{"id":10592,"date":"2020-09-15T13:04:00","date_gmt":"2020-09-15T11:04:00","guid":{"rendered":"https:\/\/www.marian.coach\/odhady-umeni-spatne-odhadnout\/"},"modified":"2024-05-06T17:40:32","modified_gmt":"2024-05-06T15:40:32","slug":"odhady-umeni-spatne-odhadnout","status":"publish","type":"post","link":"https:\/\/www.marian.coach\/cs\/odhady-umeni-spatne-odhadnout\/","title":{"rendered":"Odhady: Um\u011bn\u00ed \u0161patn\u011b odhadnout"},"content":{"rendered":"\n
\"\"<\/figure>\n\n

Probl\u00e9m<\/h1>\n\n

\u010casto sl\u00fdch\u00e1m, jak si zainteresovan\u00e9 strany nebo PM st\u011b\u017euj\u00ed, \u017ee t\u00fdm in\u017een\u00fdr\u016f opakovan\u011b nedokon\u010duje svou pr\u00e1ci v\u010das. Je\u0161t\u011b v\u00edce by tla\u010dili na dodr\u017eov\u00e1n\u00ed term\u00edn\u016f v domn\u011bn\u00ed, \u017ee to pom\u016f\u017ee. Ve skute\u010dnosti to jen vytv\u00e1\u0159\u00ed dal\u0161\u00ed nap\u011bt\u00ed, kter\u00e9 nikomu nepom\u016f\u017ee. Pokusili by se tak\u00e9 p\u0159idat um\u011bl\u00e9 pufry, co\u017e by probl\u00e9m je\u0161t\u011b v\u00edce zkomplikovalo.<\/p>\n\n

Samotn\u00e9 odhady jsou v\u0161ak probl\u00e9mem jen z\u0159\u00eddka a neust\u00e1l\u00e1 zpo\u017ed\u011bn\u00ed jsou obvykle zp\u016fsobena jin\u00fdmi p\u0159\u00ed\u010dinami ne\u017e nep\u0159esn\u00fdmi odhady. Nenechte se proto dr\u017eet jako rukojm\u00ed odhad\u016f!<\/p>\n\n

P\u0159ijat\u00e1 opat\u0159en\u00ed<\/h1>\n\n

Sna\u017eil jsem se zd\u016fraznit, \u017ee odhady by nem\u011bly b\u00fdt ch\u00e1p\u00e1ny jako z\u00e1vazky, ale m\u011bly by b\u00fdt vykl\u00e1d\u00e1ny \u0161\u00ed\u0159eji a zahrnovat hodnotu procesu v\u00fdvoje a nem\u011bly by b\u00fdt vystavov\u00e1ny a diskutov\u00e1ny mimo v\u00fdvoj softwaru.<\/p>\n\n

Pokud se t\u00fdmu neda\u0159\u00ed dod\u00e1vat v\u010das, jsou jeho odhady pova\u017eov\u00e1ny za nep\u0159esn\u00e9 nebo nep\u0159esn\u00e9. Pouze spr\u00e1vn\u00e9 poznatky n\u00e1m mohou pomoci zjistit, v \u010dem probl\u00e9m skute\u010dn\u011b spo\u010d\u00edv\u00e1. Obvykle bych zjistil, \u017ee syst\u00e9m doru\u010dov\u00e1n\u00ed nefunguje spr\u00e1vn\u011b. B\u011bhem pl\u00e1nov\u00e1n\u00ed sprintu se toti\u017e t\u00fdm zav\u00e1zal k n\u011bkolika probl\u00e9m\u016fm, kter\u00e9 se maj\u00ed realizovat, ale objevilo se velk\u00e9 mno\u017estv\u00ed nepl\u00e1novan\u00fdch v\u011bc\u00ed – od chyb a\u017e po r\u016fzn\u00e9 typy blok\u00e1tor\u016f – kter\u00e9 nebylo mo\u017en\u00e9 p\u0159edv\u00eddat. \u0158e\u0161en\u00edm by bylo identifikovat mo\u017en\u00e9 nepl\u00e1novan\u00e9 nebo ad hoc probl\u00e9my a jasn\u011b definovat, za jak\u00fdch okolnost\u00ed by mohly p\u0159eru\u0161it prov\u00e1d\u011bn\u00ed pl\u00e1novan\u00fdch \u010dinnost\u00ed. To by zahrnovalo definov\u00e1n\u00ed v\u0161ech t\u011bchto nepl\u00e1novan\u00fdch probl\u00e9m\u016f; nap\u0159\u00edklad co by p\u0159edstavovalo blok\u00e1tor atd. Pokud t\u00fdm \u0159e\u0161\u00ed n\u011bkolik ad hoc probl\u00e9m\u016f m\u011bs\u00ed\u010dn\u011b, v\u011bci by pravd\u011bpodobn\u011b prob\u00edhaly podle pl\u00e1nu. Pokud by v\u0161ak nepl\u00e1novan\u00e9 probl\u00e9my tvo\u0159ily a\u017e 20-30 % produkce, zp\u016fsobilo by to v\u00e1\u017en\u00e9 probl\u00e9my a n\u00e1sledn\u011b zpo\u017ed\u011bn\u00ed.<\/p>\n\n

