<img height="1" width="1" style="display:none" src="https://q.quora.com/_/ad/ad9bc6b5b2de42fb9d7cd993ebb80066/pixel?tag=ViewContent&amp;noscript=1">

User Experience Guide

How to improve your eCommerce site-search

Book a demo

user-experience-guide-hero

Introduction

The ability to search is important for guiding visitors to the right products. So important, in fact, that visitors who buy are 90 percent more likely to use search than those merely browsing.

A poor search experience can have devastating effects on the bottom line since visitors who receive null-results (i.e. no results at all) are three times more likely to leave and, more often than not, never to return (source).

Search_box

There is a lot you can do to improve the function of your search box

The basic function of the search box is for visitors to enter a search query and receive a sorted list (from most relevant to least relevant) of the right products. Unfortunately, because people don't always search the way you want them to and in ways computers can easily interpret (i.e. they often misspell words, use other product names or search in ways that do not produce a match in the product catalogue), effective search functionality requires more than searchable product descriptions, it needs to understand and interpret the intended meaning of the search query.

How and what are your visitors searching?

In a physical store, people have become accustomed to navigating the different sections and aisles of the store. For example, in a grocery store, people generally know that the milk is at the back of the store and that the candy is located near the checkout. In an online store, products can be organized and categorized in a similar fashion. Using filters, a visitor searching for milk can fast-track their way to the right milk product by filtering available products down to the dairy category.

When it comes to search, however, the visitor may not use the same vocabulary as what is used in the product catalogue. But by analyzing search patterns over time, it is possible to establish a good foundation for making improvements to search functionality.

"It's difficult for computers to properly interpret human behaviour and intent. The basic product search engine should understand the words entered in the search box and show the correct products as a result. But that's trickier to do than one might think.

For example, the visitor might search for 'running tights' but in the catalogue, they are referred to as 'sports leggings' or 'running pants'. In this case, few search engines would return results. This leads to incorrect assumptions about the assortment of products available and lost sales revenue.

In a physical store, this wouldn't be a problem because the store staff would know what was meant by term 'running tights'. But in eCommerce we're forced to find another way to handle the interpretation." - Robin Mellstrand, CEO at Loop54

Excluding irrelevant search results

It is just as important to get the correct search results as it is to exclude products that are not relevant to the search.

A visitor who searches for 'spaghetti squash' would naturally expect that a product like 'squash soup' would come up in results, but would not expect a 'squash racket' or baby food containing squash to appear.

It's important for the search engine to decipher the intended meaning of the search, not simply look for matching words in the product metadata.

Autocomplete_featuresAutocomplete ( incremental search ) means that you get the idea for keywords or phrases while you fill in the search box. It also expedites the search experience on mobile, where screen real estate is limited and chubby fingers are prone to error.

Different search methods and visitor expectations

Baymard Institute conducted an e-commerce search usability study and identified several search methods used by their test subjects.

The search selection:

  • Exact search - where the user's search query matches exactly to the product name.
  • Product type search - where the user searches for a type of product rather than a specific item (e.g. laptop computer -vs- Macbook air).
  • Problem-based search - the ability to search for solutions (i.e. products) by entering a problem or an experienced symptom. This is especially prevalent in certain industries, such as pharmaceuticals, where the users know the problem but not how to solve it.
  • Non-product search - where the user is looking for auxiliary information, such as opening times, terms or delivery options.

The limits of search:

  • Feature/function search - when the visitor includes one or more product features as part of their search query, such as colour or format (e.g. blue jacket).
  • Thematic search - when visitors search for thematic product categories or by intended usage (e.g. party dress, winter jacket).
  • Relational search - when the visitor searches for a person or object that is linked to the product, such as a book's author, music artist or clothing designer. Particularly important because they often don’t exist as categories in the site’s hierarchy and therefore can’t be easily accessed any other way than through search.
  • Compatibility search - when the visitor searches for products that are associated with another product, such as a phone charger or a computer case.
  • Subjective search - when a visitor uses subjective words (e.g. quality, beauty, value, etc.) in their search queries (e.g. cheap backpack or high-quality toaster). Requires the search engine to interpret the subjective meaning and form its own opinion.

The structure of search:

  • Slang words, abbreviations and symbols - for technical equipment, brand names are often shortened (e.g. HP for Hewlett Packard). The dimensions of products can also be described differently (e.g. 0.5 litres or 500 ml).
  • Natural language - when visitors write search queries the way they would explain them to verbally to another person (i.e. as a full spoken sentence).
  • Implicit search - when visitors submit partial search queries where certain aspects are only implied. Often a search that is linked to the context the visitor is in (e.g. searching for pants when in the women's category would imply a search for women's pants).

There are many parameters to take into account when designing good search and it can be hard to cover all bases. If you can not implement every best practice, select the ones that you believe are most relevant to your visitors.

Baymard Institute recommends that the first priority should be exact search, product-type search, feature/function search, thematic search and relational search. Begin by analyzing your search statistics and do usability testing when possible.

If your search statistics indicate that customers are looking for products you do not carry, maybe you should consider adding them to your assortment.

If you already carry the product, but under a different name, you will need to tag the product with synonyms or use a solution like Loop54 that will learn synonyms automatically.

Polivore_faceted_search

This search function gives results, but also the categories and brands visitors can use to refine their search results. A feature called - faceted navigation.

How can you practically improve your search functionality?

Always start by analyzing how visitors search your site today. You do this by activating Google Site-Search in Google Analytics.

The next step is your eCommerce platform. There is probably some form of basic search functionality included with your platform. Although this is a good place to start, you will quickly notice the limitations of this option and will want to adopt a more sophisticate search engine.

6 tips for designing a great search box:

  1. Design a box so it stands out graphically and is clearly visible. Visitors should find it immediately.
  2. Include helpful text in the search box, such as: "search for products, categories or brands". The text should disappear automatically when the user starts typing.
  3. Keep the search box populated with the visitor's search query. That way, the user can narrow their search without having to go back and remember what they previously searched.
  4. Allow the keyboard enter key to start the search.
  5. Adjust the size of the search box. If you have long product names a larger box makes things easier for visitors.
  6. Use auto-complete to suggest searchable keywords and phrases after only a few letters have been typed. Include the various categories to help refine search results.

Loop54 offers true personalised on-site product search

cta-shoes

Shoppers expect the same level of relevance and personalisation online as they experience in-store. Powered by Machine Learning and built exclusively for eCommerce, Loop54 delivers that exceptional online shopping experience.

  • Automated: Automatically learns words and merchandises search and category listing
  • Relevant: Interprets search intent to deliver truly relevant results
  • Personalised: Sorts results according to popularity and personal taste

Request a demo today!