Loiloi Marcom: kahi ʻokoʻa i ka hoʻāʻo ʻana o A / B

poepoe dimensional

No laila makemake mau mākou e ʻike pehea kamepiula (Ke kūʻai aku ʻana i nā kamaʻilio) ke hana nei, ma ke ʻano he kaʻa a no kahi pā kaua pākahi. I ka loiloi ʻana i ka marcom he mea maʻamau ka hoʻohana ʻana i ka hoʻāʻo A / B maʻalahi. ʻO kēia kahi hana kahi e noho ai nā laʻana ʻālualua i ʻelua mau hunaola no ka hoʻomaʻamaʻa kaua

Hoʻokahi o nā kolamu i loaʻa ka hoʻāʻo a ʻaʻole loaʻa i kekahi huna ʻē aʻe. A laila hoʻohālikelike ʻia ka helu pane a i ʻole ka loaʻa kālā ma waena o nā hunaola ʻelua. Inā ʻoi aku ka maikaʻi o ka cell test i ka cell control (ma waena o nā palena hoʻokolohua o ka hāpai ʻana, hilinaʻi, a me nā mea ʻē aʻe.

No ke aha e hana ai kekahi mea ʻē aʻe?

Eia naʻe, ʻaʻohe o kēia hana i ka hanauna ʻike. ʻAʻole maikaʻi ia i kekahi mea, hana ʻia i kahi hakahaka, hāʻawi ʻole i nā hopena no ka hoʻolālā a ʻaʻohe mana no nā mea hoʻonāukiuki ʻē aʻe.

ʻO ka lua, pinepine pinepine, hoʻohaumia ʻia ka hoʻokolohua ma ka liʻiliʻi o ka loaʻa ʻana o nā keena ʻē aʻe i nā keena ʻē aʻe, nā leka lepili, nā kamaʻilio ʻana, a pēlā aku. Ehia mau manawa i manaʻo ʻia he hopena ʻole nā ​​hopena o ka hoʻāʻo, a ʻaʻole hoʻi he sensical? No laila ua hōʻoia hou lākou i mau a hou. ʻAʻole lākou e aʻo i kekahi mea, koe wale nō ʻaʻole holo pono ka hoʻāʻo.

ʻO ia ke kumu e paipai ai i ka hoʻohana ʻana i ka regression maʻamau e kaohi no nā mea hoʻonāukiuki ʻē aʻe āpau. Hoʻohālikelike hoʻohālikelike hāʻawi pū i nā ʻike i ka loiloi marcom i hiki ke hana i kahi ROI. ʻAʻole kēia e hana ʻia i kahi hawewe, akā hāʻawi i nā koho ma ke ʻano he portfolio e hoʻonui ai i ka moʻohelu kālā.

He laʻana

E ʻōlelo mākou e hoʻāʻo nei i ʻelua mau leka uila, hōʻoia vs. kāohi a ua hoʻi nā hopena ʻole ʻole. A laila ua ʻike mākou ua hoʻouna hewa ʻole kā mākou ʻoihana inoa i kahi ʻāpana leka pololei i (ka hapanui) o ka pūʻulu kaohi. ʻAʻole i hoʻolālā ʻia kēia ʻāpana (e mākou) ʻaʻole i helu ʻia no ke koho wale ʻana i nā ʻāpana hoʻāʻo. ʻO ia, ua loaʻa i ka hui ʻoihana e like me ka maʻamau ka leka pololei maʻamau akā ʻo ka pūʻulu hoʻāʻo – kahi i mālama ʻia – ʻaʻole. He mea maʻamau kēia i kahi ʻoihana, kahi e hana ʻole ai kahi hui a kamaʻilio pū paha me kahi ʻoihana ʻē aʻe.

No laila ma kahi o ka hoʻāʻo ʻana kahi mea kūʻai aku kēlā me kēia lālani, ʻōwili mākou i ka ʻikepili ma ka manawa, e ʻōlelo i kēlā me kēia pule. Hoʻohui mākou, ma ka hebedoma, ka helu o nā leka uila hoʻāʻo, nā leka uila kaohi a me nā leka pololei i hoʻouna ʻia. Hoʻokomo pū mākou i nā loli binary e helu no ke kau, i kēia hihia i kēlā me kēia hapahā. Hōʻike ka TABLE 1 i kahi papa inoa hapa o nā hōʻuluʻulu me ka hōʻike leka uila e hoʻomaka ana i ka pule 10. I kēia manawa hana mākou i kahi hiʻohiʻona:

net \ _rev = f (em \ _test, em \ _cntrl, dir \ _mail, q_1, q_2, q_3, a pēlā aku)

ʻO ke kumu hoʻohālikelike regression maʻamau e like me ka mea i hoʻolālā ʻia ma luna nei e hoʻopuka i ka hopena TABLE 2. Hoʻopili i nā loli kūʻokoʻa ʻē aʻe o ka hoihoi. ʻO kahi leka hoʻomaopopo e hoʻokaʻawale ʻia ke kumukūʻai (net) ma ke ʻano he hoʻololi kūʻokoʻa. ʻO kēia no ka loaʻa kālā ka hoʻololi kūhelu a helu ʻia me (upena) kumu kūʻai * nui.

