Software

Agile/Lean Startup IT Portfolio Management - No Longer Oxymorphic!

Description
Centralized development resources are commonplace in many of the late adopter organizations that struggle with Agile transformations. These centralized development resources tend to be doubly damming on the organization trying to become Agile: they lead to project thinking and management (rather than product thinking and leadership), and they lead to small incremental product development optimizations (rather than broad scale new and disruptive product breakouts). This session focuses on the mechanics of a portfolio management technique intended to guide the organization into the use of Lean Startup thinking to challenge Product Managers with, paired with the right type of fiscal rigor to make CFOs happy. Examples of how a Lean Startup Business Model Canvas can be paired with lightweight cost projections and revenue forecasts will be presented, with encouragement for organizations to specialize the artifact to suit their particular needs.
Categories
Published
of 118
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Transcript
  • 1. Agile/ Lean Startup IT Porffolio Management — No Longer Oxymorphic! / "ix f'* , I' . z ( [j"("', -4) ‘ 7 , , _ V .1 . r-2 . ''H_, : r’ -—; --x: ,—-—1' ‘: '‘>>’ F I I u u u u Inspiring Organizational Agility 4 Howard Deiner Principal Agile Coach
  • 2. How Too Many Organizations The Problem Approach the Problem ‘ , , mm, r. nrt my-r. n, , .5. . :, . _ I —. '9' . g what we Should Be Doing Final Thoughts . . “'a. r' - u luimx - So Long. and Q[‘: e:; wnsS' 3-. Thanks for All lfebai: ! 3 = ' - th. Fish ‘ ‘ | | . - . 9 Depart
  • 3. The Problem what S The: Joal of PorI'l"olio Managemnnfi The R5]-3fi0Y'l5hiP of Portfolio L, . . Management to Budgeting w, x.mmg. a-mm Pei: :l‘C: . mm 7. Gzinaaianrrm um rnerkcrs r 2- as mmv tram en: : ~._m my D»: mm mm logy .3 . .- J. |;mw: nL 'nr>mq ni~"'€ '! qr'la'nM'vl: v'r7w1:_:6 2r rm-rr what Happens if we D0 a poor what Makes Portfolio Management Job of Portfolio Management? O"Y"‘°’"P’“‘3 Wm’ A5115/“‘3"“""7 PRIMITIVE ROAD N0 WARNING SIGNS
  • 4. what’S The Goal of‘ Portfolio Management’? Make sure that we are working on the best things for the business. Maximizing value for the business. Gain balance in risks, markets, categories, project types, etc. Ensure that the entire business's strategy is in alignment. Working on the right amount of things at once.
  • 5. The Relationship of Portfolio Management to Budgeting It's a given that there is never enough money to do everything that everyone thinks up. I‘ Incremental innovation is what business and IT management are concerned with. Even when business leadership calls for disruptive change, IT management is so busy with their lighting chores that they never get to implement the real changes needed. We are so used to annual budgeting cycles that our portfolios tend to pretend to have perfect 12 month vision as to needs and implementation against them. Just keeping the lights on occupies way too many IT cycles.
  • 6. what Happens if we Do a Poor Job of Portfolio Management’? I Number of emergency projects that must start perpetuates Number of new projects I d t reduces that start ea 5 0 Balancing Feedback Stabilizing Loo / Number of active projects Loop reduces leads to Ease of managing the portfolio Number of completed projects Reinforcing Feedback Loop reduces ‘Ability oI people to finish projects quickly’ Competition for people's time re iohanna Roih-nan, Mnimge Vnw P/ i]_, IEt{ Pt: ur1_{oIiu: i, 1011, The Pragmatic Eonkshel‘, page 31
  • 7. what Makes Portfolio Management Oxymorphic with Agile/ Scrum’? 1" Agile/ Scrum focuses on the tactical apsects of delivery ‘~ I and gives virtually no advice on what do do with strategic items, such as Portfolio Management. Many companies practice Portfolio Management in a R O A D vacuum and ignore the realities of both the business N 0 needs as well as the technical realities. WARNING S| G,NS
