{"id":2801,"date":"2026-08-20T12:43:16","date_gmt":"2026-08-20T10:43:16","guid":{"rendered":"https:\/\/mairateam.com\/?post_type=glossary&p=2801"},"modified":"2026-08-20T15:50:19","modified_gmt":"2026-08-20T13:50:19","slug":"churn","status":"publish","type":"glossary","link":"https:\/\/mairateam.com\/cs\/slovnik\/churn\/","title":{"rendered":"Churn"},"content":{"rendered":"

Churn<\/strong> je pod\u00edl z\u00e1kazn\u00edk\u016f, kte\u0159\u00ed za dan\u00e9 obdob\u00ed ode\u0161li: zru\u0161ili p\u0159edplatn\u00e9, p\u0159estali nakupovat, nechali vypr\u0161et smlouvu. Je to z\u00e1porn\u00e1 strana retence a u p\u0159edplatn\u00fdch model\u016f nejd\u016fle\u017eit\u011bj\u0161\u00ed \u010d\u00edslo v\u016fbec: ur\u010duje, jak dlouho pr\u016fm\u011brn\u00fd z\u00e1kazn\u00edk plat\u00ed, a t\u00edm celou hodnotu akvizice.<\/p>\n

V kostce<\/h2>\n\n\n\n\n\n\n
Vzorec<\/strong><\/td>\nChurn = odch\u00e1zej\u00edc\u00ed z\u00e1kazn\u00edci za obdob\u00ed \/ z\u00e1kazn\u00edci na za\u010d\u00e1tku obdob\u00ed<\/td>\n<\/tr>\n
D\u016fsledek pro LTV<\/strong><\/td>\nPr\u016fm\u011brn\u00e1 d\u00e9lka vztahu je zhruba 1 \/ churn. M\u011bs\u00ed\u010dn\u00ed churn 5 % znamen\u00e1 asi 20 m\u011bs\u00edc\u016f<\/td>\n<\/tr>\n
Dv\u011b podoby<\/strong><\/td>\nZ\u00e1kaznick\u00fd churn po\u010d\u00edt\u00e1 lidi, revenue churn pen\u00edze. Mohou se v\u00fdrazn\u011b li\u0161it<\/td>\n<\/tr>\n
Kde ho m\u011b\u0159it i bez p\u0159edplatn\u00e9ho<\/strong><\/td>\nV e-commerce jako pod\u00edl z\u00e1kazn\u00edk\u016f, kte\u0159\u00ed u\u017e nenakoupili v obvykl\u00e9m cyklu<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Jak se po\u010d\u00edt\u00e1<\/h2>\n

Z\u00e1kaznick\u00fd a tr\u017ebov\u00fd churn<\/h3>\n

Z\u00e1kladn\u00ed vzorec pracuje se dv\u011bma \u010d\u00edsly za jedno obdob\u00ed: po\u010dtem z\u00e1kazn\u00edk\u016f na za\u010d\u00e1tku a po\u010dtem t\u011bch, kte\u0159\u00ed b\u011bhem obdob\u00ed ode\u0161li. Z\u00e1kazn\u00edci z\u00edskan\u00ed v pr\u016fb\u011bhu obdob\u00ed do v\u00fdpo\u010dtu nepat\u0159\u00ed: nejsou v z\u00e1kladu, a i kdy\u017e stihnou odej\u00edt, do odchod\u016f dan\u00e9ho obdob\u00ed se nepo\u010d\u00edtaj\u00ed. Jinak nov\u00e1 akvizice churn \u0159ed\u00ed a ten vych\u00e1z\u00ed ni\u017e\u0161\u00ed, ne\u017e je.<\/p>\n

