Covid-19 quickly necessitated the closure of brick-and-mortar stores, forcing everyone online. Now shoppers are accustomed to the convenience of online retail – from quick delivery to free returns.
To cope with this exponential growth of ecommerce channels, Artificial Intelligence (AI) personalisation is an effective way to support online businesses scaling up. By using huge amounts of customer data, AI helps deliver the right product to the right shopper at the right time.
However, can personalisation thrive in the age of GDPR, where data privacy is a growing concern? It’s a defining time for big data, with technology behemoth Google announcing the phase-out of cookies by 2022.
For online brands and retailers, this means that gathering accurate shopper data won’t be easy.
Unfortunately, data privacy isn’t the only issue retailers face when it comes to data.
Historically, product data has never been easy to manage, but it’s becoming more of a challenge as product catalogues continue to grow from a few thousand to millions of items.
Additionally, the concept of ‘seasons’ is less prevalent, with products being pushed and removed from the website at a greater cadence – in some cases daily. Data harmonisation across sources (i.e. suppliers) is another key challenge, often resulting in inconsistent product information. These factors put substantial pressure on trading teams, who are struggling to provide a relevant shopper experience across touchpoints.
In parallel, we’ve witnessed the emergence and expansion of new online retail channels such as ‘social commerce’ and Google shopping. This means the ecommerce store is no longer the only place to browse. Brands and retailers are not in control of how products are merchandised across these new channels.
What’s more, shopper behaviours are changing. Long search queries are becoming more common as shoppers are providing additional details of the context of the search in pursuit of more relevant results. For example, instead of simply searching ‘yellow striped dress’, customers are searching with longer phrases: “I’m looking for a yellow striped dress for a cocktail party”.
The role of trading is no longer to focus on optimising top searches or category pages like ‘dresses’. Instead, ecommerce sites need to be optimised for every interaction.
So, how can you embrace these challenges with the support of AI, in a quest to still push the right products in front of the right customer?
Well, you need to optimise product discovery.
AI technologies have recently developed understanding of the products and their purchasing context.
In its early years, AI was solely used to link a lot of customer information to support customer-centric personalisation. But with the development of natural language processing and image analysis, AI technologies are now able to build deep understanding of products.
We’re now able to identify logos on shirts, colours and patterns. The application of this technology brings on a brand new level of understanding of the catalogue which could drastically support the optimisation of relevancy of products brought to customers.
An example of this is ‘Shop the Look’ recommendations. A popular type of recommendation that suggests the remaining products displayed in a styled model shot, to encourage upsell and entice the shopper to – as the name indicates – shop the entire look.
While effective, these recommendations used to be manually created, requiring a significant human investment from the trading team.
Now with AI, these recommendations can be automatically generated on the fly, requiring no manual adjustments. This not only saves substantial human costs but also enables stores to scale up great experiences to the whole website in seconds.
Natural language processing and artificial neural networks have also enabled search platforms to better understand search queries – even the most complex ones. Through customer interactions, AI learns the connections between the context of a shopping intent and the key characteristics expected in the product they’re looking for.
For example, during the start of the pandemic, we’ve seen an increase of the term ‘COVID’ in multiple search queries on fashion websites. Without the use of AI, this term would usually bring up no results, as it’s extremely unlikely that stores would have this in product descriptions. However, with the use of AI and no other adjustments, we’ve seen the engine naturally bringing back face masks that were part of the catalogue. Our AI powered search engine learned from shopper behaviours and adapted the response to meet a new intent.
Bridging product understanding and context understanding through AI, is an efficient way for trading teams to make sure they are offering customers relevant products. Especially as shoppers’ requests as well as the product catalogues, are growing in size and complexity every day.
Another benefit of optimising relevancy and context understanding, is the ability to scale outside of your owned channels where personalisation tools used to be focused. By making sure your product offering is relevant to the customer, you’ll get a strong conversion whether your shoppers are coming from Google Shopping or social channels. By optimising your product discovery, you’ll be able to scale up across your full online channel ecosystem.
With recent advances in AI and the ability to connect disparate data sources, product discovery goes beyond presenting the right product to the right shopper at the right time. In today’s fast-moving retail space, where channels are constantly evolving from Amazon to Google Shopping, relevancy in product discovery is no longer limited to the website. In order to deliver a seamless experience, it now needs to fuel all the channels and touchpoints relevant to your shoppers.
By Pierre Bizeul, Head of Product Management – Attraqt
Request a PDF copy of our Business in 2022 report.
Want to join The Collective, and contribute to the debate?
Email us at: The.Collective@lewissilkin.com