NĀ KULA 1

pule em_test em_cntrl dir_mail q_1 q_2 q_3 net_rev
9 0 0 55 1 0 0 $1,950
10 22 35 125 1 0 0 $2,545
11 23 44 155 1 0 0 $2,100
12 30 21 75 1 0 0 $2,675
13 35 23 80 1 0 0 $2,000
14 41 37 125 0 1 0 $2,900
15 22 54 200 0 1 0 $3,500
16 0 0 115 0 1 0 $4,500
17 0 0 25 0 1 0 $2,875
18 0 0 35 0 1 0 $6,500

E hoʻohui i ke kumukūʻai ma ke ʻano he kūʻokoʻa kūʻokoʻa ke kumu o ke kumukūʻai ma nā ʻaoʻao ʻelua o ka hoʻohālikelike, kahi kūpono ʻole. (ʻO kaʻu puke, ʻIkepili Kūʻaiwai: Kahi Alakaʻi Pono i ka ʻepekema ʻIke Kūʻai ʻoi, hāʻawi i nā laʻana he nui a me ka hoʻopili ʻana o kēia pilikia kālailai.) ʻO R2 i hoʻoponopono ʻia no kēia k modelkohu he 64%. (Ua hoʻokuʻu au i ka q4 e hōʻalo i ka pā dummy.) Emc = kaomi leka uila a me emt = leka uila hoʻāʻo. He mea nui nā loli āpau ma ka pae 95%.

NĀ KULA 2

q_3 q_2 q_1 dm emc EMTs ke kū
coeff -949 -1,402 -2,294 12 44 77 5,039
st mis 474.1 487.2 828.1 2.5 22.4 30.8
lakio-pālākiō -2 -2.88 -2.77 4.85 1.97 2.49

Ma nā ʻōlelo o ka hōʻike leka uila, ʻoi aku ka maikaʻi o ka leka uila hoʻokolohua e ka 77 vs 44 a ʻoi aku ka nui. Pēlā, e helu ana i nā mea ʻē aʻe, ua hana ka leka uila hōʻoia. Hiki mai kēia mau ʻike ke hoʻohaumia ʻia ka ʻikepili. ʻAʻole hiki i kahi hōʻike A / B ke hana i kēia.

Lawe ʻo TABLE 3 i nā coefficients e helu ai i ka waiwai o marcomm, kahi makana o kēlā me kēia kaʻa e pili ana i ka loaʻa kālā. ʻO ia, e helu i ka waiwai o nā leka pololei, hoʻonui ʻia ka coefficient o 12 e ka helu mean o nā leka pololei i hoʻouna ʻia o 109 e kiʻi i $ 1,305. Hoʻolilo nā mea kūʻai aku i ka nui awelika o $ 4,057. Penei $ 1,305 / $ 4,057 = 26.8%. ʻO ia hoʻi, hāʻawi ka leka uila pololei i ka 27% o ka nui o ka loaʻa kālā. Ma nā ʻōlelo o ROI, 109 mau leka uila e hoʻonui i $ 1,305. Inā he $ 45 ke kumu kūʻai o ka waihona nui ROI = ($ 1,305 - $ 55) / $ 55 = 2300%!

Ma muli o ke kūʻokoʻa ʻole o ke kumukūʻai, hoʻoholo pinepine ʻia e kanu ʻia ka hopena o ke kumukūʻai i ka manawa mau. I kēia hihia ʻo ka paʻa mau o 5039 e pili ana i ke kumukūʻai, nā loli ʻē aʻe i nalowale a me kahi hemahema maʻamau, a i ʻole ma kahi o 83% o ka loaʻa kālā.

NĀ KULA 3

q_3 q_2 q_1 dm emc EMTs ke kū
ʻO Coeff -949 -1,402 -2,294 12 44 77 5,039
keʻano 0.37 0.37 0.11 109.23 6.11 4.94 1
$4,875 - $ 352 - $ 521 - $ 262 $1,305 $269 $379 $4,057
waiwai -7.20% -10.70% -5.40% 26.80% 5.50% 7.80% 83.20%

Panina

Hāʻawi ka regression maʻamau i kahi koho ʻē aʻe e hāʻawi i nā ʻike i ke alo o ka ʻikepili ʻeleʻele, e like me ka mea maʻamau i kahi hoʻolālā hoʻokolohua ʻoihana. Hāʻawi pū ka Regression i kahi hāʻawi i ka loaʻa kālā a me kahi hihia ʻoihana no ROI. ʻO ka hoʻohaʻahaʻa maʻamau kahi ʻano ʻē aʻe e pili ana i ka waiwai o ka marcomm.

ir? t = marketingtechblog 20 & l = as2 & o = 1 & a = 0749474173

2 Comments

  1. 1

    ʻO kahi koho ʻē aʻe i kahi pilikia kūpono, Mike.
    I ke ala āu i hana ai, kuhi wau ʻaʻohe mea pili o nā mea kamaʻilio māka i nā pule i hala iho nei. Inā ʻaʻole ʻoe e loaʻa i kahi mea auto-regressive a me / a i ʻole manawa i lohi ʻia e ka manawa?

  2. 2

    Lawe i kāu mau loiloi e pili ana i ka optimization i ka puʻuwai, pehea e hoʻohana ai kekahi i kēia k modelkohu e hoʻonui i ka lilo o ke kahawai

Pehea kou manaʻo?

Ke hoʻohana nei kēia pūnaewele i ka Akismet e ho'ēmi i ka spam. E aʻo pehea e hanaʻia ai kāuʻikeʻikepili.