Technology moves on apace. In this blog I consider the developments in the area of facial recognition and detection and how it can impact on the ongoing relationship between retailers and shoppers.
This technology has had its origins in the area of police and security protection. There is ongoing debate as to the efficacy of the use of facial recognition and detection when it is used to scan people going about their business in the streets and highways. Some argue that it is a gross invasion of privacy. Others maintain that it helps the police to detect criminals and terrorists and that it can lead to successful prosecutions. Clearly this leads to polarised views as to how facial recognition is used. While police detection rates may improve as a consequence, the ever-encroaching influence of “Big Brother” moves closer. In terms of personal freedom and democracy, powerful arguments can be put forward to restrict its usage.
It is not the intention of this blog to enter into sustained discussion on political or social matters here. Our focus relates specifically to how the retail sector approaches such technology.
Clearly it has some potential benefits for retailers.
Software such as FaceFast technology can scan faces as far as 50 to 100 metres away. Many retailers use variants of such technology to identify existing shop-lifters and dishonest people. This can help to decrease the losses from theft and arguably leads to lower prices for shoppers (as retailers traditionally pass on this cost in the form of higher prices). This use of radio facial recognition might be called “shallow learning” – from the perspective of the retailer. By this I mean that it is not used for “deeper learning” such as linking individual faces to data on that customer which has been gleaned from loyalty cards and previous purchases in the store.
Such a linkage can then be used more strategically in the form of triggering sales associates in outlets to make use of this information to personalise their approach to shoppers as they come in close proximity to them.
This conjures up an interesting experience for that shopper to say the least. For instance if you enter a Zara retail outlet (presupposing that you have downloaded the relevant Zara app and you have your smart phone switched on) and a sales associate approaches and says “Hi Bill, Welcome back”, How would you react? Would you feel reassured and relaxed about being part of the “Zara community”? Or would you feel uncomfortable about such familiarity at being referred to by your first name? Many of us probably would not feel too aggrieved by such a tactic as, superficially at least, it poses no threat to us and may engender a “feel good” factor.
It is when we move to the “deeper learning” aspects that some shoppers might feel uncomfortable. For instance Seven Eleven has around 11,000 stores in Asia. It uses facial recognition technology to address the following issues.
- Identify shoppers that hold loyalty cards
- Analyse in-store traffic
- Monitor inventory levels on the shelves
- Suggest appropriate products to shoppers, based on their previous purchasing patterns and preferences
- Measure the individual emotions and moods of individual shoppers as they walk around the store
- Monitor the most popular areas that shoppers visit within the stores.
This provides some “wins” for the retailer as it can use such information to create a more relevant and personalised experience for shoppers. Likewise customer can benefit from a slicker and more relevant visit to a store.
We can see yet again (a common theme running through the blogs) the confluence between technology and data.
In some ways such developments as facial recognition technology bridge the gap between what goes on in the retailer-shopper interface within online channels and physical outlets or stores. Up to now it could be argued that online retailers have enjoyed a major advantage in being able to capture the data about individual shoppers and use it to promote and provide personalised offers.
Now, with in-store technology, retailers can also put together customised offers. Online and offline information can be merged.
For instance the International Finance Centre Mall in Seoul, South Korea uses its information kiosks for such a purpose. As a customer approaches the kiosk, the cameras identify the person’s age and gender in real-time. It can then personalise the interactive advertising surrounding the kiosk accordingly.
We also see increasing use of cameras hidden in digital billboards to effect a similar response and experience for the shopper. For instance in the Westfield shopping centres throughout Australia, such cameras capture age, gender and also the mood of the shopper as they pass by the digital billboards. They can then conjure up personalised messages or adverts on the screens of the billboards.
Westfield uses software developed by Quivindi (a French software firm) in 2015.
Tests have demonstrated that it is accurate in 90 per cent of the cases.
It can identify five categories of mood, ranging from very happy to very unhappy. Clearly mood is an important measure for advertisers as it can indicate the level of sentiment towards a brand. While clearly more difficult to measure, when compared to age (accurate to within five years of a person’s age), it can glean some useful information nonetheless.
It is important to note that this is an example of facial detection, NOT recognition. In this case all of the data captured is anonymous – it is not linked to the individual’s past purchases and preferences. It identifies the characteristics of the individual: not who they actually are.
It does not take a genius to recognise that it does not take much work to take this technology to a higher level of “deeper learning”: where the face of the individual is recognised and is linked to all of the existing and ongoing information that is already captured. In such cases the retailer can shape and groom the shoppers to specific and highly personalised messages, promotional offers and brands.
A study by RichRelevance (2015) indicated that 68% of respondents described facial recognition technology as being “Creepy”. 63% favoured a mobile personalised app which would identify item locations within the store.
Although the survey was conducted in 2015, it indicates a degree of resistance on the part of shoppers to such technology. However it is also likely that such attitudes can change, as shoppers become more comfortable with it and can see likely benefits to their overall shopping experience.
We are clearly going to see more usage of such technology going forward.
It will be interesting to see if legislators address issues such as the privacy (and possible intrusion thereof) of shoppers as they engage with retailers within an in-store experience.
Amazon’s pioneering cashless stores are beginning to roll out. In this case shoppers can pick up items, place them in a bag and leave the store without have to check each item out individually. Neither do they have to queue up at a check-out to pay. They simply walk out of the store, with the payment automatically conducted with relevant technology. A large part of this process relies on facial recognition and arguably it works well for the shopper as it speeds up the process of shopping. For many of us check-outs, whether manned or self-checkouts are a “pain-point”. It causes delays and for many of us who are “time-poor”, this represents a major improvement.
The debate about the intrusion of privacy and “Big Brother” will remain a constant issue. Let’s see how it pans out in the future.