  • 8. The Problem what S The Joal of Porfioiio Managemenfi The R5]-3fi0Y'l5hiP of Portfolio . , . . Management to Budgeting mximingiauerm Pal: :I‘’, ‘:. . .i. m V. QJII'iG€ialIKK4n um 'Ier('lS r 2- as umxv : i-um en: r~. _m my D»: mm mm logy I) . .- J. I;mw: nL 'nVHnq Hi‘ We 'lq"l arnmwl 9» mm; 2! Wlfi‘ what Happens if we D0 a poor what Makes Portfolio Management Job of Portfolio Management? O"Y"‘°’"P’“‘3 Wm’ A5115/“‘3"“""7 PRIMITIVE ROAD N0 WARNING SIGNS
  • 9. How Too Many Organizations The Problem Approach the Problem , , , mime, r. nrt Vixwv"-ri, , .5. . :, . . I —. ‘ii . s what we Should Be Doing Final Thoughts . . “h. ” - u luimx - So Long. and Q[‘: e:; wnsS' 3-. Thanks for All lfebai: ! 3 = ' - th. Fish ‘ ‘ | | . - . 9 Depart
  • 10. How Too Many Organizations Approach the Problem we Know rhat We Have way More work Than Capacity software is a wicked probjem So_ our Approach Must Be So. why Do we Continue to Act Poorly‘? Adoptive. Not Predictive! 5I‘), UiItir DICDICV“ I‘ . i awn vi Lm: I'. ml at We minrlvtrs mww
  • 11. we Know That we I-lave way More work Than Capacity Plan For Silo Efficiencies 50 break oui me specialisls. because they luipw llinii ]u|7S well‘ In aummallvl manulacluiing, me Splnallsls may be mncepl des qners. veh he aesignels, mmpanenl designers. manufacture aesqneis (O’I'p0i'IEnl man| ifa(lIAring personnel. lpgislizs aria sliippirig spe(i. 'ilv5ls, lin. 'il assembly utoducllan persnnnei, nualily curlliol peiiannei, Hi’. We coula go as was the . lime and melior‘ Studlts liial mini and Gllbfelh am in. Ford. .p get maximum out 0! worker labor. or. By the way They weren‘t too Cullclmed with Lean Defect wasies. Create a Plan for Each Project. Then Roll Them All Together a ‘s Get the mini Maiiageis busy with nearing the lasks. liming dependentlts between tasks and mark em. as tllhel FS, FF, 55, pi 5: (Is there even a good example oiwhal a '5lail to linish' rlruendency mulo be in lhe real waild7) coiieu all Ihl? moi: -u plans and create pioginii. plans lipiii them. iopiliig ip. nipie |1EDE| Vdefl€IE$ and iiuslialians Several weeks is usually siilliricnl l. usli. ii. cn ipi iliis Matrix and load balanu people and male some teams‘ Put in some milssrones along vie way in. good nieasuie The less cost me overall plan implies, and the hlghel everyone s utilization is ovei the Louise uflhe pines. plan, the belle! you oiai NDW,9IinlEVElylI'Iil1g out lei aislilbuxian, saw. me moment Use Software Manufacturing Specialists ii. software dfl‘/ ElaprPIEVlL we Iypimlly use spenallsls such as Marketing Mai'lal)el5_ Product M. .nnqn. s_ Project Managers. ii/ is, Q/ ls. S‘/ SIEIVI Aizliiluls, Business / lnliilms Sotlwale Architects, neualppeis, Daiahaun Admlnlalidlnls, ux design peisniinet ux developers. an We feel ii we use specialists ‘pr Kl'l€ work rhalweii be able to gain EWCIEHEIES in: me organization through me use ot ‘me best experienced ample for the ion‘ We [hlllk ma. eHl(IeVV silos at dlvilflpmlrll mm is. elincnyp Ofgamlallnns "l! £1 >u‘Dl.5‘ii rxiu in N»VElIiI’1.. A'4l[‘l in my ». . . .i But This Doesn't work! [Div p. ..l nevi» IU1iIK . ii nnFl is in. .. inner mni. -n nii [hr pinim plan lhal ynu pinilm pui, n was out nldale Why7 The reason . s that me am style manufacturing model aoesnl woik lo. spilwaie development When we angina. a sai, an(e the niuepiinls ale worked pul up liniii, win {an stain and [using I can Manulazluilng techniques) prfldlllr as many Lenin ul ll. .~ dl‘VI(l‘ ni. ii. -pupil win. Iiiqh uuuliiy 5l. N{l-lid) However tough engirleellng good (315 . s, engln('EVing good software . s laugh. .. sorlwaie . s tougher than ms roi one really gm ieasoi.
  • 12. Plan For Silo Efficiencies So break out the specialists, because they know their jobs well! In automative manufacturing, the specialists may be concept designers, vehicle designers, component designers, manufacture ‘b desgners, component manufacturing personnel, logistics and shipping specialists, final assembly production personnel, quality . ’~' 3 5_ "‘~ ', r ~, control personnel, etc. IKX, ’ We could go as far as the old time and motion studies that Taylor ' ; and Gilbreth did for Ford, to get maximum out of worker labor. Oh. I ‘ By the way. They weren't too concerned with Lean Defect wastes. http: // www. sarahanglis5.corn/ wp-content/ uploads/2011/O2/Frank+Gilbreth+-+rnotion+study+2.jpg
