We’re always on the look out for new and exciting tools that can help transform our client’s websites. Search has always been high up on our list and we’ve tried and tested a number of different solutions. We were introduced to Klevu about a year ago when their CEO and co-founder, Nilay, came to visit us and gave us a great demonstration on the potential of Klevu. Since then, we’ve met a number of times with their Solutions Architect, Paul Rogers, and our partnership has flourished simply due to the fact that we can safely say this is the best search tool we’ve ever seen for Magento. We know, those are some big words, but we think Klevu deserves them. It’s brilliantly simple to implement and manage, it can be transformed into a beautiful frontend feature, and the experience for the customer is lighting fast and painless.
What’s more, it’s a constantly evolving piece of software that gets smarter with every day your customers are using it to helping to connect shoppers with the products they want to buy, the information they need to find or the content that will inspire them. We’ve seen fantastic results for our clients who have started to use Klevu and we think it’s an essential piece of the Magento package.
Product, information and inspiration is the perfect combination for our client, Sous Chef. They have a huge catalog of speciality cooking supplies prepared for their customers who are often armed with an ingredient list for their next dinner party. A simple search for a food stuff shows the customer a number of options helping them to find the perfect ingredient including top products and related categories. A simple integration allows us to crawl their Wordpress blog along with their Magento catalog to help index their amazing range of recipes concocted to inspire their customers which display when a search is made. Finally, an information required like delivery times and shelf life, is just a click away thanks to ability to index CMS pages created within Magento.
Hey Paul, thanks for joining us today. Can you give us a brief overview of the services Klevu offer?
Thanks very much for having me!
So, Klevu is an advanced, NLP-based eCommerce search solution, that has developed a market-leading product, built around the following pillars (although these pillars do change depending on who you’re speaking to at Klevu haha):
Natural language processing - so far, this has probably been the biggest focus for Klevu and is often the real differentiator for us when competing against other providers. Our NLP offering allows us to provide more accurate results by extracting more meaning from queries and matching this to the data we have from a product catalog. I’ll talk more about this later on.
Machine learning - Klevu uses machine learning to help provide more effective search results, by boosting items that are performing better (based on search-led clicks, add-to-carts and purchases) in real-time. We use this alongside merchandising logic, dictated by the user.
Advanced search merchandising - this has been a big focus for us recently and we now allow merchants to add a lot more complex logic to promote different products and groups of products. We now support individual product boosting, query-level visual merchandising and attribute-based promotions (e.g. boosting products that are “blue” and are in the “shirts” category.
We have various other key features, but they mostly fall into these areas - we’re also doing a lot more with personalised search, which is likely to be a big focus for us in the future.
The services Klevu provide are exclusive to eCommerce search and we only focus on this - we have looked at various other areas (such as category merchandising), but for now we’re solely focusing on improving the core Klevu product and our core integrations.
We’d love to learn the history around the product. How did Klevu get started and what has been the company’s journey until this point?
The first iteration of Klevu’s search solution was launched in 2013 (by the core team out of Finland), but the first couple of years were very focused on developing the product, as opposed to sales and marketing. It wasn’t until 2015 that Klevu really started pushing on these fronts, bringing in various people to help grow the sales and partner side of things, as well as growing the brand.
Nilay and Niraj, who are now both based in London, are the two main Founders and they both have a very strong background in linguistics and artificial intelligence. The primary objective at the start (which is still often referenced in the UK office today) was to bring the optimal eCommerce search experience to retailers of all sizes.
Prior to founding Klevu, Niraj was working on The General Architecture for Text Engineering (GATE - open source group focused on large-scale text analysis), which has been a big part of Klevu’s history also (helping to support our catalog enrichment capabilities). Today, Klevu has a team of more than 20 and has offices in the UK, Finland, India and The US - we also work with a host of global brands, across various eCommerce platforms.
Klevu has some fantastic features as part of the software and we’ve seen sneak previews of some of the new features you guys are due to introduce. Which is your favourite feature about Klevu currently and what are you most excited to introduce?
I personally really like the merchandising side of things, which gives users with a lot more control over which products are being made more visible generally and for specific queries. We recently released “query-level visual merchandising”, which essentially allows for hero-SKUs to be assigned against specific keywords and groups of keywords (as per the screenshot below).
This has been really popular so far! Another one I really like is the personalised suggestions when search is activated, as can be seen below on the Soho Girl store.
This feature also leads us onto the new feature I’m most excited about, which is to see more personalisation logic in our automated boosting capabilities. Currently, the machine learning capabilities of Klevu are geared around promoting products that are performing well generally, we’re now looking to use 1:1 personalisation more in our algorithms, which will essentially mean that we’re promoting products to a user based on the items they’ve personally interacted with and the queries they’ve used.
I’m really excited by this one!
The way customers use websites to shop online and consumer habits are constantly evolving. Have you seen any changes in the way customers are using search tools?
Definitely! There are two big ones that I tend to talk about a lot - these are:
Search as the primary user journey on mobile, which is something that we’re seeing more and more from retailers because users have used search more and also converted better via search (over category-led navigation). Users seem to expect to be served with an Amazon-like experience on mobile - which has generated really strong results for a number of our customers. The example from O’Neills below is a really good one in my opinion (in terms of pushing search and a really good, fast overall experience).
The second point is that users are starting to be more detailed in search queries, which is often attributed to sites like eBay and Amazon, which encourage more advanced queries. We’ve seen this trend when analysing our data and reports like this one from Baymard also suggest that this is becoming more common.
This is also where the benefits of our NLP and catalog enrichment capabilities come in, as can be seen with the below example taken from one of our largest ever customers.
It’s easy to overlook search as a less important feature when it comes to UX and conversion. Why do you think search is so important to a commercial website?
In my experience, users who complete a search are considerably more likely to convert and this is something that is only increasing, along with usage of search. I’ve done a couple of projects recently where we’ve made the search bar a lot more prominent, with the intention of driving more usage and increasing the overall conversion rate across the site - which is something that was achieved in both cases.
Although the conversion rate of search dropped as we pushed more people to use it, we did find that the overall conversion rate increased and there was an incremental uplift through driving search as the main navigation route.
Searching directly for a product or product type is a far quicker and less complex way of finding what you’re looking for, which is why I consider it so important. I also think search represents a far more mobile-friendly experience - which is something more and more retailers are adapting to, as mentioned above.
On average, we find that the conversion rate of users who complete a search is around 4x higher than users who don’t - obviously there’s an argument to suggest that there’s more purchasing intent from users who search, but overall I’d suggest that a user is likely to have a more positive experience by using a well-optimised, fast search function over browsing categories.
What is the most impressive stat you’ve heard from a client using Klevu?
Klevu once worked with a big B2B customer (previously using Magento Enterprise native search) and the uplift in search revenue in one full day paid for their Klevu enterprise license for the month.
Thanks for your time, Paul. Is there anything you’d like to add?
Thanks very much for giving me the opportunity to talk about Klevu and search in general! If you need any more information on the benefits of Klevu or driving improvements to your search function, you can read the following:
If you're interested in utilising Klevu on your website, why not take them up on their free trial offer and ask us to implement this on your Magento instance. Paul can be found writing about and discussing Magento, Shopify and everything e-commerce over on his blog.