocity but I'm going to tell you thatteams that care and are motivated aboutwhat they're building outperform teamswith good acceptance criteria 5 to 1everyone needs to spend face time withusers and keep users in mind I want tofor questions we've talked about myfifth point here is just to stop hidingwork from the team if you are proud ofthese organics or the five waste productownership gets fixed if yourorganization understands that we havewell we have a learning work thatlearning work it's this well as patternI drew for you here where we start withhypotheses and figure out what we needto learn and we test it's all right it'sall experiments this is the work we wecall discovery and the way we dodiscover easilystart with ideas and we pull out and wego through one of those build measurelearn cycles where we get informationsometimes they are first once orsomething vlog to get our bearings andthen sometimes they're short but wemeasure these things in hours or daysnot sprinkling not weeks and sometimesthe result of those is a trash or killour ideas that if we kill an idea youknow there's always another one and pullit back out and we go through thesecycles and do this and when we finallyhave something we are confident to buildthen we invest in software we canactually ship then we build a differentkind of backlog that we move to a sortof different way of working that focuseson well development that focuses onthose thing that scrum and agilepatterns as we follow them at best thatthey're great at predictivepredictability at quality and that'sawesome and we go from working in thingsthat take hours and days to things thatwould go into these shorter strongthings but the problem if you've got aproduct there's always lots of ideas andkind of passed these forward thesepeople don't wait for these people tofinish this we build things in smallchunks we work at a feature at a timeand while we were doing this work wellthe team was building the last featureand once we finish this work we've gotto pick up the next thing and startbuilding that or start doing discoverywork on thatwe have two types of work and when thismodel gets drawn that often getsreferred to as dual track developmentand the mistake that people make is tothink that it is - it's called dualtrack because there are two kinds ofwork in the same team it is two tracksit is not two teams now when we're doingthis type of work we do have to get outthere and we do have to get out thereand spend time with customers and itisn't just UX people or people thatexperience that we interview this guyshall read if I showed you before fromAtlassian and a rule of thumb there thatlast year is that we don't do customerinterviews without having a developer inthe room so that's the way it works herethese people are doing an interview andoddly the one actually doing theinterview they're normally we bringother team members along to take notesand pay attention but the guy actuallyin the middle of talking to a customeron the left is a QA person turned outafter doing this for a while he reallygot good at it really liked doing it butat minimum I got saying also the reasonthis picture here is so blurry it'sbecause I'm sitting across the room inStarbucks with the lead engineer who wasdefinitely afraid to sit with customersit doesn't relieve him theresponsibility of going and at leastlistening in and trying to take notesand the goal is to build some apathy andwhen we come back we socialize whatwe've learned we build sharedunderstanding about anything simplethings like simple protocol journey Mapsand these quick ad-hoc personas arethings that the people that last you canbuild you recognize that guyby now how many people were here at thedesign sprint thing that was here Iheard somebody that was the last yeahhow many were here for that look whenyou do that kind of work that's quickand collaborative and it means weeverybody pitches in to figure out howthis product looks and works that meansnot just UI designers design everybodydesigns everybody gets time to sketchand comes back together and shares theirideas it is up to a designer tointegrate ideas and make tough decisionsabout what really are good ideas andwhat really aren'tbut everybody contributes that's how thedesign works and when we come up when itcomes time to build prototypes peoplehelp with that especially when it comesto start building live data prototypesthings like that look discovery work isteam work the organizations that I haveto do this start every sprint planningmeeting by talking about what discoverywork we have to do and how much of thatwe want a time box in this sprint andhow whole teams will participate at thesprint review as we talk about what welearned in discovery work everyday dailystand-ups we talk about the workdiscovery work we're doing and thedevelopment work we're doing that'slet's here's let's end up here this isthese are five big things and for a lotof organizations these are five horriblyimpossible things the product teachercross-functional team they're not aproduct owner that works with a bunch ofdevelopers and those aren't separategroups they work together they areaccountable for that outcome and impactthey know that their job is to buildless so they build us by making smallbets doing the learning activities to dothat they actually get out and spendtime with customers themselves and theybuild empathy andplan on doing all this stuff togetherand so it works it looks different todaywas that's where this thing is now[Applause]actually we're teaching a class todayArif