On-site search represents an important part of online retail, particularly for merchants with complex products (e.g. branded cameras with different lens dimensions etc) and user journeys. In addition to the benefits to the end user, merchants can also benefit from the data e.g. the sorts of searches that consumers make being provided and the ability to merchandise and optimise search results.
Up to 30% of users will be using your site search on an ecommerce store and according to a research from Econsultancy, conversion rates through site search can be up to 50% higher than the average. This is testament to the importance of site search in online retail in 2016 and beyond and, if you’re not dedicated to ensuring you make the most of your site search, you are missing out on a lot of potential conversions.
It’s important to remember that not all retailers will have the same need for on-site search, SKU-heavy merchants for example will want to encourage users to search in order to find a product that may be difficult to find otherwise – whereas some merchants might use it only as a fall-back when users struggle to find what they’re looking for via categories.
This article will look at on-site search best practice, new technology, optimisation and how to measure the impact and performance of site search effectively.
Klevu is a leading provider of search technology – which is used by ambitious merchants of all sizes all over the world. Klevu has built a solution designed to change search completely – our technology is cloud-based and results are usually served in under 300 milliseconds. It is, however, the self-learning capabilities and ability to understand NLP in the ecommerce domain where Klevu’s search truly excels.
Klevu can be integrated with most ecommerce platforms, including Magento and Shopify (where we have existing modules / integrations available). Klevu can also be integrated quickly and easily with more enterprise-level platforms, such as Demandware, Hybris and IBM Websphere.
Site search is something that almost all ecommerce stores will have, but just having the functionality doesn’t provide the full value for conversions and customer experience. In addition to search being functional, you need to think about how it should work in terms of product serving, product discovery, merchandising etc.
The use of search can be largely dependent on your industry or type of products but, generally speaking, there is still a best practice that should be adhered to that will help ensure your search is as effective as it can be.
The statistics speak for themselves when it comes to site search and its importance to your ecommerce store. You can expect that up to 30% of ecommerce shoppers will interact with your site search – something that makes your conversion rate optimisation even more crucial.
Additionally, just 40% of ecommerce stores offer faceted search (though four out of five of the top ecommerce stores offer it) and this is a factor which has shown to lead to increased conversion levels and needs to be taken into account.
There are some fundamentals to site search which remain fairly consistent and which should be adhered to in order to yield the best results in terms of conversion and customer satisfaction. Outlined below are ten of the most pivotal features to remember when creating the site search functionality on your own ecommerce store:
Users that perform a search are in most cases more likely to convert, as they’re going directly to a selection of products that meet their criteria without having to browse, so making your search bar easy for them to find is really important. This is something that needs to be addressed as part of a UX / CX / design consideration, but it’s one of the fundamental elements to get right from an ecommerce conversion point of view.
A different colour or a clear outline can also help users distinguish the search box from the rest of your site’s design.
In addition, a small piece of copy such as ‘enter a product name or brand’ can also make your search stand out and guide users quickly to the functionality they need. If you do go down this route, it’s important to ensure that the text represents a clear call to action and ideally entices the user to input. Similarly, an icon such as a magnifying glass is commonly used to represent search functionality and can provide further visual guidance for your users. These both seem simple, but lots of merchants move away from these.
There are plenty of examples of ecommerce stores that don’t make their online search bar a key component of their design, restricting the number of users using search as part of their user journeys. Another major appeal of search is that it gets the user to the product they want far quicker than any other route, which is why it’s important to showcase the functionality (as long as your search works).
Some platforms, including enterprise-level options such as Magento and Demandware, provide very basic search functionality out of the box, which is likely to be far less impactful for complex queries and searches that don’t include parts of the product name. They also don’t provide the same level of reporting as the more advanced solutions and SaaS offerings.
There are lots of examples of online retailers who have done a really good job of encouraging users to use their search functionality:
These are just a few examples of merchants that showcase their search box well – although if you look at different ecommerce websites, you will see that lots of the retailers position their search box to the left of the main header, which appears to be a recent trend.
Another consideration when it comes to designing your search box is making sure that you get the size of it right, as users will want to be able to see their full search query whilst they are typing it and review it for any typos or mistakes.
There are certain sectors (such as electricals) which will have long, specific product names that might not fit within conventional search boxes. It may look like a small consideration but one that is worth bearing in mind because it can be a subconscious factor in a conversion.
An example of a product name that will be too long for a conventional search bar would be something like “Sony Cyber-Shot WX350 Compact Camera 18.2MP Black” which, as you can see, is a highly specific search term but if you’re looking for a product with lots of similar models, users will be more likely to input more detail into the search box.
You can see from the Currys example above, that they’ve made the search box wider than you’d usually see, which is likely to in order to facilitate for more detailed queries. Amazon is another very good example of this.
