Wɔ wɛbsaet no mfiase no, na search bar no yɛ ade a ɛyɛ fɛ, na wɔde ka wɛbsaet bi ho bere a abɛyɛ “kɛse dodo” sɛ wobɛfa so denam kliki a wobɛkyere so no. Yɛfaa no sɛ index a ɛwɔ nhoma bi akyi: nsɛmfua a wɔahyehyɛ no ankasa, sɛnea nkyerɛwde te a ɛkyerɛ nkratafa pɔtee bi. Sɛ wokyerɛw asɛmfua pɔtee a ɔkyerɛwfo no de dii dwuma no a, wubehu nea wuhia. Sɛ woanyɛ saa a, wɔde “0 Results Found” screen a ɛte sɛ dijitaal dead end hyiaa wo. Mfeɛ aduonu anum akyi no, yɛda so ara reyɛ search bars a ɛyɛ adwuma te sɛ 1990s index cards, ɛwom mpo sɛ wɔde nhama ahyɛ nnipa a wɔde di dwuma no mu titiriw de. Ɛnnɛ, sɛ obi a ɔde di dwuma no si wo sait no so na ontumi nnya nea ohia wɔ wiase nyinaa akwantu mu wɔ sikɔne kakraa bi mu a, ɔmmɔ mmɔden sɛ obesua wo taxonomy. Wɔde wɔn ani kyerɛ adaka a wɔhwehwɛ nneɛma mu no. Nanso sɛ saa adaka no di wɔn huammɔ, na ɛhwehwɛ sɛ wɔde wo brand nsɛmfua pɔtee no di dwuma, anaasɛ wɔtwe wɔn aso wɔ typo bi ho a, wɔyɛ biribi a ɛsɛ sɛ ɛma UX designer biara da anadwo. Wofi wo wɛbsaet no so, kɔ Google, na wɔkyerɛw site:yourwebsite.com [query]. Anaasɛ, nea enye koraa no, wɔkyerɛw wɔn asɛmmisa no kɛkɛ na wowie wɔ akansifo bi wɛbsaet so. Ɛkame ayɛ sɛ m’ankasa mede Google di dwuma wɔ site’s search so bere biara. Eyi ne Site-Search Paradox no. Wɔ bere a yɛwɔ data pii ne nnwinnade a eye sen bere biara mu no, yɛn mu nhwehwɛmu osuahu ahorow no taa nyɛ papa araa ma wɔn a wɔde di dwuma no pɛ sɛ wɔde wiase nyinaa nhwehwɛmu afiri a ne bo yɛ dɔla ɔpepepem di dwuma de hwehwɛ kratafa biako wɔ mpɔtam hɔ wɛbsaet bi so. Sɛ́ Information Architects ne UX designers no, ɛsɛ sɛ yebisa sɛ, dɛn nti na “Big Box” no di nkonim, na yɛbɛyɛ dɛn atumi asan agye yɛn a yɛde di dwuma no? “Syntax Tax” Ne Owu a Ɛfa Nsɛdi Pɛpɛɛpɛ Ade titiriw nti a site search di nkogu ne nea mefrɛ no Syntax Tax. Eyi ne adwene mu adesoa a yɛde to wɔn a wɔde di dwuma so bere a yɛhwehwɛ sɛ wɔsusuw nkyerɛwde ahorow a yɛde adi dwuma wɔ yɛn database no mu no ho. Nhwehwɛmu a Origin Growth yɛe wɔ Search vs Navigate so kyerɛ sɛ bɛyɛ 50% a wɔde di dwuma no kɔ search bar no so tẽẽ bere a wɔasi fam wɔ site bi so no. Sɛ nhwɛso no, sɛ obi a ɔde di dwuma kyerɛw “sofa” gu furniture site bi a wɔakyekyɛ biribiara mu wɔ “couches” ase, na site no nsan mfa hwee mma a, nea ɔde di dwuma no nsusuw sɛ, “Ah, ɛsɛ sɛ mesɔ asɛmfua a ɛne ne ho di nsɛ hwɛ.” Wosusuw sɛ, “Sait yi nni nea mepɛ.” Eyi yɛ Information Architecture (IA) huammɔdi. Yɛakyekye yɛn nhyehyɛe ahorow no sɛnea ɛbɛyɛ a ɛne nhama (nkyerɛwde ahorow a ɛtoatoa so ankasa) bɛyɛ pɛ sen nneɛma (nsusuwii ahorow a ɛwɔ nsɛmfua no akyi). Sɛ yɛhyɛ wɔn a wɔde di dwuma no ma wɔne yɛn mu nsɛmfua nhyia a, na yɛrebɔ wɔn amemene tumi tow.