  • 13. Use Software Manufacturing Specialists In software development, we typically use specialists such as Marketing Managers, Product Managers, Project Managers, BAs, QAs, System Architects, Business Architects, Software Architects, Developers, Database Administrators, UX design personnel, UX developers, etc. We feel if we use specialists for the work that we’ll be able to gain efficiencies for the organization through the use of “the best experienced people for the job”. We think that efficient silos of development make for effective organizations.
  • 14. Create a Plan for Each Project. Then Roll Them All Together Get the Project Managers busy with creating the tasks, finding dependencies between tasks and mark each as either FS, FF, SS, or SF. (Is there even a good example of what a "Start to finish" dependency could be in the real world? ) Collect all the project plans and create program plans from them, looking for more dependencies and frustrations. Several weeks is usually sufficient frustration for this. Matrix and load balance people and create some "teams". Put in some milestones along the way for good measure. The less cost the overall plan implies, and the higher everyone's utilization is over the course of the project plan, the better you did! Now, print everything out for distribution. Savor the moment.
  • 15. But This Doesn’t work! The bad news is that as soon as that toner cooled on the project plan that you printed out, it was out of date. Why? The reason is that the old style manufacturing model doesn‘t work for software development. When we engineer a car, once the blueprints are worked out up front, we can scale and (using Lean Manufacturing techniques) produce as many copies of the device as needed with high quality standards. However tough engineering good cars is, engineering good software is tougher. Software is tougher than cars for one really good reason.
  • 16. we Know That we l-lave way More work Than Capacit Plan For Silo Effiiciencies so break out me specialists, because tney linow tnerr jabs weili In automative manutacturing, the specialists may be concept des gners. ven c. e designers. component des gners. manuracture _! -' slllpplrlq specialists rrnal assemoly production personnel guo. rty control personnel. etc I . . we could go as ratas tne . time and Mollor‘ studies tnat Taylor v“. and crlorettr did tor rord. so get maximum out 0! worker laoor. on By tne wav ttiey weren‘t too concerned witn lean Defect wastes. Create a Plan for Each Project. Then Roll Them All Together cet tne Wojetl Managers ousy witn (reallV‘glh€laSk_1indIng dependencies oetween tasks and mark earn as t-rtne- rs, rt, 55, or . sr (is rnere even a good example orwnat a ’SlarI to tinisn- ctr-irenctency could be in tne real worlrfl) Cullen all irre project plans and create program plans from ttrem. lootrirg for more dependencies and truslrations several weeks is usually siitrrtleni rrustratron tor tnis Matrix and load tralance people and create some teams Put in some milestones along tne way tor good measure rite less cost trie overall plan males, and me nigner everyones utilization is over tne course oilhc project plan. tne belle’ you did Now, print everything out tor distribution. savor ine mnmlv‘: Use Software Manufacturing Specialists in sorcware development. we typically use specialists suctr as Manietrrrg Managers. Product Managers. protect Managers. B/ s. qns. system nrcnriects, Business / irtnitecis sortware / lrcriitects. Developers. Database / sdininislintor-. ux design personnel. ux developers, etc we reel it’ we use specialists ~: or trre work tnatwell be able to gain e'V‘! ClEHElES lot tne organization tritougn tne use or ‘the best experienced ample for the lab‘ we tnrirrc ttiat erricient silos of development make ror ertective orgamzallnns But This Doesn‘t work! lire oars nr-wi lV inat . i-, xnnrl mi rn. ri inner cont. -n ciri lhr mnjrrl plan trrat you printed out. it was niit or date Why7 rrie reason is that me old style marrurarturrng model ttoesnt work tor sottware development j : '- . " r g , When wt: engineer a car, once tne olueprints are worked out up V‘ _ . riont. we can scale and {wring I ran Manuractui ng totnnioueii . _. priiclure . rs irrany toyit-i. u! the device at ricedetl witti liigri riuality | t!MldH}'t However coirgrr engineering good cars is. engineering good sottwaie is tougrier. sortware is totrgner tnan tars rot one really good reason
