The rise of the mobile personal shopper - how your phone camera can predict your next fashion purchase

Online shopping in the UK is booming. Last month, the UK Cards Association revealed internet spending had grown to £154bn in 2016 - an average of £422 million a day. The average UK household spends more shopping online than any other country, writes Pixoneye CEO Ofri Ben-Porat.

Growing smartphone usage, greater spending power among millennials, and innovations in digital wallets, one-click payments, and enhanced security will only increase the nation's propensity to opt for online commerce rather than visiting brick and mortar stores.

While online retail has undoubtedly provided a host of benefits for shoppers, allowing them to purchase products from the relative comfort of their sofa rather than competing for a parking spot at their local business park, department store, or fighting the crowds being loaded with bags on public transport, we risk losing a key benefit of physical shopping.

The key attraction of High Street shopping, especially in the fashion industry, has been the ability of the in-store experience to satisfy all the senses ahead of making a purchase decision. With such varied influences and styles available, stores offer the benefit of a personalised service - having someone on hand that can help select the clothing combinations that best match your desired look.

This physical experience cannot adequately be replicated online. Your package from ASOS or Boohoo arrives only for you to decide that the dress you bought on your mobile to relieve the boredom of the train journey home doesn’t really fit with your wardrobe. Indeed, it is estimated that two thirds of shoppers who bought women’s clothes online send at least one item back. This also represents an expensive issue for retailers who’s bottom line is hit hard by honouring free returns through costly postage and pick and pack facilities.

For consumers that enjoy the service experience of in-person shopping and the convenience of shopping online, is it possible to receive a personal shopper experience in a digital world?

Smartphone nation

The UK is one of the world’s most mobile-centric nations. As many as 81% of UK adults - and 91% of 18–44 year olds - use smartphones.

Our smartphone image galleries have in many ways become a reflection of our lives. Every day we document our lives and activities across multiple channels, whether it be sharing holiday snaps on Facebook or getting a friend’s opinion on a new purchase via WhatsApp or Messenger. Our phone gallery captures many of the details, decisions and nuances of our lives.

Each person’s gallery is unique and distinctively theirs. If someone were to look through your pictures they’d speedily know more about you than from a simple conversation. They’d get an accurate depiction of your interests, tastes, favoured possessions and desires.

Each image you have stored contains a range of data attributes and meta tags that reveal a 360 degree perspective of you; your purchasing habits, interests, preferences and desires. It can also be the key to predicting what purchases you’re likely to make in the future. This rich trail of data, combined with browsing and purchasing activity, gives retailers an opportunity to personalise their offers to our tastes and preferences.

Image galleries provide a more accurate profile of a consumer than "skewed" social media

Predicting your next purchase

Using image recognition and understanding technology, retailers can – with the consumer’s consent – anonymously gain access to the data contained within image galleries. The retailer (or, more accurately, its marketing team) will not see the images themselves, but simply be able to use the data contained within for a full 360 degree, contextual understanding of the individual, their lives and motivations. A fundamental USP of this technology is that the images are not extracted from the device to a server or otherwise, but only a graphic representation of the images themselves in the form of a feature vector.

In the fashion industry, this means that smartphones can effectively play the role of a mobile personal shopper, with the power to understand the customer’s desires and predict their next purchase. The retailer can analyse its customers’ styles, helping them to identify key upcoming moments that will influence imminent purchasing decisions and tailor their offering accordingly.

Are you starting a new job and need suitable professional attire? Are you about to go on holiday and need a new set of beach clothes? Have you been on a fitness bootcamp ahead of the holiday and will be looking to refresh your wardrobe with slimmer fitting clothing? Do you need appropriate dress clothes for wedding season?

By layering deep learning and artificial intelligence on top of on-device image recognition, retailers can use your smartphone gallery to predict these moment. Combined with existing datasets and strategies, this data brings greater relevancy to suggestions and offers that will mean the consumer receives a more tailored offering. Finally, brands can answer the consumer’s desire for a truly personalised, curated shopping experience.

Satisfying demand through personalisation

Using the data contained within images to predict an individual’s future needs supports retailers to get the most out of m-commerce and satisfy their customers’ needs, before they even arise.

The quality of this data is far superior to the demographical and hyperlocal (e.g. weather, events, etc.) data that retailers have traditionally relied upon to make marketing decisions, and adapts to changes in customers’ lives in real-time.

This level of data sophistication exists only on device. A full offline view of the individual, rather than a skewed and unrealistic social media perspective image analysis gives only a snapshot of the range of data available. (We all recognise social media is a poor reflection of real life).

The key is personalisation. Online, consumers take personalised experiences as a given, and this expectation is increasingly being mirrored in physical environments. Consumers expect brands to understand them, and use the data they provide to offer products and services that are right for them. Indeed an IBM report from 2015 noted that only 22% of customers considered that brands understood them. What a poor return for their marketing departments versus marketing spend invested.

Image recognition and contextual understanding presents the opportunity for retailers to gain an unprecedented perspective of their end users and present the right content, products and services to them when most relevant, regardless of whether they’re shopping online or in-store.

By offering a more curated and personalised shopping service within a digital environment, retailers can improve the shopping experience for their end users and in turn, foster greater loyalty and potentially even reduce returns.

In a crowded retail market, it's the brands that are able to predict their customer’s next desires that will reap the benefits of the mobile age.