Nea Enti a Google Di Nkonim: Ɛnyɛ Tumi, Ɛyɛ Nsɛm a Ɛfa Ho Ɛnyɛ den sɛ yɛbɛtow yɛn nsa agu soro na yɛaka sɛ, “Yɛrentumi ne Google’s engineering nsi akan.” Nanso Google’s nkonimdi no nyɛ raw power kɛkɛ; ɛfa nsɛm a ɛfa ho ntease ho. Bere a yɛtaa bu hwehwɛ sɛ mfiridwuma mu mfaso no, Google bu no sɛ IA asɛnnennen. Nsɛm a efi Baymard Asoɛe no da no adi sɛ e-commerce wɛbsaet ahorow no mu 41% ntumi nnye nsɛnkyerɛnne anaa nsɛmfua ntiantiaa titiriw mpo ntom, na eyi taa ma wɔn a wɔde di dwuma no gyae wɛbsaet bi bere a wɔabɔ mmɔden pɛnkoro sɛ wɔhwehwɛ nneɛma mu a entumi nyɛ yiye no. Google di nkonim efisɛ ɛde stemming ne lemmatization di dwuma — IA akwan a wohu “mmirikatu” ne “mmirikatu” no yɛ adwene koro. Nhwehwɛmu dodow no ara a wɔhwehwɛ wɔ mu no yɛ “anifuraefo” wɔ saa tebea yi ho, na wɔfa “Mmirikatu Mpaboa” ne “Mmirikatu Mpaboa” sɛ nneɛma soronko koraa. Sɛ wo site search no ntumi nni dodow kabea a ɛnyɛ den anaasɛ nkyerɛwde a ɛnteɛ a wɔtaa de di dwuma ho dwuma a, wɔ ɔkwan a etu mpɔn so no, woregye tow afi wɔn a wɔde di dwuma no mu esiane sɛ wɔyɛ nnipa nti.
The UX Of “Ebia”: Designing For Probabilistic Nea Efi Mu Ba Wɔ atetesɛm IA mu no, yɛdwene wɔ binaries mu: Kratafa bi wɔ category bi mu, anaasɛ ɛnte saa. Search result yɛ match anaasɛ ɛnte saa. Nnɛyi hwehwɛ a mprempren wɔn a wɔde di dwuma no hwɛ kwan no yɛ nea ebetumi aba. Ɛka “ahotoso dodow” ho asɛm. Sɛnea Forresters kyerɛ no, wɔn a wɔde hwehwɛ di dwuma no taa dannan mpɛn 2–3 sen wɔn a wɔnyɛ saa, sɛ hwehwɛ no yɛ adwuma a. Na 80% a wɔde di dwuma wɔ e-commerce sites so fi site bi mu esiane nhwehwɛmu a enye nti. Sɛ́ adwumfo no, yɛntaa nyɛ mfonini mma mfinimfini fam. Yɛyɛ “Results Found” kratafa ne “No Results” kratafa. Yɛpa tebea a ɛho hia sen biara no: “So Wopɛe?” Ɔman. Ɛsɛ sɛ nhwehwɛmu kwan a wɔayɛ no yiye ma “Fuzzy” nhyiam ahorow. Sɛ́ anka yɛde “0 Results Found” screen a ɛyɛ nwini bedi dwuma no, ɛsɛ sɛ yɛde yɛn metadata no di dwuma de ka sɛ, “Yɛanhu saa wɔ ‘Electronics’ mu, nanso yehuu 3 matches wɔ ‘Accessories’ mu.” Ɛdenam nhyehyɛe a yɛbɛyɛ ama “Ebia” so no, yebetumi ama nea ɔde di dwuma no akɔ so atra nsu a ɛsen no mu. Nsɛm a Wɔayɛ no Ho Nhwehwɛmu: Nsɛm a “Aniwa Nhu” Ho Ka Sɛ yɛbɛte nea enti a IA yɛ pɛtro masearch engine no, ɛsɛ sɛ yɛhwɛ sɛnea wɔahyehyɛ data wɔ akyi. Wɔ me mfeɛ 25 a mede ayɛ adwuma no mu no, mahunu sɛ “findability” a ɛwɔ krataafa bi mu no bata ne metadata a wɔahyehyɛ no ho tẽẽ. Susuw adwumakuw kɛse bi a me ne no yɛɛ adwuma a na wɔwɔ mfiridwuma ho nkrataa bɛboro 5,000 ho hwɛ. Na wɔn mu hwehwɛ no resan de aba a ɛho nhia aba efisɛ na “Title” tag a ɛwɔ krataa biara mu no yɛ emu SKU nɔma (e.g., “DOC-9928-X”) sen sɛ ɛbɛyɛ din a onipa betumi akenkan. Ɛdenam nhwehwɛmu a yɛhwɛɛ mu no so no, yehui sɛ wɔn a wɔde di dwuma no rehwehwɛ “installation guide.” Esiane sɛ saa kasasin no anpue wɔ SKU-based title no mu nti, engine no buu n’ani guu fael ahorow a ɛfa ho kɛse no so. Yɛde Controlled Vocabulary dii dwuma, a na ɛyɛ nsɛmfua a wɔahyɛ da ayɛ a ɛma SKU ahorow kɔ nnipa kasa mu. Wɔ asram abiɛsa mu no, “Exit Rate” a efi kratafa a wɔhwehwɛ no so tew 40%. Na eyi nyɛ algorithmic fix; na ɛyɛ IA fix. Ɛda no adi sɛ search engine ye te sɛ asase mfonini a yɛde ma no nkutoo. Kasa mu Nsonsonoe a Ɛwɔ Mu Wɔ mfe aduonu a mede ayɛ UX no nyinaa mu no, mahyɛ asɛmti bi a ɛsan ba nsow: emu akuw taa hu amane wɔ “Nimdeɛ nnome” ho. Yɛde yɛn ho hyɛ yɛn ankasa adwumakuw nsɛmfua mu araa ma, anaasɛ ɛtɔ mmere bi a wɔfrɛ no adwumayɛ mu kasakoa, araa ma yɛn werɛ fi sɛ nea ɔde di dwuma no nka yɛn kasa. Bere bi me ne sikasɛm adwumayɛbea bi a na n’abam abu wɔ telefon a wɔfrɛ wɔn frɛ kɛse a wɔde kɔ wɔn mmoa beae no yɛɛ adwuma. Na wɔn a wɔde di dwuma no nwiinwii sɛ wontumi nnya “bosea a wotua” ho nsɛm wɔ wɛbsaet no so. Bere a yɛhwɛɛ search logs no, “loan payoff” ne asɛmfua #1 a wɔhwehwɛe a ɛde zero hits bae. Adɛn? Efisɛ na institution’s IA kuw no akyerɛw kratafa biara a ɛfa ho wɔ asɛmfua a wɔde di dwuma wɔ ɔkwan a ɛfata so “Loan Release” ase. Wɔ sikakorabea no fam no, na “akatua” yɛ adeyɛ, nanso na “Loan Release” ne mmara kwan so krataa a na ɛyɛ “ade” a ɛwɔ database no mu. Esiane sɛ na search engine no rehwehwɛ nkyerɛwde nhama ankasa nti, ɛpowee sɛ ɛde nea