To understand the importance of analytics, you first have to understand that life is full of patterns and we are surrounded by this daily. The ability to tell how some things are likely to happen is probably what keeps most of us sane. Patterns, as you probably already know (or don’t), are the sequence in which things consistently happen. In humans, patterns become a behaviour. In financial markets, they become trends. Over time, these sequences become the definition of the medium used in carrying them out and then stereotypes develop. 

In small doses, patterns are easy to follow and discern. An example is your understanding of a best friend's behaviour - their repetitive patterns - and your understanding of their intentions. For the sake of this article, let’s refer to this as small data analysis. However, understanding the behaviour of a large collective of people, over a longer period, is more difficult. We would refer to this as big data (see the article on why Africa needs to get big on big data). Imagine a million different people walking into your life with different intentions that may or may not change every time they walk in. You couldn’t possibly keep track of this, not unless you’re employing analytics. 

Analytics is the derivation and communication of information (actionable insight) by the systematic analysis of data or statistics. It is the bridge from gathering data to you making decisions that affect your business. Usually, analytical tools use mathematics, statistics, predictive modelling and machine learning techniques to find meaningful patterns. There are software solutions dedicated to this - I’ll highlight some further down the line. By putting these tools to work, you can ‘clear your doubt’ - find out if what you think to be true is actually true, as well as start to answer questions that you did not even think to ask. 

The three predominant types of analytics I’ll be referring to in this article are:

• Descriptive - Understanding the status quo - what happened and why it happened? 

• Predictive - Predicting customer behaviour, through techniques like [jargon alert*] support vector machines, neural networks and random forests, by determining the probability that something will happen. 

• Prescriptive - Answering the question on the best action to take if a predicted scenario comes around. The key here is being able to use big data, contextual data and smart computing algorithms to produce real-time answers.

Great Picture 

The African retail industry is slowly transitioning from an informal sector to a more organised and structured formal sector. 

This is thanks to our popular ‘rising middle class’ (Bearing Point, the Dutch consultancy, believes this group of people could further rise from 350 million people to 900 million people in 2024- more than the middle-classes of China and India combined) which continues to encourage the modernisation of the retail process, as well as provide experimental opportunities for those in the marketplace. According to Deloitte, most countries in East and West Africa are registering a retail revenue contribution to GDP that is greater than 50% of total GDP. You might have already noticed an emerging shopping culture and the non-stop springing up of shopping centres. All of this is exciting. A.T. Kearny believes that Sub-Saharan Africa will remain the next big thing for the next several decades as it is one of the few markets with an annual growth rate of more than 5 percent. A recent report from the same firm further attests to the potential of popular markets like Nigeria and Ghana, as well as dynamic markets such as Gabon and Angola. 

Again, all nice and exciting. 

I feel a ‘but’ coming…

It, however, takes more than just being excited about an encouraging market to win in it. You have to prepare and carefully implement a strategy that works. For all the talk, Africa’s growth is influenced, and fuelled by a rising consumer market (population) than an abundance of resources. The consumer facing industries represent the single-largest business opportunity on the continent, McKinsey reveals. 

So why do some companies still struggle in turning this potential into profit? 

Africa is far from being a homogenous market. It can be divided up by cultural differences, purchasing power and even political stability - all of which vary in various nations and are key drivers of consumer behaviour. Across the continent there has been consistent migration of rural populations into cities (making it easier to target consumer groups) but this hasn’t automatically resulted in the automatic change of taste and capacity. Therefore, making business decisions on macroeconomic perceptions without further depth can be dangerous. For example, national level data on GDP per capita can very easily mislead on the national average - due to a high concentration of wealth in certain places. Also, simply because there is a ‘rising middle class’ and favourable demographics, does not mean that it can be translated as higher consumer spending. 

What then affects spending?

Leading retailers let us in on the fact that changes in consumer lifestyles and ambitions are what actually influence purchasing behaviour. By paying close attention, they have been able to deduce that South African consumers are becoming more aspirational and brand-conscious while in Ghana consumers are increasingly attracted to products that are well packaged and well documented. This increase in discernment has affected the quality of goods consumers expect and their willingness to pay for it. Analytics has been the secret behind deriving this. 

Adding colour 

“50% of companies who master the art of customer analytics are likely to have sales significantly above their competitors.” - McKinsey. 

