Websites offer different strategies for navigation. This article briefly explains the differences between search, browse and filter and how they connect together.
BROWSE
‘I don’t have a specific product in mind, but I have some general criteria I want met and I want the website to suggest some products that meet those criteria. ‘
We can think of ‘browsing a website’ as like being taken on a guided journey. I have an idea of what I want – a general destination – but I’m not quite sure which product meets my needs and need some guidance. The hope is that by following the paths laid out by the website I end up on a page that hits the spot.
Websites have different features for this type of navigation – including editorial pages, drop down menus with links, hyperlinked buttons/text. A user might start their journey at the front page, click on something of interest which takes them to a second page, then follow a hyperlink to a third page and so on. Browsing options on websites must play the role of both physical store layout and store staff who help the customer find a product to meet their needs.
Since the consumer can only follow the paths offered, browsing is a useful way for a website to direct the consumer journey. Browsing involves curation: it is the website which sets out the journeys the customer may follow. Which paths are offered will inevitably depend on the available meta-data, but customers’ use of the site can offer clues as to what journeys to offer. For example, if a certain search term is consistently entered at a high frequency, it can be an indication that this is an item that ought to feature in the navigation menu.
On its own, browse only really works well for websites with a small product range as the customer needs to be able to navigate around the whole product set with only a few clicks. For websites with large product ranges, browse can work effectively in conjunction with search and filter to get the customer to a desired destination.
SEARCH
‘I, the customer, know what I want. I can articulate that in search terms that are globally understood. The website can match these terms and give me relevant results’.
Search can be defined as ‘the act of finding something by looking or otherwise seeking carefully and thoroughly’. The goal of web search is to take the customer’s search terms, map them to terms attached to pages and return the pages that match in order showing the most ‘relevant’ results first. The efficacy of a search engine depends on the data it is using and what it does with it.
I, the customer, can articulate what I want in search terms that are globally understood. Successful search demands that I, the customer, know what I am looking for, am able to express that want using words that (most) other people would use to represent that want, and can enter those words successfully into the search box.
The website can match these terms In order to match the customer’s search terms the website needs its own database of terms to search. Where do those ‘terms’ come from? In general web search – as provided by the search engines such as Google – those terms come from the words that appear on webpages. Site-specific engines search only their own database of terms. They have the flexibility to decide which terms make it into their searchable database, i.e. they can add terms that don’t appear on their webpages and remove ones that do.
Once the database of terms has been created, the search engine looks for a match between the words entered by the customer and the terms in the database. At its simplest it just looks for an exact match between the search term and the term attached to the product, and only with an exact match is the product page returned.
More sophisticated search engines can deal with inexact matches, such as spelling errors (‘their’ versus ‘there’), orthographic differences (‘colour’ versus ‘color’), close meaning matches (‘car’ versus ‘automobile’) and linguistically related words (‘kitten’ versus ‘cat’, ‘leaf versus ‘tree’). The cleverer the system, the more likely a match will be found and the less likely the engine will deliver irrelevant or zero results.
The website can give me relevant results Customers want to the see the results that ‘best’ match their needs before the ones that are less helpful. Relevance is then about ordering of results. Once search has found the ‘hits’ – all the pages that map to my search words – relevance is about prioritising them for me. How do websites prioritise their search results? That depends on the search engine they are using. Bolt on search engines may take into account relative number of occurrences of the search term, its significance in the data set, whether related words occur alongside etc. Site specific search engines – such as that used by TheBookseekers – can be designed specifically around the meta data that is relevant to that product. Since they are bespoke they can decide how to prioritise results based on the needs of their customers and the business they represent.
FILTER AND SEARCH IN SEARCH
‘I don’t want to look at all these results. I’m only interested in the subset that apply to my particular requirements which I’ve not yet told you about’
Successful search should produces lots of results. Browse too may also take the customer to a large set of results. Perusing multiple results takes time and customers are unlikely to want to browse them all, even if they are ‘relevant’. They need help to reduce the results to a manageable number.
Filtering and search-in-search are useful tools that assist the customer in drilling down into the result set, allowing them to ‘find the needle in the haystack’. Filtering uses criteria preselected by the website to split out a subset from a full set of results. Search-in-search is effectively Boolean search with the operator ‘and’, so takes the results from matching the first term and pulls out only the results that match the second search term.
Using multiple filters together can be particularly effective in cutting the number of results set down to size. Since a website cannot offer infinite filtering options – there is just not enough space on the page – the choice of filters should reflect the demands of the majority of customers. If the the meta data is not available to fully deliver the filter, then a website needs to decide whether a partial filter is better than no filter at all.
Search-in-search leaves the choice of category in the hands of the customer. This can be effective if their choice of search terms matches how products are coded in the database. It is therefore less likely to produce results the fewer search terms the database contains.
Filtering and search-in-search methods can work both independently and together alongside browse to navigate around a lot of products, to find the ‘needle in the haystack’. They are therefore invaluable tools for websites with huge product lines. This is increasingly important for companies moving into the marketplace model of ecommerce.
Keywords: search, browse, filter, ordering, relevance, needle, haystack
© Bridget Martindale 4-10-2022