Ecommerce site search matters. Consumers across the globe now expect the same search capabilities they get with Google or Amazon when they go onto a retailer’s website.
Time pressed customers want every keyword search to count and return a close or exact match for their queries. This is why market leaders are increasingly deploying artificial intelligence in ecommerce.
Right now, online customer experience is one area that retailers are investing heavily in. Covid lockdowns have funnelled huge swathes of the population through ecommerce sites. These habits aren’t going away as stores reopen. A surge in online traffic means there’s increasing pressure to convert those clicks and generate revenue. Yet e-commerce search solutions are often overlooked.
For many retailers on-site search is a broken experience. A large-scale usability study of ecommerce search last year found that 61% of all sites performed below an acceptable level of performance, 46% of sites didn’t support themed search queries, 27% of sites wouldn’t yield useful results if users misspell a single character in a product title, according to research from Baymard Institute.
Many merchants still don’t realise that they now can deploy the same level of sophistication that Google has done when it comes to answering complex Internet queries, but with on-site search - which has in fact changed very little in over a decade. They can do this by utilising AI in retail.
Several years back Google introduced new algorithms that allowed consumers to type in long-drawn out search queries and get accurate results back. Over time people around the world have changed their behaviour, they often now type in large strings of text when sourcing a product. However, traditional internal search engines cannot handle such complexity. Retailers have been slow to realise this.
Now AI in commerce is being deployed to the on-site search function with incredible results. It’s easy for a retailer to offer a product page to a customer if they type in “red dress,” but if someone types in “red cocktail dress for a party,” there’s a lot more words with different meanings. An on-site search engine needs to be able to understand this correctly and return an accurate result.
This is what AI search is now able to do. Machine learning makes it a lot easier to serve fast and relevant results by analysing millions of customer interactions every day. Natural language processing, which is how machines understand our communication, and semantic understanding, where computers are able process meaning and context behind human information, means that retailers can make sense of real-world complexity when it comes to ecommerce site search.
As shown in the latest case study, the AI-powered search results are also outstanding, they can reduce the number of searches that return “no product found” to almost zero, conversions are up by 20 to 50% on search queries, since an ecommerce site has a better knowledge of what to serve customers, and revenue is also up for retail brands who deploy this tool. This is adding much more value to the on-site search function.
It’s the future of what this technology should look like. But why?
AI works hard behind the scenes for a retailer. It continually learns from every search, conversion and checkout ever made, looking for patterns. AI systems don’t just process search words instead they try to understand their real meaning, and how these terms consciously and unconsciously express purchase intent. In the process artificial intelligence in retail can improve the customer experience and refine this over time. This really matters for product discovery.
It can self-learn what products are the most impactful for any particular query. AI can segment user types allowing retailers to find search and purchasing patterns that the merchandiser cannot do manually. Software compares how different consumers interact with the website and creates groups of similar user behaviour. The AI can then match customer journeys that mirror each other.
By comparing a consumer’s behaviour to a vast database of all other behaviours the retailer can predict the action of a specific customer - users who bought this, also bought that - this demonstrates common patterns. AI therefore makes it possible for retailers to target products to customers with similar behaviour. This leads to better conversion rates.
AI is not just revolutionising the world of on-site search, it can also be used to speed up the process of product attribution. The biggest challenge of any retailer is to have enough data in their product descriptions and attributes. The more products in a catalogue, the less data there will be on individual products. New collections and launches creates new rounds of problems. If products don’t have enough information or tags, then on-site search won’t be optimised.
Artificial intelligence can now be used to speed up the manual process of attribution and in some cases replace it, especially with product images. AI-powered image recognition software can detect colours, patterns and other details enriching the product database. AI can now populate the information in real time as soon as the product arrives in the catalogue and is uploaded to the ecommerce platform.
An algorithm can then be applied to the newly created product data and the AI driven on-site search engine is automatically enriched with the fresh information. If this is done manually it can take several teams a lot of time. It takes time to create this workflow. Artificial intelligence in retail can auto-populate product description data at speed.
The sweet-spot when using this technology is the ability to use the AI tools, which allow for a high degree of automation, with the ability to override or tweak how they work manually. Human control is crucial. This allows you to channel and personalise how the customer views and interacts with the product range. For the luxury sector or homeware as an example, this can be crucial to match certain products, looks and accessories.
Attraqt provides a full suite of AI driven on-site search, merchandising and personalisation capabilities to ensure that retailers and brands of all tiers can have a Google-like shopping experience on their website.