Considering that retail businesses run on the model of trends and ‘hot items’, it would be a bit of a self inflicted joke to ignore analytics in its entirety. Furthermore, in an increasingly ultra competitive retail sector, relying solely on past sales data and intuition can land you in trouble. The African Continental Free Trade Agreement is set to establish the largest free trade area in the world since the creation of the World Trade Organisation in 1995. As other countries join,   “AfCFTA will cover more than 1.2 billion people and over $3 trillion in GDP” the Council on Foreign Relations reveals. This means the game just got a lot more crowded. To survive, maximise profit and gain customer loyalty, retailers must understand their customers better while becoming lean and efficient. Today, big retail businesses use data analytics at all stages of the retail process. They are using data analytics to forecast future demand and sales, track emerging popular products and optimise product placement. 

Here’s how it works

As a retailer, you’re either online, offline or on both. 

As an online retailer, your channel of contact with the consumer is solely over the internet (E-commerce) which means that all of your interactions with them can be measured. The most important points of interactions are associated with site traffic (how many people are ‘coming into’ your online shop), order transactions (how many of these people are buying or attempting to buy something), marketing & promotions (how many more people are coming in after a promotion or engaging with a promotion), customer services (what are my ratings like and what feedback am I getting) and inventory management (how often do I need to restock particular items). 

As an offline retailer, all of the above listed metrics are just as important. You just have to go about gathering the data to be used for analytics in a different way. Your channel of contact with the consumer here is face to face so the most important form of interactions happen as so. Customer feedback is one, invite customers to discuss with staff and employ technology that helps you to structure the feedback that they want to give. People counting and dwell time is another, understanding the movement of customers through your store will help you to better implement store layouts, product placement while gauging customer response to your marketing efforts. Queue measurement can be used to monitor trends as they develop and inform your staff to act in real time when queues are getting too long.

There are various other data gathering methods made possible by people counting cameras and monitors that gauge how frequent customers return by reading IP addresses. Unique situations and goals will influence which solution you choose to adopt. However, neglecting to put these measures in place could cause you to lose valuable customers who you are unable to better serve. Gathering data on all of the above listed will help you to understand what is going on within your business (descriptive analysis). Once you understand the current situation you are ready to move on to predictive analysis. Based on the data you decide to gather (there are a lot of predictive analysis you can choose to run - usually with the help of technical experts and/or with available software) your predictive analysis options will vary. An online retailer can choose to incorporate a predictive search on the site. Click through analysis reveals the products that most customers are after. By being intelligent enough to predict what the consumer is looking for, you’re saving them time and getting them to the buying stage quicker. Offline retailers, through computing data gathered, can go on to predict promotional impact. You’re now also able to offer recommendations and promotions - machine learning can help you understand a customer’s behaviour, using their purchasing history, and prompt recommendations that have a higher probability of being appealing to the customer.  Predictive analysis also helps you to manage pricing, showing you the correlation of various prices with sales performances over time, allowing you to set the right price at the right time. 

Another key area, within online retailing, is fraud management. By employing industry-specific models that identify potential fraud scenarios, predictive analysis is able to reduce the overall cost of retail fraud. Supply chain management can also be managed using predictive analysis. If you’re able to gauge next months demand then you’re in a position to plan, forecast, source, fulfil, deliver and return inventory. 

Now that you know what’s going on in your business and can even predict what will most likely happen, you’re set to make a decision. But you’re not done with analytics yet. Prescriptive analysis can be used to validate all your key decisions. It helps the manager to uncover business opportunities that human data analysis cannot achieve because of how improbable or impractical it is. In these cases, retailers must use big-data solutions to harbour vast quantities of data in a structured way and then employ data analysts and solutions to help validate decisions as well as identify problems within datasets (how things are currently done) that could be holding back sales.

Analytics, when used effectively, can bring about greater competitiveness. It is, however, still important to highlight the vital role humans play. From the required expertise needed to analyse the data gathered to human judgement in having the last word. Humans are vital throughout the chain, in understanding when analytics is needed, what kind of analytics is needed and when it isn’t. So when you are thinking about the tools and machines you need to get to the point of better-informed decision-making, remember that the people involved in the process are also key to successfully leveraging analytics. 

Tolu Oni is a Director and Lead at Thread Strategy.