ɔde di dwuma no ahiade kɛse no bɛbata adwumakuw no ano aduru a ɛyɛ aban de no ho. Eyi ne baabi a ɛsɛ sɛ IA adwumayɛfo no yɛ adwuma sɛ nsɛm asekyerɛfo. Ɛdenam “bosea a wotua” a yɛde kaa ho kɛkɛ sɛ metadata asɛmfua titiriw a ahintaw wɔ Loan Release nkratafa no so no, yedii mmoa ho haw bi a ɛyɛ dɔla ɔpepem pii ho dwuma. Ná yenhia server a ɛyɛ ntɛm sen saa; na yehia nkyekyem ahorow a ɛwɔ tema kɛse. Anamɔn 4 a ɛwɔ Site-search Audit Framework no Sɛ wopɛ sɛ wosan gye wo search box no fi Google so a, wuntumi “nhyehyɛ no na wo werɛ mfi” kɛkɛ. Ɛsɛ sɛ wubu hwehwɛ sɛ ade a ɛte ase. Nhyehyɛe a mede di dwuma de bu nhwehwɛmu na mayɛ nhwehwɛmu osuahu ahorow no yiye ni: Ɔfã 1: “Zero-result” Nhwehwɛmu no Twe wo nhwehwɛmu nsɛm a woakyerɛw afi nnafua 90 a atwam no mu. Filter ma nsɛmmisa a ɛsan de zero aba no nyinaa. Fa eyinom bom yɛ bokiti abiɛsa:
Nokware gapsContent a ɔde di dwuma no pɛ sɛ wo nni kɛkɛ (sɛnkyerɛnne ma wo content strategy team). Synonym gapsContent a wowɔ, nanso wɔakyerɛkyerɛ mu wɔ nsɛmfua a ɔde di dwuma no mfa nni dwuma mu (e.g., “Sofa” vs “Couch”). Format gapsObi a ɔde di dwuma no rehwehwɛ “video” anaa “PDF,” nanso wo hwehwɛ no de HTML nsɛm nkutoo na ɛkyerɛw.
Ɔfa 2: Asɛmmisa Adwene Mapping Hwehwɛ nsɛmmisa 50 a ɛsen biara a wɔtaa bisa no mu. So wɔyɛ Navigational (a wɔrehwehwɛ kratafa pɔtee bi), Informational (a wɔrehwehwɛ “sɛnea wɔbɛyɛ”), anaa Transactional (a wɔrehwehwɛ ade pɔtee bi)? Ɛsɛ sɛ wo hwehwɛ UI no yɛ soronko ma emu biara. Ɛsɛ sɛ ɔkwan a wɔfa so hwehwɛ nneɛma mu no “Quick-Link” de nea ɔde di dwuma no kɔ baabi a ɔrekɔ no tẽẽ, na ɛtwa nea efi mu ba no kratafa no ho koraa. Ɔfã 3: “Fuzzy” Nhyiam Sɔhwɛ Bɔ da kyerɛw wo nneɛma 10 a ɛsen biara no wɔ ɔkwan a ɛnteɛ so. Fa dodow kabea, mfomso a wɔtaa de kyerɛw, ne Amerikafo ne Britania Engiresi kasa mu nkyerɛwde (e.g., “Color” vs. “Colour”) di dwuma. Sɛ wo hwehwɛ no di nkogu wɔ sɔhwɛ ahorow yi mu a, wo engine no nni “stemming” mmoa. Eyi yɛ mfiridwuma mu ahwehwɛde a ɛsɛ sɛ wokamfo kyerɛ wo mfiridwuma kuw no. Ɔfa 4: Scoping Ne Filtering UX Hwɛ wo results krataafa no. So ɛde filter ahorow a