  • 17. How Too Many Organizations Approach ‘the Problem we Know rhat We Have way More work fhan Capaclfy software is a wicked problem So_ our Approach Must Be So. why Do we Continue to Act Poorly‘? Adoptive. Not Predictive! anew plcalcr‘ I‘ “ almnc n un: I'. ml 0) we m: :<L‘lvt's mww
  • 18. Software is a lxrlhltod Problem hnp / /www mindsamiorkxom au/ problemrsolvxng—2/on—wIcked—problerns/ Wicked Problems are: - hard to precisely define - involve multiple contrbutors and benefactors - have conditions that change during the analysis and solutioning phases - are easy to misunderstand until a solution make the problem evident - problems which tend to morph into new and different wicked problems after a solution
  • 19. How Too Many Organizations Approach ‘the Problem we Know rhat We Have way More work fhan Capacliy software is a wicked problem So_ our Approach Must Be So. why Do we Continue to Act Poorly‘? Adoptive. Not Predictive! Simvtmr DICMCV“ I‘ A avhnr n um-. ml 0; we m: :<L‘lvt's mww
  • 20. So. Our Approach Must Be Adoptive. Not Predictive! Since our problems are wicked, small changes in requirements can produce large varitions in development. We lose the ability to predict duration, even if we had perfect and unchanging requirements. This means that we have to have an iterative approach to the solution, gnawing off one piece at a time, and then seeing what's next to consider in the solution. We don't want to paint the Mona Lisa from top to bottom in final oils, lest we find that her hands don't look right and require reworking the sholders. We start with a sketch and constantly ask ourselves, "how are we doing? "
  • 21. How Too Many Organizations Approach the Problem we Know rhat We Have way More work lhan oapaczty software is a wicked problem So_ our Approach Must Be So. why Do we Continue to Act Poorly‘? Adoptive. Not Predictive! Simvtmr DICMCV“ F A avhnr n oa: I'. mi 0; We m: :<L‘lvt's mww
  • 22. So. why Do we Continue to Act Poorly’? The Annual Slog And what's wrong with That’? So. Even with the Best Intentions. .. . < m7VlHl. SlAi| '0fD1l‘Va| iD VViH)q(IIiA‘ViY pinning in am w. .. m'nVFl1tA veremi . .uF wtutiarvx :1! Nu‘: as us! mnmmm JHdDh! y{<l| llJVIl’ll'W| ThrnIi>eir. d'w3 H Weluve1><lVV'iH| ){yr'lfl. n:lE 'VIIElIW( was - gm M nnqrh mi mm, Mal‘ , m:alrs: '~eH5Iiaav. ni. >:v- mmmam ma. Vrw r. M‘ Ihqrqu xwvrnrm ? w)1lWNiA<OLi"E{i>~ (mm u . .Fr. irr, .m. l-ulnrivlqmmlV. ..rvu. »vn. u.u mm. .. r-um Ilfli mm. .. m. vnl at [Ft my H anettv mu rm». pining: my snarl Paw IrhvmvtuH| ri4Fi' we : e$J| r, ‘full new mesa Plant , ,Iiuiw'imES wfla . mnm. .m at [new 3ui<L'1 LVIIXPL. in-urniNr)fllViIi)‘ibl1‘nd( NW1 brmm M it/ at we um . -mam ocsrlu: -rrvw max : vva‘r< m mum- DrtrMi'D| a*' 'lrAjrlhr' ‘il(. VV('VvVVg(VAfiYnr'| l’rql' lhwqs kflli xam mu: Fe. ..» iEv‘lluuVi; 9u"r‘V1'. M 4'15 . u um pummg . IIK m. .. mmnrg link: av! mu pr: -xm i1 -‘ilk-KVl'Vv'lNVl- M. :. ... u.F. ..F. ..u. . JHPKVM ll1.4| r. .m, .r. .n». ».a. u.- . ..r. . in! .m. Fu. ,.. . n r. .. .., ..r 'lfl ii an-1| mm 9! II I! - Mm: Am. '~,1.m berm-m . mm mm oi . mu, umale F . Fm ; amg. .n. . nouwztt mu ur, WI“! a. .,. .u. =,m M. wnnir : :x1 Mg Agih-. '<i VH1 r. Jffiri VH1rVfl'V'. r'Vf1! wrgvilm um gt. rmluiurd VVGD urn i7iJ| r4l‘/ 1. . u., .r. ... .r. ... ,i. ,., ..r. .F. .. r.. .igr. ,». .. Fr. ..r inutiuiiu WA . ,. 'lVH| .u u. or. .. arulvnz nwlmul . ,.. ... ., . . tn!