It is, of course, important to consider site search on mobile devices and responsive design needs to be considered to ensure that your search functionality isn’t downgraded in any way for those consumers viewing your site on a mobile or tablet.
Mobile site search will grow in importance exponentially as more and more consumers begin to shop online with their smartphones. It is best practice to have a mobile optimised website but it is important to retailers as well because if a consumer can’t locate your search functionality, you might just miss out on a conversion as a result.
With mobile transactions making up 30% of sales in online retail in 2015 and this anticipated to increase in 2016, now is the time to make sure your ecommerce store is not just mobile friendly but that your search functionality is optimised as well.
Auto-complete tools have become really popular over the last couple of years, as have suggested and popular queries. Auto-complete, in particular, saves users a considerable amount of time by suggesting different types of results as they input their query.
Another trend you may want to consider is offering images alongside product names and price to help users qualify the results more. Lots of Klevu implementations, as below, house search results in an expanded window, whilst providing images and pricing info.
Another fundamental of search usability is trying to avoid 0 results pages – you should always try to deliver results to the user, which is where natural language processing comes in very useful. You could also consider providing alternative suggestions with or using ‘Did you mean Product X’ to avoid users being left frustrated.
A page displaying ‘No matches’ or ‘No matching products found’ also provides an opportunity to feature complimentary products and still carry the user onto a product page. There is always scope to serve up products of some description even if you don’t stock exactly the item the user is looking for, you can cross-sell to them instead.
There’s no excuse in 2016 for not delivering some actionable product result from a site search on an ecommerce store and a “no products found” page simply isn’t good enough any more.
The use of rich search results has been another big trend over the last 12 months, as retailers continue to embrace the whole “content & commerce” focus. Rich search results provide a broader search experience for the consumer and, resultantly, will help improve conversions in the long run. Below you can see an example of rich search in action on the Red’s Gear Store:
As you can see, it provides search results in real time that includes category pages, different materials etc which hones in on the user’s search query and makes it much more straightforward to find exactly the product that they are looking for. Some merchants also include guides and blog posts within this.
Your users will not always know the exact name of the product they are searching for. In order to make your search function as helpful as possible, ensure your search can handle typos, colloquial names of products, and takes cultural alternatives into account. A user searching for “trousers” in the UK will need relevant results for “pants” in the USA, for example.
Klevu’s search functionality is tailored towards issues of this nature and the ranking of products is decided based on where the terms you are searching appear in the product. In practice, what this means is that if a word is appearing in the name, it will receive the highest relevancy score.
If a word appears in the description, it receives the least relevancy score. Attributes configured to be displayed in the search layered navigation are given better weighting than those selected as other attributes to index – the end result of which is a search functionality that is always learning and providing the most relevant results.
Auto-complete will also help guide users with smart suggestions and reduce errors, but your search should still be able to return relevant results based on everyday language and slang. This is by no means an easy feat, of course, and that’s why solutions like Klevu can help to improve the functionality of your search bar and deliver genuinely useful and applicable results to the consumer.
We’ve already covered the importance of making sure your search doesn’t deliver a ‘no results found’ page but it is equally as important to ensure that the search results that you do provide are as accurate and effective as possible. For example, if someone searches for “32” black trousers” and they’re served black trousers that aren’t available in that size, it will be a damaging experience from a conversion point of view.
Lots of search solutions use things like meta keywords for product association, whereas more modern solutions are thriving on descriptions covering wide variety of use-cases. Users searching for detailed queries are likely to be fairly far along the buying process and anything that negatively impacts their experience (such as inaccurate search results) needs to be addressed.
Accuracy is one of the most crucial components of site search and that’s why tools like Klevu are such an important weapon in any ecommerce store’s armoury as they make it much more straightforward to configure the search and provide results that aren’t questionable or lacking in accuracy. In addition to features that enhance search performance, Klevu also provides advanced reporting around the effectiveness of results and any errors that have occurred, helping to address these.
There’s no telling how specific or detailed users are going to be when they’re searching your catalog, so it’s important that you provide options for them to further refine the results. It’s good practice to provide layered navigation filtering on search results, to give users filtering options based on your product attributes.
For example, if the search term is something quite vague, like “footballs”, providing filtering around the size, colour, brand etc will make it easier for them to find the correct product.
This is another good example of where, leading nicely onto the next section, you should be using the data to understand what people are doing on search results pages. In an ideal world, you’d use a solution that learns from how users are responding to queries, which would automatically optimise the results being returned.
Search data can provide lots of insight into how users are using your store, although most merchants only look at the search term data.
One of the important things to understand is how search is influencing the conversion path and how journeys that include a search differ from those that don’t. Generally, the merchants that are actively encouraging users to search have done this analysis and they’ve found that search has a very positive influence on conversion rate, as users are able to find their desired product much faster.
