Iyo Lamoda iri kushanda sei pane algorithms inonzwisisa zvishuwo zvemutengi

Nenguva isipi, kutenga online kuchave musanganiswa wesocial media, mapuratifomu ekurudziro, uye capsule wardrobe kutakura. Oleg Khomyuk, mukuru wedhipatimendi rekutsvagisa nekusimudzira rekambani, akaudza mashandiro anoita Lamoda pane izvi

Ndiani uye sei muLamoda anoshanda papuratifomu algorithms

KuLamoda, R&D ine basa rekuita mapurojekiti mazhinji anofambiswa nedata uye kuita mari nawo. Chikwata ichi chine vanoongorora, vanogadzira, data sainzi (mainjiniya ekudzidza muchina) uye mamaneja ezvigadzirwa. Iyo muchinjika-inoshanda timu fomati yakasarudzwa nechikonzero.

Sechivanhu, mumakambani makuru, nyanzvi idzi dzinoshanda mumadhipatimendi akasiyana - analytics, IT, madhipatimendi ezvigadzirwa. Iko kumhanya kwekuita mapurojekiti akajairwa nenzira iyi kunowanzo kuderera nekuda kwekuomerwa mukuronga pamwe chete. Basa racho pacharo rakagadzirirwa sezvinotevera: kutanga, rimwe dhipatimendi rinobatanidzwa mune analytics, uye imwe - budiriro. Mumwe nemumwe wavo ane mabasa ayo uye nguva dzekupedzisira dzemhinduro yavo.

Chikwata chedu chemuchinjika chinoshandisa nzira dzinochinjika, uye zviitiko zvenyanzvi dzakasiyana dzinoitwa zvakafanana. Nekuda kweizvi, iyo Time-To-Market chiratidzo (iyo nguva kubva pakutanga kwebasa reprojekiti kusvika pakupinda mumusika. - Trends) yakaderera pane avhareji yemusika. Imwe mukana weiyo muchinjika-inoshanda fomati kunyudzwa kwenhengo dzese dzechikwata mune yebhizinesi mamiriro uye nebasa reumwe neumwe.

Project Portfolio

Iyo purojekiti portfolio yedhipatimendi redu yakasiyana, kunyangwe nezvikonzero zviri pachena inorerekera kune chigadzirwa chedhijitari. Nzvimbo dzatiri kushanda madziri:

• katalogi uye kutsvaga;
• recommender systems;
• personalization;
• optimization yemukati maitiro.

Catalog, yekutsvaga uye yekukurudzira masisitimu maturusi ekuona ekutengesa, nzira huru iyo mutengi anosarudza chigadzirwa. Chero kukwidziridzwa kwakakosha kwekushandiswa kweichi chiitiko kune zvakakosha pakuita kwebhizinesi. Semuenzaniso, kuisa pamberi pezvigadzirwa zvinofarirwa uye zvinokwezva kune vatengi mukuronga kwekatorike kunotungamira mukuwedzera kwekutengesa, sezvo zvakaoma kuti mushandisi atarise huwandu hwese, uye kutarisisa kwake kunowanzo kugumira kumazana akati wandei anoonekwa zvigadzirwa. Panguva imwecheteyo, mazano ezvigadzirwa zvakafanana pamakadhi echigadzirwa anogona kubatsira avo, nokuda kwechimwe chikonzero, vaisada chigadzirwa chiri kuonekwa, kuita sarudzo yavo.

Imwe yenyaya dzakanyanya kubudirira dzataive nadzo yaive yekuunzwa kwetsvakiridzo itsva. Musiyano waro mukuru kubva kune yakapfuura vhezheni iri mune yemitauro algorithms yekunzwisisa chikumbiro, icho vashandisi vedu vakanyatsoona. Izvi zvakava nemigumisiro inokosha pahuwandu hwekutengesa.

Mazano ese anoedzwa

Pamberi pekushanda kutsva kunowanikwa kune vashandisi veLamoda, isu tinoitisa A / B bvunzo. Iyo inovakwa zvinoenderana neyechinyakare chirongwa uye kushandisa echinyakare zvikamu.

• Chikamu chekutanga -Tinotanga kuyedza, tichiratidza mazuva ayo uye chikamu chevashandisi vanoda kugonesa izvi kana izvo kushanda.
• Chikamu chechipiri - Isu tinounganidza zviziviso zvevashandisi vanotora chikamu mukuyedza, pamwe nedata nezve maitiro avo pasaiti uye kutenga.
• Chikamu chechitatu - pfupikisa uchishandisa yakanangwa chigadzirwa uye bhizinesi metrics.

Kubva pamaonero ebhizinesi, zvirinani maalgorithms edu anonzwisisa mibvunzo yemushandisi, kusanganisira ayo anoita zvikanganiso, zvirinani zvinokanganisa hupfumi hwedu. Zvikumbiro zvine typos hazvizotungamira kune peji isina chinhu kana kutsvaga kusiri kwayo, kukanganisa kwakaitwa kunozojeka kune yedu algorithms, uye mushandisi achaona zvigadzirwa zvaaitsvaga mune zvekutsvaga. Somugumisiro, anogona kutenga uye haazosiyi nzvimbo yacho pasina chinhu.