  • 23. The Annual Slog It's not unusual for portfolio management planning to take a season or two. We start with requirements, plan solutions, estimate and cost/ benefit them, and project plan them. Then the real fun starts. We have a seemingly endless series of meetings where fight out ho gets what money. Many times it takes the HSITR approach to figure out who gets what. That is, the Highest Salary in the Room makes the decisions. Finally, we rollup the project plans into the master plan. Then we setup dependencies. The rest of the year is pretty much "we've planned our work, now let's work our plan. "
  • 24. And what’s wrong with That’? The results from most of these plans are many times just a combination of cherry picked projects which (hopefully) provide high benefits and low costs. Not product development that creates the future. Once that plan is put together, it takes an emergency to change things. See tailspin into negative reinforcing patterns. And, after all that planning, all sorts of risks are still present. In particular, - Risk that estimates have become hard deadlines, and that quality may have to be sacrificed to make the dates. - Risk that final integration and release will be delayed due to latent defects.
  • 25. So. Even with the Best Intentions. .. QB II I! The net result is that while we may be trying to be Agile in delivery, the strategic planning of the organization has already made annual decisions as to who is available when to do what. Agile/ Scrum becomes a mere shadow of what it can be — mostly daily status updates in a more collaborative workspace. That's why most organizations have trouble applying Agile/ Scrum to larger "enterprise” type programs that use centralized resources. That's why it's hard for most organizations to use Agile/ Scrum and portfolio management in the same sentence without getting raised eyebrows from all around.
  • 26. So. why Do we Continue to Act Poorly’? Ihe Annual Slog And what's wrong with That’? So. Even with the Best Intentions. .. . < m7VlHI. SlAi| V0fD1l‘Va| iD VViH)q(IIiA‘ViY pinning in am w. .. m'nVFl1tA verems . .uF v. a.mm :1! Nu‘: m :45! nnmmm Jfldvfdjzilulsnfllfln Thrnli>eie.4l'mv3 H Weiuve: ><(VV'm| ;{yr'ifl. n:le 'VIIElIW( was - gm M nnqrh mi mm, Mal‘ , m:aie: -»msiram. .»; .- mmmam ma. Vrw r. M‘ W. .. xwvrnrm ? w)1lWNiA<OLi"E{i>~ (mm u . .Fr. i:. r,. m. i-ujrxijvlq-inuvlV. ..rvH. >in| H.u um“. .. r-iii<r: Ir~jr4>x1rr . . m. vnl at [Ft my H anettv mu rave vinvxtc my snarl Paw lrhvr. mwv| rim' we : e$J| r, ‘full new mesa Plant , ,-iunvnines was . mnm. .m at [new 3.. M LVIIXPL. in-urniNr)fllViIi)‘ibl1‘nd( NW1 brmm M it/ at we um rwzrduu ocsrlu: -rrvw mar : mi‘r< in mm. .- DrtrMi'D| a*' 'lrAjrlhr' ‘il(. VV('VvVVg(VAfiYnr'l’rql’ lhwqs 9?“! Iain mil: Fe. ..» irvllnurifi 9u"r‘V1'. M 4'15 . ii um pummg . IIK m. .. mmnig link: arr mu pr: -xm r1 -‘ilk-KVl'Vv'lNVl- M. :. ... u.F. ..F. ..u. . . ..Fm. ll1.4| r. .m, .r. .n». ».a. u.- . ..r. . in! .m. Fu. ,.. ‘A ii i. .. .., ..r 'lfl ii an-1| mm 9! II I! - Mm: Am. '~,1.m berm-m . mm shadow oi . mu, name F . Fm ; amg. .n. . nouwztt mu ur, WI“! a. .,. .u. =,m M. wnnir : :x1 Mg Agrii-. '<i mm m Jffiri VHirVfl'V'. r' W wrgrilm um gt. rmluiurd VVGD urn i7iJ| r4l‘/ 1. . u., .r. ... .r. ... ,i. ,., ..r. .F. .. r.. .igr. ,». .. Fr. ..r jnuliuliu WA . ,. 'lVH| .u u. or. .. arirlvnz nwlnuul . ,.. ... ., . . tn!
  • 27. How Too Many Organizations Approach the Problem we Know Phat we Have way More work man capacity software is a wicked probiem So_ our Approach Must Be So. why Do we Continue to Act Poorly‘? Adoptive. Not Predictive! Si‘), UCIiif p-eoiee F A1JViAH(’i UD: I'i mi (7) we D. u(L‘l‘Il". mww
  • 28. How Too Many Organizations The Problem Approach the Problem , , , inane, n. nit “1K1>""-Fir ; -L4 . :. . . I —A '1' A s what we Should Be Doing Final Thoughts . . 1.1‘ -
  • We Need Your Support
    Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

    Thanks to everyone for your continued support.

    No, Thanks
    SAVE OUR EARTH

    We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

    More details...

    Sign Now!

    We are very appreciated for your Prompt Action!

    x