Modelov\u00fd p\u0159\u00edklad: slu\u017eba m\u00e1 na za\u010d\u00e1tku m\u011bs\u00edce 2 000 plat\u00edc\u00edch z\u00e1kazn\u00edk\u016f. B\u011bhem m\u011bs\u00edce jich 100 zru\u0161\u00ed p\u0159edplatn\u00e9 a 300 nov\u00fdch p\u0159ibude. Z\u00e1kaznick\u00fd churn = 100 \/ 2 000 = 5 %. \u017de m\u011bs\u00edc kon\u010d\u00ed s 2 200 z\u00e1kazn\u00edky (2 000 minus 100 plus 300), na churnu nic nem\u011bn\u00ed: z\u00e1kladna roste, a p\u0159esto ka\u017ed\u00fd m\u011bs\u00edc odch\u00e1z\u00ed dvacetina.<\/p>\n

Tr\u017ebov\u00fd churn (revenue churn) po\u010d\u00edt\u00e1 tot\u00e9\u017e v pen\u011bz\u00edch: opakovan\u00e9 m\u011bs\u00ed\u010dn\u00ed tr\u017eby (MRR) ztracen\u00e9 za obdob\u00ed d\u011bleno MRR na za\u010d\u00e1tku. Kdy\u017e 2 000 z\u00e1kazn\u00edk\u016f platilo dohromady 1 200 000 K\u010d m\u011bs\u00ed\u010dn\u011b a 100 odch\u00e1zej\u00edc\u00edch neslo 90 000 K\u010d, tr\u017ebov\u00fd churn = 90 000 \/ 1 200 000 = 7,5 %. Odch\u00e1zej\u00ed tedy nadpr\u016fm\u011brn\u011b velc\u00ed z\u00e1kazn\u00edci a tr\u017ebov\u00fd churn je vy\u0161\u0161\u00ed ne\u017e z\u00e1kaznick\u00fd. Hrub\u00fd tr\u017ebov\u00fd churn zahrnuje i downgrady, \u010dist\u00fd od n\u011bj ode\u010d\u00edt\u00e1 upsell u st\u00e1vaj\u00edc\u00edch z\u00e1kazn\u00edk\u016f.<\/p>\n

P\u0159epo\u010det mezi m\u011bs\u00edcem a rokem<\/h3>\n

M\u011bs\u00ed\u010dn\u00ed churn se na ro\u010dn\u00ed nen\u00e1sob\u00ed dvan\u00e1cti, proto\u017ee z\u00e1klad se ka\u017ed\u00fd m\u011bs\u00edc zmen\u0161uje. Ro\u010dn\u00ed churn = 1 minus (1 minus m\u011bs\u00ed\u010dn\u00ed churn)^12. P\u0159i 5 % m\u011bs\u00ed\u010dn\u011b: 0,95^12 = 0,54, ro\u010dn\u00ed churn je tedy 1 minus 0,54 = 46 %, ne 60 %. Opa\u010dn\u00fdm sm\u011brem: m\u011bs\u00ed\u010dn\u00ed churn = 1 minus (1 minus ro\u010dn\u00ed churn)^(1\/12).<\/p>\n\n\n\n\n\n\n\n
M\u011bs\u00ed\u010dn\u00ed churn<\/strong><\/td>\nRo\u010dn\u00ed churn<\/strong><\/td>\n<\/tr>\n
2 %<\/td>\n21,5 %<\/td>\n<\/tr>\n
3 %<\/td>\n30,6 %<\/td>\n<\/tr>\n
5 %<\/td>\n46,0 %<\/td>\n<\/tr>\n
10 %<\/td>\n71,8 %<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Kohorty a e-shop bez p\u0159edplatn\u00e9ho<\/h3>\n

Jedno souhrnn\u00e9 \u010d\u00edslo m\u00edch\u00e1 z\u00e1kazn\u00edky star\u00e9 m\u011bs\u00edc a t\u0159i roky. Proto se churn po\u010d\u00edt\u00e1 po kohort\u00e1ch: skupina z\u00e1kazn\u00edk\u016f z\u00edskan\u00e1 ve stejn\u00e9m m\u011bs\u00edci se sleduje zvl\u00e1\u0161\u0165 a churn se vykazuje pro m\u011bs\u00edc 1, 2, 3 a dal\u0161\u00ed od akvizice. Teprve srovn\u00e1n\u00ed kohort uk\u00e1\u017ee, jestli se odchody koncentruj\u00ed na za\u010d\u00e1tku vztahu a jestli se nov\u011bj\u0161\u00ed kohorty chovaj\u00ed jinak ne\u017e star\u0161\u00ed.<\/p>\n

