This is a Guest Post and does not necessarily reflect the thoughts and opinions of Techzim. We have a strong filtering process of what makes it to our blog and are confident that you’ll enjoy the article below.
Big Data, Machine Learning and Artificial Intelligence are the most common buzzwords that are often thrown around even by some novices in the tech industry just to sound smart or to evade having to explain the things they don’t even understand themselves [ I hope it’s not the case with me here 🙂 ]. Nevertheless, in this post, I want to talk about a niche application of big data analytics and computer vision which is an emerging use case of foot analytics by retailers to understand/influence customer’s buying behaviour.
Disclaimer: while this post focuses on the use of foot analytics to understand customer behaviour in brick & mortar stores, the author acknowledges the benefits of online shopping and digital marketing tools like web traffic and click analytics. The author suggests the use of foot analytics in offline stores to complement & not replace the digital marketing tools.
But first things first, let me explain what do these terms; Foot Analytics, Computer Vision,… mean.
Computer vision — this is a field of computer science that works on enabling computers to see, identify and process images in the same way that human vision does, and then provide appropriate output e.g. automatically detecting & counting people in a store & determine their gender from a security camera footage.
Foot analytics — The set of technological tools that enable the collection, analysis & visualization of pedestrian traffic into a place of interest (for example a supermarket) so as understand and retain shoppers through understanding what drove that visit and re-target them to influence a purchase.
Customer Behavior — Refers to the totality of a customer’s decisions, concerning the purchase, consumption and disposal of goods over time, Customer behaviour is dynamic in nature & can be influenced through pricing, product assortment, product placement, store location, in-store promotion among other stimuli.
Why Foot Analytics in a Digital Economy?
Offline Retail vs Online Shopping
Now we have cleared the air on the basics of foot analytics, the question in your minds could be, ‘why bother about foot analytics in offline stores when the world is increasingly going towards online shopping & e-commerce’. Well, even though it’s true that online shopping & e-commerce in developing countries, particularly in Africa, is rapidly growing; offline shopping or brick & mortar shopping is still dominant in this part of the world. Here are some of the statistics on African consumers:
- 88% of connected consumers use the Internet to find products they want
- 40% check prices online, then buy in a traditional store
The Challenges of going digital in Africa
The major reasons why African consumers prefer offline/brick & mortar vs online shopping to a greater extent are not due to illiteracy or lack of knowledge as one might assume on face value but they are mainly due to exorbitant mobile internet data costs, poor logistics (delivery services) and a general lack of trust in divulging personal information owing to the threat of identity theft & online fraud or scams.
High data costs — Depending on where you are in Africa, purchasing a gigabyte (GB) of mobile internet data can cost as much as an equivalent of USD $35. The data alliance for affordable internet(A4AI) conducted a survey of 60 low and middle-income countries and found Africans pay at least 5% more than their counterparts in other developing countries such as the Middle East and Asia as shown below;
Lack of / unreliable logistics services — last mile delivery of products is an important factor in determining the success of online stores & e-commerce. In developing countries, particularly in Africa, there are no/very few reliable logistics providers that offer door to door delivery of items purchased online at affordable costs. Large courier providers like DHL & FedEx, even though they may be present in the aforementioned regions they often do not offer affordable door to door delivery in African countries and online buyers will have to resort to collecting their items at the courier depot which will be not less than 5 km, on average, from the customer’s place of residence.
The fragmented logistics & courier service in Africa is one of the major deterrents to online shopping as it introduces hustles in the form of delays and additional costs to consumers as compared to offline shopping where a customer goes into a supermarket purchase their product, pays in cash and proceeds with their life.
Now that we have talked about some of the challenges that are being faced in online shopping in Africa & why offline shopping is there to stay in the foreseeable future, lets now look at why foot analytics is a huge opportunity to brick & mortar retail businesses in developing countries.
The Benefits of Foot Analytics
By utilizing foot analytics, retailers open themselves to a wealth of new knowledge and insights on the performance of their businesses through measuring the following key performance indicators (KPIs) which they were not able to measure or track before:
- Dwell time in the store
- Average shop times across a particular time of day or day of the year
- Parts of the store that customers visit the most and the least — for A/B testing of new shelf displays & potential for up-selling/cross-selling
- Gender distribution of customers — measure the common customer behavioural characteristics by each gender type
- Cross-store data comparisons — for retail chain stores
- Determine the relationship between pedestrian traffic volume and sales volumes/value at any given point in time
- Optimize Your Staffing — determine your staffing needs e.g. security, till operators & store assistants
All these and many other use-cases or advantages that are unlocked by foot analytics enable accurate data-driven decision making by managers and store owners in predicting and influencing customer behaviour by altering the following:
Product pricing — by knowing the demographic & behavioral characteristics of the potential customers who frequent the store the challenge of product pricing is lessened when the store owner or manager can price products appropriately to suit the customers whilst maintaining a healthy profit margin and avoid overpricing (which could lead to loss of sales) or under-pricing(which could lead to loss of additional revenue)
Product assortment — When determining which products to range in your stores, there is no doubt that you need to consider your target customers. By knowing which type of customers frequent your store, you then to look at your specific retail format and then match your product assortment to the expectations of your customers.