ntease wom ankasa ma? Sɛ obi a ɔde di dwuma hwehwɛ “mpaboa” a, ɛsɛ sɛ ohu filters a ɛkyerɛ Size ne Colour. Generic filters betumi ayɛ bɔne te sɛ nea filter biara nni hɔ. Reclaiming The Search Box: Akwan a Wɔfa so Yɛ Ma IA Adwumayɛfo Sɛ yɛbɛgyae atutra a yɛrekɔ Google no a, ɛsɛ sɛ yɛkɔ akyiri sen “Adaka” no na yɛhwɛ scaffolding no. Anamɔn A: Fa semantic scaffolding di dwuma.Don’t just return a list of links. Fa wo IA no di dwuma de ma nsɛm a ɛfa ho. Sɛ obi hwehwɛ ade bi a, kyerɛ no ade no, nanso kyerɛ wɔn nhoma no, FAQs, ne afã horow a ɛfa ho nso. Saa “fekubɔ” hwehwɛ yi suasua sɛnea onipa amemene yɛ adwuma ne sɛnea Google yɛ adwuma. Anamɔn B: Gyae sɛ wobɛyɛ nhomakorabea sohwɛfo, fi asebeing a concierge.Nhomakorabea sohwɛfo ka baabi pɔtee a nhoma no wɔ wɔ shelf no so kyerɛ wo. Ɔpon ano hwɛfo tie nea wopɛ sɛ wunya na ɔkamfo wo kyerɛ. Ɛnsɛ sɛ wo search bar no de nsɛm a ɛkyerɛ nkɔm di dwuma ɛnyɛ sɛ wode wie nsɛmfua nko, na mmom de kyerɛ adwene a wowɔ. Search Bar a Google na ɛma ahoɔden a wode bedi dwuma “Google-powered” search bar a wɔde bedi dwuma, sɛnea yehu wɔ Chicago Sukuupɔn wɛbsaet no so no, yɛ gye a wogye tom titiriw sɛ wɛbsaet bi mu ahyehyɛde abɛyɛ nea ɛyɛ den dodo sɛ n’ankasa akwantu betumi adi ho dwuma. Bere a ɛyɛ “nsiesie” a wɔyɛ no ntɛm ma ahyehyɛde akɛse de hwɛ hu sɛ wɔn a wɔde di dwuma no benya biribi no, mpɛn pii no ɛnyɛ nea wɔpaw yiye ma nnwuma a emu dɔ.
Sɛ wode hwehwɛ no hyɛ Google nsa a, wode osuahu a ɔde di dwuma no bɛma abɔnten algorithm bi. Wohwere tumi a wode bɛhyɛ nneɛma pɔtee bi ho nkuran, woma wɔn a wɔde di dwuma no hu aguade ho dawurubɔ ahorow a ɛto so abiɛsa, na wotete w’atɔfo ma wofi w’abɔde a nkwa wom mu bere a wohia mmoa no. Wɔ adwuma bi fam no, ɛsɛ sɛ hwehwɛ yɛ nkɔmmɔbɔ a wɔayɛ no yiye a ɛkyerɛ adetɔfo kwan kɔ botae bi ho, na ɛnyɛ link ahorow a wɔahyehyɛ no nyinaa a ɛpia wɔn san kɔ wɛbsaet a wɔabue no so.
Nhwehwɛmu a Ɛnyɛ Den UX Nhwehwɛmu Nhwehwɛmu a etwa to a wode bɛhwɛ bere a worekyekye hwehwɛ osuahu no ama wo dwumadiefoɔ ni. Wo ne wo nneɛma kuw no nyɛ adwuma na hwɛ hu sɛ wo ne kuw no mufo a wɔfata redi nkitaho.