V e-shopu bez p\u0159edplatn\u00e9ho \u017e\u00e1dn\u00e9 \u201ezru\u0161en\u00ed\u201c neexistuje, proto se odchod definuje neaktivitou: z\u00e1kazn\u00edk, kter\u00fd nenakoupil v okn\u011b odvozen\u00e9m z obvykl\u00e9ho n\u00e1kupn\u00edho cyklu, se po\u010d\u00edt\u00e1 jako ode\u0161l\u00fd. V\u00fdpo\u010det: z kohorty 1 000 z\u00e1kazn\u00edk\u016f, kte\u0159\u00ed poprv\u00e9 nakoupili v lednu, jich do 12 m\u011bs\u00edc\u016f nakoupilo znovu 380. Churn kohorty = (1 000 minus 380) \/ 1 000 = 62 %, retence 38 %. Okno mus\u00ed b\u00fdt pro v\u0161echny kohorty stejn\u00e9, jinak se \u010d\u00edsla nedaj\u00ed srovnat.<\/p>\n

Pro\u010d mal\u00fd rozd\u00edl v churnu m\u011bn\u00ed cel\u00fd byznys<\/h2>\n

Modelov\u00fd p\u0159\u00edklad: p\u0159edplatn\u00e9 za 500 K\u010d m\u011bs\u00ed\u010dn\u011b. P\u0159i m\u011bs\u00ed\u010dn\u00edm churnu 5 procent plat\u00ed pr\u016fm\u011brn\u00fd z\u00e1kazn\u00edk asi 20 m\u011bs\u00edc\u016f a p\u0159inese 10 000 K\u010d. P\u0159i churnu 3 procenta plat\u00ed asi 33 m\u011bs\u00edc\u016f a p\u0159inese 16 500 K\u010d. Dva procentn\u00ed body rozd\u00edlu zvednou hodnotu z\u00e1kazn\u00edka o dv\u011b t\u0159etiny, a t\u00edm i strop, kolik sm\u00ed st\u00e1t akvizice. \u017d\u00e1dn\u00e1 optimalizace kampan\u00ed neum\u00ed ud\u011blat pro ekonomiku tolik, co sn\u00ed\u017een\u00ed churnu o p\u00e1r bod\u016f.<\/p>\n

Z na\u0161\u00ed praxe<\/h2>\n

Churn je d\u016fvod, pro\u010d p\u0159i auditech p\u0159edplatn\u00fdch a model\u016f zalo\u017een\u00fdch na slu\u017eb\u00e1ch za\u010d\u00edn\u00e1me reten\u010dn\u00ed k\u0159ivkou, ne kampan\u011bmi: kdy\u017e z\u00e1kazn\u00edci odch\u00e1zej\u00ed po prvn\u00edm m\u011bs\u00edci, akvizice jen nahrazuje odch\u00e1zej\u00edc\u00ed z\u00e1kazn\u00edky a spr\u00e1vn\u00e1 investice je do produktu a onboardingu, ne do m\u00e9di\u00ed. Opa\u010dn\u00fd sm\u011br plat\u00ed tak\u00e9: n\u00edzk\u00fd churn je argument pro agresivn\u011bj\u0161\u00ed akvizici, proto\u017ee ka\u017ed\u00fd z\u00edskan\u00fd z\u00e1kazn\u00edk vyn\u00e1\u0161\u00ed dlouho. My sami jsme toho p\u0159\u00edkladem: 90 procent na\u0161ich klient\u016f s n\u00e1mi z\u016fst\u00e1v\u00e1 p\u011bt a v\u00edce let, a pr\u00e1v\u011b to n\u00e1m dovoluje investovat do vztahu v\u00edc ne\u017e agentury pracuj\u00edc\u00ed na jednor\u00e1zov\u00fdch zak\u00e1zk\u00e1ch.<\/p>\n

\u010cast\u00e9 chyby<\/h2>\n