Here are some other metrics that you should be looking at:
Regularly examining this data will give you a much better understanding of how users are interacting with your site and whether search is being used as optimally as it could be. Getting an overview of the type of searches that are regularly carried out on your ecommerce store can often be surprising and the sorts of queries which you might expect to have been prevalent might not be so and, conversely, you might discover that people are searching for products that you don’t have – which can be used to inform future stock purchases.
Lots of search solutions will provide comprehensive data around the use of search, but it’s also important to remember that Google Analytics can be utilised to provide lots of additional data points. For example, you can use events to track things like 0 results pages, enhanced e-commerce can provide lots more detail around search results page performance etc.
These guides are very useful for getting more data via Google Analytics:
Despite its prevalence within search queries, a lot of ecommerce stores don’t specifically allow for abbreviations and symbols within their online search and this is a fact which inevitably has a knock on effect on their conversions.
An example of this is something like a search for 42 inch television and 42” television which, theoretically, should both provide identical search results but, in actuality, an awful lot of ecommerce stores don’t have the relevant search functionality to provide identical results for these queries even though the consumer is clearly looking for the same product with each.
Having a solution in place that effectively deals with abbr*eviated terms or ones with symbols is something that can have a big bearing on conversion and from an online search point of view, it is certainly best practice to make sure your search box allows for this manner of query and delivers the same set of results regardless of how the query is structured.
The auto-completion of searches for consumers is a feature which appeals and there are a number of sites who do this extremely well such as Zimmermann Wear where, for example, if you start typing “black” it provides results like the below:
The more that you type, the more specific the results it displays will become so for example, if you type “black shoes” it will immediately filter out every product which isn’t relevant for this query.
Another site which provides great functionality with their search is that of Baby Bunting who provide a similar user friendly approach to auto completing search queries and providing various different types of content including product, category and blog pages within the delivered search results:
It’s important to understand the importance of keywords but not to become over-reliant upon them for the configuration of your onsite search. For example, you might have a product such as “red Nike trainers” but you need to consider the other terminology that consumers might use such as “sneakers” or “kicks” and this is where software like Klevu really comes into its own because it learns the language of the consumer and adapts the results that it supplies accordingly.
Technology is, by definition, always advancing and improving and this is no different so far as ecommerce site search is concerned. The last couple of years has seen lots of innovative and intuitive search tools introduced, which have helped improve conversions and user experience – the most obvious examples are self-learning technology, significantly faster search results, rich search and the way that results are displayed.
Accommodating slang and colloquial search terms can be a heavy undertaking. New tools based on artificial intelligence (AI) can help and have become very popular. A concept known as “choice paralysis” is becoming increasingly used in ecommerce and it essentially refers to how consumers are often overwhelmed by the sheer volume of choice offered by today’s online retailers that they find it difficult to make an actual decision. This is where artificial intelligence site search might just come into play in 2016.
AI-powered search, or “intelligence search” can handle queries of varying length, abbreviations, slang, synonyms and typos. Significantly, AI technology does not rely on product teams to build metadata or invest heavily in tagging to ensure the right results are displayed. Instead, AI interprets meaning from your users’ search terms and delivers the relevant results.
The most advanced search tools now use machine learning to record each query, and importantly, which results users preferred. As more and more users click on results, the technology learns more about its users and delivers results based on real user behaviour rather than marketing assumptions.
There are all sorts of exciting changes afoot with site search in the future as a result of the dawn of AI and just some of the most relevant and beneficial from an ecommerce point of view include things like image recognition whereby a consumer can input an image of a product they like or are looking for and the AI search will serve up the most relevant results based on this image.
Mobile search is another consideration, as the user experience is very different to desktop. This is another area where data is important, as it may be that search is more or less effective for mobile users.
Lastly, new technologies will emerge for online retailers to complement their site search technology to help them act more proactively, in real time, to how people are engaging with their website and they will lead to stores reaching out to customers as they shop as opposed to waiting for customers to come to them with queries etc.
Conversion rates are commonly higher in bricks and mortar stores than they are with online stores but brands will try and redress the balance with the new technologies that become available to them that will let them offer real time engagement for web shoppers such as more intuitive live chat and offering real time assistance if a consumer seems to have hit a brick wall on their customer journey.
Effective onsite search needs constant refinement and you’ll always be looking for (and finding) ways to improve and enhance the way in which you implement your onsite search. Put simply, test, test, and test again.
If you’re looking for a good place to start with optimising your own onsite search then consider trialling the following elements to determine what works best for your users and your ecommerce store.
Take the time to trial different result prioritisation and then monitor how the user interacts with the results that you’ve served. Consider ranking results by relevance, product availability, or customer ratings to see what has the most traction with your users.
This sort of segmenting of results will be imperative to understanding what it is consumers are looking for when they use your site search and changing it up to find the most effective layout will have a big impact on the amount of users you are able to convert into sales.