For example, if your pedestrian analytics reveal that 75% of your customers who frequent your shop are female, you as the shop manager or owner now know that these customers expect to find sanitary wear in your shop at any given time so as to maximize customer value & avoid loss of sales to your competitors since sanitary wear is often bought with other accompanying products.
Product placement — While the product pricing and range can influence customer behaviour, foot analytics can provide insights on how to present them to the customer in the most effective way to please and not to frustrate the buyer/customers.
When products are merchandised correctly, it not only means they are not only easy to access but easy to see. More than that, you create stores placements that are aesthetically pleasing to a point where shoppers spend more time in them (and buy more products) than they had initially expected to.
Store location — Picking the right store location isn’t only about finding a place with the most foot traffic because foot traffic doesn’t automatically amount to paying customers. Instead, you need to choose a location that is accessible to your target market.
For example, if your target market is a lower Living Standard Measure (LSM), you would want to find out where they live, work and shop and then place a store in an area that they can reach easily, there is also a need to consider their restrictions as well such as the fact that they might need to use public transport to get to your store.
In-store Promotions — These are an excellent opportunity to entice shoppers in, and when used effectively, can keep your customers in your stores for longer. More than that, they can attract new customers, who wouldn’t have thought of coming in, and spark excitement.
That said, it’s not merely the fact that you should run promotions, as much as you should think about where you place them. this is where foot analytics come to use, if the marketer knows the demographic characteristics of the intended audiences in-store displays can be directed to the demographic that yields the most value, for example the children’s toy banner display targeted at fathers can be significantly different from a children’s toy banner display targeted at mothers and foot analytics helps the marketer in deciding which in-store promotional display to put up in the store.
Why Use Computer Vision? — and not just count in person or use cellphone movement data?
Large retail chains in developed countries such as in Europe & USA have been using foot analytics to better understand their customers but they use entirely different technologies from the ones that are being suggested by the author for use in developing countries. The technologies in current use in developed countries such as cellphone movement tracking and others are not accessible or are difficult to use in developing countries due to the following setbacks:
- Exorbitant data costs
- Mobile network operators are reluctant to share location usage data with vetted/trustworthy third parties due to lack of skill or appreciation of the beneficial effects
- Not all customers have smartphones— smartphone penetration rate of 33% so mobile phone tracking excludes children and those without smartphones, etc
In spite of these challenges, in developing countries, almost all retail shops have CCTV security cameras installed and are in current use. Through the use of these CCTV cameras, computer vision algorithms can be used to collect customer data for use in conducting foot analytics through tracking the behaviour & movement of customers from the point they enter the store, select purchases up to the point where they pay & exit the store all without any need for active human involvement.
This enables accurate data collection that is consistent thus trustworthy to yield credible insights that are indicative or reflective of customer behaviour. Also, the presence of stored video feeds means that the foot analytics tasks can be re-run to verify any curious data patterns that deviate from the norm & that could have raised any questions with the business owner, analyst or researcher.
And, above all, the main advantage of conducting foot analytics through computer vision algorithms on security cameras is that apart from the software and computational additions there is no additional cost of deploying or implementing foot analytics using foot analytics in the form of data licensing fees to Mobile network operators or third party data providers but also opens up the possibility for the business to monetize the data it could have generated to third parties such as analysts and researchers, or the media.
The major drawback of conducting foot analytics using computer vision is that it is limited to analysis of in-store customer interactions (within the CCTV camera’s field of view) and thus cannot be used to get a complete picture of the customer journey from their home or workplace. Also to note is that appropriate notices must be put in place to inform customers that they are being recorded for analytics or security purposes and vigilant data & privacy protections must be put in place through anonymizing customer information or redacting facial information to prevent hackers from accessing sensitive customer data for malicious intent thereby risking erosion of trust by customers which will in turn scare them from patronizing your store ,in the end, defeating the purpose of foot analytics in the first place.
In conclusion, the application of computer vision in conducting foot/pedestrian analytics is a HUGE opportunity for retailers (particularly retail chains) in developing countries in Africa & Middle East to unlock customer value and influence customer behaviour through data-driven decision making with little/no room for error.
Isaac .T. Chikutukutu is Software developer-turned-data-Scientist, holder of an Honors in Business Studies & Computing Science, participated in several projects spanning from Web & Mobile Software development, Networking, Data Collection & Analysis and Computer Vision for over four years.