Search đ in e-commerce today
An in-depth look into Search: Behaviours, Query types, Results and Usability
Offline in person we have always been searching. Asking a waiter which meal they would recommend. Asking store staff where can you find the eggs..or a particular department? Asking a passer-by for directions.
These all share the same motivations â immediacy â° (I need to know something right now) and intent đ đ(so I can buy, browse, compare). Search has always navigated us towards something. đș
Online Search also answers these basic human motivations. Today, it would be hard to imagine life without a Search for your holiday destination, hotel, product, address for maps, train times, flight times, taxis⊠right from your fingertips on mobile.
There are multiple parts that make up the Search experience. In a nutshell, you enter a query, result(s) will be matched and presented to you.
To begin with, weâll look at search query types.
Search Query Types
Primary (Query types) â these define the range of search
There are 4 main query types. These are Exact, Product type, Symptom and Non-product.
- Exact: when you have an exact product in mind. Eg. âPink Lady Appleâ âProduct SKU 1243452342â
Product type: when you donât yet know the exact product, but you know the type. Querying by a category. E.g. âFruitâ
Symptom: when you donât know the product solution, but you search by symptoms. Eg. âFeverâ âThirstyâ
Non-product: when you are looking for something that isnât a product. Instead, is navigational or informational. E.g âContact supportâ âAccountâ âOrdersâ
Secondary (Query Qualifier) â modifier of primary query
Secondary query types are more sophisticated and modify the primary query
- Feature: when including one or more features of the product they are looking for. E.g.. Colour, material, brand. âgreen appleâ âOrganic appleâ âBritish apples. Typically, these after Search, these are applied as dynamic filters.
- Thematic: when looking for more conceptual or theme. E.g âSummer fruitâ âBirthday cakeâ. âPartyâ âSummer dressâ. For this type of query modifier to be supported, it requires thematic tagging of the product catalogue to determine which products match themes.
- Relational: when youâre looking by a person related to the product, instead of a specific product in mind. For example, a film director âRidley Scottâ, or designer âCath Kidstonâ .
- Compatibility: when you donât know the specific product (accessory or spare part) that you need, but have the details of the product you already own. For example, âiPhone 6 chargerâ. âRed wine paired foodâ
- Subjective: when you are looking by more subjective terms. For example: âCheap applesâ could return apples sorted by price. Or apple products on promotion surfaced to the top. Best applesâ, Search could infer they are looking for highest-rated and highest selling apples.
Query structure
- Slang, abbreviation, typo and symbol Searches: when looking by linguistic shortcuts. Mapping between different terms and pairing. For example âbevvyâ â alcohol. Inch unit and the â character. Example: mlk we could assume the user means âmilkâ.
- Implicit Searches: when looking for a product within a category. For example, if a user is on a âFruitsâ category, then searches for âAppleâ we can assume they are looking for fruit apple, not the technology brand Apple. Purchase history, products in basket, demographic information. If the user has browsed men clothes, then searched âbagsâ, we can assume intent of looking for bags within the Men category.
- Natural Language Search: when a query is submitted by spoken language. This can contain all of the above query types. The combined queries are then interpreted and a meaningful response is returned. Voice assistants.
New searches methods worth mentioning
- AR (augmented reality) search: using real-time image analysis, combined with other queries to interpret and return results. Examples, using phone camera above restaurant menu â popular choices highlighted. Using AR lens on a sofa, that returns sofa models and buying choices. Real-time translation by interpreting images.
- Image search: Taking a photo, that is analysed and interpreted by a search engine, what the product type is and combining with other query types. Example: Photo of a tomato returns ideas for tomato recipe (compatibility query) suggestions. Taking a photo of a plant = Returned results of the identified plant.
Avoiding âNo Resultsâ Dead End
A âdumbâ Search will return âNo Results Foundâ and give the user no alternative options or direction. This is known as a dead-end experience, and should aim to be avoided. Typically, dead-ends will have a high amount of user-friction and abandonment.
Implementing Search and Expected Behaviours
A sophisticated Search will be able to interpret as many query types as possible.
- Auto-complete search: when matched (or partially) search queries. Example: typing âappâ auto-compete could suggest âapplesâ âpink lady Appleâ. If the user has historically viewed or purchased apple crumble, âApple crumbleâ could feature on the auto-complete (implicit search).
Final thoughtâŠ
As Search becomes more sophisticated as technology advances, userâs mental framing of the potentials of Search grow. E-commerce sites must evolve to match Search expectations, or collaborate with major Search Engines to optimise their content. As users become more impatient, the efficiency and relevance of Search is more important than ever to create frictionless experiences.