is in the class that we've got andI can ask you I just described a wholebunch of things they're true to routeproduct about how a product ownership orproduct that's what really works in aproduct company and a lot of the waysthat's when if we're getting better atthis we find our language an awful lotbut if in a weird way I'm trying to getpeople back to what I was doing in 1998and it is what companies like a lastyearright now today but let me ask you on ascale of one to five how does what youdo every day match any stuff I justtalked about where five is that's theway we do things I learned nothing heretoday three is what we do some of thosethings but a lot of things we don't doand one is you're talking smack that'sscience addiction there's no way we cando any of those things in the worldwhere I work on a scale of one to fivehow much of these kind of things you'redoing one two threegive me I'm curious whether read is okayI see twos all the way up to the boardsmeans anybody that answers four five islie look up the dunning-kruger effectleaders but there are some people thereprobably are awesome that answered forand the roomful of people for the classwe're teaching what's interesting for meis that diversity we have some peoplewho answer one to that question and somepeople answer high numbers to thatquestion there is a real disparitybetween the haves and have-nots andbetween the Spotify Atlassian zandgovernment fix price fix fill contractsome of the baby things about seequestions questions so we spend I thinkwe're we're back the technologies we wastrying to do a lot of this stuff but wealso have to spend a lot of time swagingup for people like us work in this wayso in especially in terms of likestarting with the user needs some typesof people find that technique and makingtime for learning and therefore not likeconsistently delivering do you have likewhat one two things could do what can wetell people our policies we're buildingthis there are lots of variations of mypolicy statements we're building thisand this is we've got this is theoutcome we get start measuring evensubjectively one thing I'll ask peopleto do is one thing I've done in roomseven full of stakeholders is to havethem name the last five features or fivethings are several things that we'vedelivered and I will ask them to ratethem on a scale up awesome we really gotthe value out of this to all not onlydid we not get the value that actuallyhurt us to ship it and they were veryquickly see that gosh a lot of what weshowed is low and almost guarantee thatif you reap things are awesome to Applethey'll see it so that's a way to exposeoutcome and impact right away and thenin design thinking terms that'ssometimes we call that co-creation andit's one thing to have them try to thinkabout a friend of mine that used to workwith Marriott and they were responsiblemarrieds or franchisees and they builtstronger for franchisees we had anexecutive that was working there thatdidn't wasn't giving a lot of time andair for this stuff - you are theexecutive to lunch and then he kidnappedthe executive took the Marion franchiseand let him back in the room and had himtalk the matter of the Marriottfranchise I didn't listen to this guycomplain for a while and do things ifyou can let them not just do not wastetime to start and data to themdo not waste time shorten stupid stufflike personas there's so many smartactual empathy as vacation photos aren'tactual vacation take them on vacationdon't show them cool pictures of exoticplaces or exotic customers actually ifthey could meet somebody or be in aninterview that's good and sometimes youhave to order straight I gave if youdon't sit there but one of the othermaybe the big first don't is a justdon't talk about processyou know that's for us I find that mostbusiness executives honest to god docare about outcome impact if you talkabout these things that matters to themhow we are in our getting them andsometimes when executives don't trustApple team members is because they starttalking about story points and velocityand where the starting hominid you askwhat are the one or two things I justgive you several impossible things to dosorry I think your possible it's thatit's a culture changeif you keep the house small they wantthey oftentimes so useful if you'redoing in a post like fashionthis is the futurewhy yes them then we found out all thetests or goods and then two years agowould easily we did what you suggestdid you leave it with them and let themuse it in their jobs when they startactually using from their actualfinancial calculations remember when Isaid see try that use part exact so youdidn't do that the trick is so I can seewhy they failed they might have liked itand predicted that they would use itjust like I just like a lot of us makepredicting we will use that exercisemachine you bought in January but theproblem is we have our time predictingour own behavior now the trick is tofigure out actually how we can test it Iworked with coming to do electronicmedical records and our HIPAAregulations in the US things like thatwe were doing building software fornewborn intensive care unit in thehospital there's a little bit of concernabout whether this software works or notthings like that but the doctors we lookfor people who are early adopters earlyadopters like being involved in thisprocess and I have a problem and I'mmotivated solvents and they agreed todouble chart meaning they use theirexisting medical records product at theend of every day to spend an extra hourcharting patients