PM a techni\u010dt\u00ed vedouc\u00ed by tak\u00e9 m\u011bli m\u00edt jasn\u00fd p\u0159ehled o tom, jak\u00fd typ pr\u00e1ce t\u00fdm vykon\u00e1v\u00e1; nap\u0159\u00edklad na kolika funkc\u00edch nebo chyb\u00e1ch pracoval, kolik podpory nab\u00edzel nebo pom\u00e1hal z\u00e1kazn\u00edk\u016fm atd. Pokud m\u00e1te ucelen\u00fd p\u0159ehled, m\u016f\u017eete upravit o\u010dek\u00e1van\u00fd pom\u011br a zahrnout jej do pl\u00e1nov\u00e1n\u00ed dal\u0161\u00edho sprintu. Tento p\u0159\u00edstup by pomohl s odhady, proto\u017ee by poskytl z\u00e1ruky pro v\u00fdstupy, jak jsou zahrnuty v pl\u00e1nu. Pokud by t\u00fdm a PM jednali transparentn\u011b a pravideln\u011b informovali z\u00fa\u010dastn\u011bn\u00e9 strany o stavu funkc\u00ed, v\u00fdsledkem by bylo pochopen\u00ed a podpora. Pokud bychom kdykoli zjistili, \u017ee se mo\u017enost zpo\u017ed\u011bn\u00ed zvy\u0161uje, m\u011bli bychom o tom neprodlen\u011b informovat z\u00fa\u010dastn\u011bn\u00e9 strany. V nejhor\u0161\u00edm p\u0159\u00edpad\u011b by se informace o mo\u017en\u00fdch zpo\u017ed\u011bn\u00edch m\u011bly uchov\u00e1vat a\u017e do samotn\u00e9ho dne doru\u010den\u00ed.<\/p>\n\n

Dal\u0161\u00ed v\u011bc\u00ed, se kterou jsem se setkal, byla nedostate\u010dn\u00e1 komunikace mezi produktem a technick\u00fdm odd\u011blen\u00edm. Produkt obvykle p\u0159ed\u00e1val po\u017eadavky a in\u017een\u00fd\u0159i je plnili. Nejen\u017ee je to sp\u00ed\u0161e p\u0159\u00edstup podobn\u00fd Vodop\u00e1du, ale nav\u00edc to brzd\u00ed produktivitu in\u017een\u00fdrstv\u00ed. M\u00edsto toho by produktov\u00fd a in\u017een\u00fdrsk\u00fd t\u00fdm m\u011bly m\u00edt mo\u017enost navrhovat v\u011bci spole\u010dn\u011b, co\u017e by in\u017een\u00fdr\u016fm pomohlo pochopit slo\u017eitost a v\u0161echny okrajov\u00e9 p\u0159\u00edpady, a t\u00edm zv\u00fd\u0161it mo\u017enost, \u017ee odhady budou spr\u00e1vn\u00e9.<\/p>\n\n

Teprve po vy\u0159e\u0161en\u00ed t\u011bchto dal\u0161\u00edch ot\u00e1zek bych se zab\u00fdval p\u0159esnost\u00ed odhad\u016f. Mohli bychom vytvo\u0159it matici, kter\u00e1 odhal\u00ed variabilitu nebo diferenciaci odhad\u016f. Pokud je odchylka vysok\u00e1, t\u00fdm zjevn\u011b neprov\u00e1d\u00ed odhady spr\u00e1vn\u011b. Abychom to napravili, mohli bychom zav\u00e9st kontroln\u00ed seznam odhad\u016f, kter\u00fd by ur\u010dil, co by m\u011blo b\u00fdt sou\u010d\u00e1st\u00ed odhad\u016f – revize k\u00f3du, QA, nasazen\u00ed atd. – a m\u011bl by zahrnovat implantaci a n\u00e1vrh.<\/p>\n\n

Z\u00edskan\u00e9 zku\u0161enosti<\/h1>\n\n