Hunhu hweiyo modhi nyowani hunogona kuyerwa neiyo errata kururamisa mhando metrics. Semuenzaniso, unogona kushandisa zvinotevera: "muzana yezvikumbiro zvakaruramiswa" uye "muzana wezvikumbiro zvisina kunyatsogadziriswa". Asi izvi hazvitauri zvakananga nezvekubatsira kwehunyanzvi hwakadaro hwebhizinesi. Chero zvazvingava, iwe unofanirwa kutarisa kuti chinangwa chekutsvaga metrics chinoshanduka sei mumamiriro ekurwa. Kuti tiite izvi, tinomhanyisa miedzo, iyo A / B bvunzo. Mushure mezvo, tinotarisa ma metrics, semuenzaniso, mugove wezvisina chinhu zvekutsvaga uye "click-through rate" yezvimwe zvinzvimbo kubva kumusoro mukuedza uye kutonga mapoka. Kana shanduko yakakura zvakakwana, icharatidzwa mumametrics epasi rose senge avhareji cheki, mari, uye shanduko yekutenga. Izvi zvinoratidza kuti algorithm yekugadzirisa typos inoshanda. Mushandisi anotenga kunyangwe akaita typo mumubvunzo wekutsvaga.

Kutarisa kune wese mushandisi

Isu tinoziva chimwe chinhu nezve wese Lamoda mushandisi. Kunyange kana munhu akashanyira saiti yedu kana kunyorera kekutanga, tinoona chikuva chaanoshandisa. Dzimwe nguva geolocation uye traffic sosi inowanikwa kwatiri. Zvido zvemushandisi zvinosiyana pamapuratifomu nematunhu. Naizvozvo, isu tinobva tanzwisisa izvo mutsva angangove mutengi angada.

Isu tinoziva maitiro ekushanda nenhoroondo yemushandisi yakaunganidzwa kwegore kana maviri. Iye zvino tinogona kuunganidza nhoroondo nekukurumidza - chaizvo mumaminetsi mashoma. Mushure memaminitsi ekutanga echikamu chekutanga, zvave kutokwanisika kutora dzimwe mhedziso pamusoro pezvinodiwa uye zvinonaka zveumwe munhu. Semuenzaniso, kana mushandisi akasarudza shangu chena kakawanda pakutsvaga sneakers, saka ndiyo inofanirwa kupihwa. Tinoona tarisiro yekushanda kwakadaro uye tinoronga kuishandisa.

Ikozvino, kuti tivandudze sarudzo dzemunhu, isu tiri kutarisa zvakanyanya pane hunhu hwezvigadzirwa izvo vashanyi vedu vaive neimwe mhando yekudyidzana. Kubva pane iyi data, isu tinogadzira imwe "maitiro ehunhu" yemushandisi, iyo yatinoshandisa mune yedu algorithms.

Maitiro ekuchinja achitevera maitiro evatengi vepamhepo

Chikamu chakakosha chekusimudzira kwechero chigadzirwa kusimudzira kwevatengi (kuyedza zano kana prototype yechigadzirwa chemangwana pane vangangove vatengi) uye kubvunzurudza kwakadzama. Chikwata chedu chine maneja echigadzirwa anobata nekutaurirana nevatengi. Ivo vanoita kubvunzurudza kwakadzama kuti vanzwisise zvisingafadzi zvinodiwa nemushandisi uye voshandura ruzivo irworwo kuita mazano echigadzirwa.

Pakati pemaitiro atiri kuona ikozvino, zvinotevera zvinogona kusiyaniswa:

• Chikamu chekutsvaga kubva kune nharembozha chiri kuramba chichikura. Kuwanda kwemapuratifomu enhare kuri kuchinja mabatiro anoita vashandisi nesu. Semuenzaniso, traffic paLamoda nekufamba kwenguva yakawanda uye yakawanda inoyerera kubva kukhathalogi kutsvaga. Izvi zvinotsanangurwa zviri nyore: dzimwe nguva zviri nyore kuseta mubvunzo wemavara pane kushandisa navigation mukatalogi.
• Imwe tsika yatinofanira kufunga nezvayo ndeye chishuwo chevashandisi kubvunza mibvunzo mipfupi. Naizvozvo, zvinodikanwa kuvabatsira kuti vaumbe zvikumbiro zvine revo uye zvine udzame. Semuenzaniso, tinogona kuita izvi nemazano ekutsvaga.

Chii chinotevera

Nhasi, mukutenga online, kune nzira mbiri chete dzekuvhotera chigadzirwa: kutenga kana kuwedzera chigadzirwa kune anofarira. Asi mushandisi, sekutonga, haana sarudzo dzekuratidza kuti chigadzirwa hachifarirwe. Kugadzirisa dambudziko iri ndechimwe chezvakakosha mune ramangwana.

Zvakaparadzana, timu yedu iri kushanda nesimba mukuunzwa kwekombuta yekuona matekinoroji, logistics optimization algorithms uye yakasarudzika feed yekurudziro. Isu tinovavarira kuvaka ramangwana re e-commerce zvichibva pakuongororwa kwedata uye mashandisirwo ematekinoroji matsva kugadzira sevhisi iri nani kune vatengi vedu.

Nyorera zvakare kuTrends Telegraph chiteshi uye gara uchienderana neazvino maitiro uye fungidziro nezveramangwana rehunyanzvi, hupfumi, dzidzo uye hunyanzvi.