You won’t, of course, be able to serve up a result prioritisation that caters to the needs of every user but by trialling different orders you’ll be able to come to a conclusion as to which offers the best user experience and conversion rate.
As mentioned earlier, mobile ecommerce is all set to increase in 2016 and ensuring that you’re ready for how consumers will be using their devices to make purchases is a must. Make sure your search function can be easily accessed and used by shoppers on all devices as having a fiddly mobile interface which makes it difficult to utilise the search bar will be a massive roadblock to conversion.
Search buttons should be large enough to be easily found and results should be large enough to select. Utilising best practice in mobile design will show you where best to deploy your search bar on the mobile version of your ecommerce store and making sure you get this right in 2016 will lead to more conversions being made from mobile devices.
Location, location, location might typically refer to property but it is also applicable to your online search because where you choose to position it will have a big bearing on its impact. Experiment where the search box works best on your site because finding the optimum position will represent a massive breakthrough so far as conversion is concerned.
Ensure that, wherever your users are throughout your ecommerce store, the search box remains visible, prominent and user friendly because, chances are, most users will look to use it at least once during their customer journey. If moving the box has implications on your site design, measure the impact with A/B testing to ensure that it does not have a negative impact.
There’s no harm in regularly testing the location of the search box but don’t change it daily because it will be frustrating for the user to come back to the site regularly and not being able to find the search box because it’s moved to a different position each time they return. When testing the location of the search box, give it some time to test its effectiveness before you trial it in a different position on the site.
Your customer service department will have a lot of insight into the exact words and phrasing that your customers prefer and this is something which you should be looking to tap into to make sure you’ve got a broad overview of what it is your customers are looking for.
Apply their knowledge of your customers to your onsite search strategy to ensure your keywords are the most applicable for your users. This is something that will take a little time to facilitate but with smart search tools like Klevu and various analytic tools at your disposal, you should be able to get to the bottom of the most popular search terms on your site and hone in on offering exactly the right search results for your consumers.
As well as finding specific products or a previous page they may have visited, your users will often be searching for customer service information. Consider including search terms such as ‘shipping’, ‘returns’, and ‘store locations’ to deliver popular information quickly without users having to dig through reams of results.
Site search is all about understanding, and often pre-empting, what it is your customers are going to be looking for from your website and whilst a lot of searches will be product oriented, there will also be a lot that surround more functional things like opening times, delivery costs etc and making sure you cater to both is essential for conversion.
As well as providing insight into the popularity of certain products, search data can show you exactly what information your users are looking for, and crucially, what information is missing from your site.
Search terms and their associated volumes can highlight where informational pages such as shipping or returns policy are lacking required information. The data can also be used to refine product descriptions or titles and ensure the user can find all the information they need in one place.
As mentioned earlier, there is a lot of data at your fingertips (as an ecommerce store owner) that can be used to make informed decisions about your consumers which can then be rolled out to help improve site conversion.
Some of the suggested metrics that you should be investigating for onsite search include the below:
This is just a selection of some of the core onsite search metrics that are worth considering when it comes to getting to the bottom of your consumers actions and how they engage with your store.
Once you’ve compiled the list of data from the metrics above you should have a comprehensive understanding on what you’re doing well with your onsite search and, conversely, the areas which are in need of some improvement.
On-site search holds massive potential for retailers in 2016 and being proactive about how you approach it is a must for all retailers large and small. Its fundamentals are often basic, but can be extremely effective so making sure you’ve got them covered is essential. Unlike many elements of ecommerce, search functionality amends can be small, but significant, arguably offering some of the highest ROI for your site.
We hope that this definitive guide to ecommerce site search will help online businesses thrive in 2016 and beyond and helps provide an insight into the inherent importance and fundamental role that it plays in driving conversions, sales and customer satisfaction in the ecommerce environment.
Our search solution is different to most of the others in the market, as it understands the natural language of shoppers – meaning the technology understands shoppers even if the keyword searched isn’t clear or doesn’t match the exact keyword in the product catalog.
To achieve that, Klevu performs two Natural Language Processing (NLP) tasks:
1. Catalog enrichment before the data is indexed
2. Query processing to understand the intent of a consumer
The query processing involves adding synonyms, inflections, semantic categories, text normalizations etc. to the product catalogs, ensuring whatever information is added is relevant to the context of the product. As part of the query processing, Klevu identifies the terms those should be searched in product catalogs vs the terms used for highlighting preferences (e.g. preference for price, color, occasion etc).
Klevu also self-learns and optimises search results – meaning it learns from shoppers search behaviour and adjusts results accordingly. This is one of the main reasons why merchants move over to Klevu.
Please enter your details below and download PDF of the definitive guide to ecommerce site search.
No obligation to continue, No credit card required.