Kill the dead-end.Nka kɛkɛ da sɛ “No results found.” Sɛ nea ɛne no hyia pɛpɛɛpɛ nni hɔ a, kyerɛ ɔfã a ɛte saa ara, ade a agye din, anaa ɔkwan a wobɛfa so ne mmoa adi nkitaho. Fix “almost” matches.Hwɛ sɛ hwehwɛ no betumi adi dodow kabea (te sɛ “afifide” vs. “afifide”) ne mfomso a wɔtaa de kyerɛw ho dwuma. Ɛnsɛ sɛ wɔtwe wɔn a wɔde di dwuma no aso wɔ wɔn nsateaa a wɔde wɔn ho hyɛ mu no ho. Predict the user’s goal.Fa “auto-suggest” menu di dwuma de kyerɛ nneyɛe a ɛboa (te sɛ “Track my order”) anaa akuw ahorow, ɛnyɛ nsɛmfua a wɔahyehyɛ kɛkɛ. Kasa te sɛ onipa.Hwɛ wo search logs no mu na woahu nsɛmfua a nkurɔfo de di dwuma ankasa. Sɛ wɔkyerɛw “couch” na wofrɛ no “sofa” a, yɛ bridge wɔ akyi sɛnea ɛbɛyɛ a wobenya nea wohia ɔkwan biara so. Smart filtering.Kyerɛ filters a ɛho hia nko ara. Sɛ obi hwehwɛ “mpaboa” a, kyerɛ no kɛse ne kɔla a wɔde yiyi mu, na ɛnyɛ generic list a ɛfa wɛbsaet no nyinaa ho. Show, don’t just list.Fa mfonini nketewa ne nkyerɛwde a emu da hɔ di dwuma wɔ nhwehwɛmu aba no mu sɛnea ɛbɛyɛ a wɔn a wɔde di dwuma no betumi ahu nsonsonoe a ɛda ade bi, blog so asɛm bi, ne mmoa asɛm bi ntam wɔ animtiaabu mu. Ahoɔhare yɛ ahotoso.Sɛ hwehwɛ no gye bɛboro sekan biako a, fa loading animation di dwuma. Sɛ ɛyɛ brɛoo dodo a, nkurɔfo bɛsan akɔ Google so ntɛm ara. Hwɛ “failure” logs.Pɛnkoro ɔsram biara, hwɛ nea nkurɔfo hwehwɛɛ a ɛsan bae a ɛnyɛ hwee. Eyi ne wo “to-do list” a wode besiesie wo site’s navigation.
Awiei: Search Bar no Yɛ Nkɔmmɔbɔ Nhwehwɛmu adaka no nkutoo ne beae a ɛwɔ wo wɛbsaet no so a nea ɔde di dwuma no ka nea ɔpɛ kyerɛ yɛn pɛpɛɛpɛ, wɔ n’ankasa nsɛm mu. Sɛ yɛantumi ante saa nsɛmfua no ase a, sɛ yɛma Google “Adaka Kɛse” no yɛ adwuma no ma yɛn a, ɛnyɛ krataafa a yɛhwɛ kɛkɛ na yɛrehwere. Yɛrehwere hokwan a yɛde bɛkyerɛ sɛ yɛte yɛn adetɔfo ase. Odi nkonim wɔ nnɛyi UX mu no nyɛ sɛ wubenya nsɛm a ɛsen biara; ɛfa sɛnea wubenya nsɛm a wobetumi ahu sen biara ho. Ɛyɛ berɛ a ɛsɛ sɛ yɛgyae toɔ a wɔbɔ wɔn a wɔde di dwuma no wɔ wɔn syntax ho na yɛhyɛ aseɛ yɛ design ma wɔn adwene. Ɛdenam ahama a ɛne ne ho hyia ankasa so kɔ ntease mu ntease so, na yɛde Information Architecture a ɛyɛ den, a ɛfa nnipa ho bɛboa yɛn search engine ahorow no so no, awiei koraa no yebetumi atoto nsonsonoe no mu.