in our system whichlessens didn't have all thecertifications and things like that sothat we can compare the results so thatthey can actually useevery day and we could iterate it andhow they felt until they said this feelsa lot better than what we have here andyou've got to be creative when we arelooking for things that don't they lookif we're looking for things that don'tfall into the stupid zone but also helpus build our confidence sometimes it'shard it's a trick to find thingssomewhere here in the middle with itactually allows us to learn we doactually end up having to build somesoftware but actually allow us to seewhether people use it that with theconsumer software it is so much easierwith banking software things like thatit's tougher the end of the day yeaheven if you just expose your wrist evenif you're able to say we've given peopleprototypes they said they like it butpeople are notoriously bad at predictingwhat they will actually use and companywe have some huge risk here just beaware everybody takes on risk justexpose it don't be surprised but maybe alonger conversation is to try to figureout how we can actually get somebody toask to God highlighted yeah sometimesyou have to pilot alongside of othersystems or sometimes you have to get ohit's your government in Finlandin Finland that's experimenting with aguaranteed guaranteed income maybe morewhere they took a group of people thatwere making unemployment they gave thesepeople over guaranteed income but ifyou're in unemployment and approveyou're looking for a job and you couldnot actually accept a job you couldn'tdidn't make sense for you to accept thejob that paid you less than unemploymentso you turn down jobs that were lowerthey give you a guaranteed income that'sbut look they won't do this guaranteedincome thing but they had to get specialpermission to because the government isobligated to provide everybody the sameservice I've actually changed the rulesif there is it seems super unsafe tohave a regulatory system that has no wayof testing what we're doing that theregulations that are proposed keep ussafe ensure one more question yeahmention one when you talked aboutoutcomes and you mentioned and measuringthose and that focus and you mentionedit was a non-profit person that could bebecause a lot of what they need to do isapply for grants and things and thoseare focused on specifying their outcomestheir project I'll start with that so sothere's a difficult one in a non-profitin social enterprise world you know youmeasure things by social impact yeahit's not just about revenue growth andthings like that so it's it can bedifficult to measure it's still aboutoutcomes but if you're starting smallyour sample sizes are small and that canbe dangerous and extrapolating to isjust going to give us the social impactwe need do is is there a way to kind ofaccurately that reject or were they justscale slowly to two things I mentionedthe it's funny you're right that theword for impact is when we talk aboutsocial impact we can do something to tryand bring about social impact buteverybody does this is mindful that theymay do something put as an event or putsomething out there that helps and theycan measure actually whether peopleattended an event or consumed someinformation maybe change behavior in ashort term but these social theseimpactful things aren't return oninvestment their big impact and we knowthat no one single thing we do will havean impact it's a lot of things that wedo over time and we also learn if we'retrying to do social impact that nosingle thing no single thing by itselfdoes it and we also know that weactually do start havingattribution post pockets incrediblydifficult it's hard to know which theydid it if you are a large soccer couplelike sapa you sell software and we thelicensing fee is this much if I make achange that leagues one part easier touse and more people use it did that meanI made more money is sav the thefiguring out how tribution at anorganizational level is equallytime-consuming and attribution is webelieve impossible we can't predict youdo this stuff so and by the way you cansee I have a lot of text they return oninvestment which right away tell me thisbird is actually an internationalproduct because return on investment isalmost impossible thing you cannot lookat the return on investment of a featurenow I can't tell you how many peopleused it and I can make some inferenceswhether it doesn't contribute to thingslike that now yeswe usually start with small soundingsand we release to small numbers ofpeople to do this and the only math thatallows you to project is pure booboo andanyone that says they have accurate mapthat does that these are also peoplethat are accurate and have two projectsstock market increases and decreases andthings like that it's just not easy ifthis were easily outsourced at all andit does require experience and intuitionor craft and scienceI think I might be saying that because Isuck at numbers and metrics there may bepeople that are better at this butthere's other people that are reallygood at this do any positive outcome orimpact this you know commonly acceptedmodels or ways to do this

Jeff Patton was our guest speaker at Open Charity. This event took place at Cancer Research UK in London on 8 February, 2018.

$23.20

Agile Software Development with Scrum (Series in Agile Software Development)

From Amazon

Sales Rank 50542 Schwaber, Ken/ Beedle, Mike

eXtreme Programming is an ideal many software shops would love…
Buy now

Related posts

Leave a Comment