At no time during the past two decades has the Shoprite Group been worth so much more than the Pick n Pay Group. The extent of this deviation is shown below, highlighting that much of the change in relative value has occurred in the past three years.

We’re calling out Pick n Pay to highlight the outperformance of Shoprite because it’s the latter’s closest and long-time rival, but we could make similar assessments about Shoprite against many other listed South African retailers.

Given this extreme deviation in market values, some investors are likely to be banking on a reversion of this trend, perhaps expecting this reversion to be driven by better trading and market share gains by their preferred retailer?

This report intends to account for some of the drivers that have caused the disparity in fortunes between Shoprite and other grocery retailers and to highlight flags that investors can look out for when evaluating if competitors are making up ground.

We do this by looking at the spending of a panel of 22seven users over the past two years. Analysing the same users over time eliminates user churn from the results. Then, to allow us to attribute changes in spending to different types of customers, we add a new dimension to the analysis: loyalty.

A user is loyal to, say, Pick n Pay if more than 45% of their total grocery spend over a period goes to a Pick n Pay branded store. Loyalty helps to group customer numbers based on their preferences for certain stores. More importantly, it allows us to evaluate where growth comes from.

The logic is simple. A retailer attracts spend from three groups of customers: those who regularly shop at the store; those who previously bought groceries from a competitor; and those who tend to shop around.

By attributing change in spend in different time periods to these three groups of customers, we can see which types of shoppers are accounting for growth at the retailer.

Let’s begin by explaining our method. We started by comparing how spending by the 22seven sample has grown compared to the reported numbers from the major grocery retailers. These are shown in the chart below.

The blue columns show Insights’ best estimate of the reported (actual) growth in sales at each retailer for the two years ending March 2023. The red columns show the growth in spending at each retailer for a very large sample of 19,170 shoppers who are also 22seven users, for 1Q23 compared to 1Q21.

Three things about this chart: the sample of 22seven users is likely to overestimate growth at Checkers and Woolworths, and also overestimate the weak performance at Pick n Pay. However, our Spar growth-rate data is similar to what the company has reported. Importantly, the sample data corresponds with the main growth trend: Checkers significantly outperforming Pick n Pay.

At the same time, we can use our sample to unpack additional trends that are not available to company outsiders, while being cognisant of the differences between the sample and the national population.

This is important because, as the next chart highlights, the divergence in spending growth at the major grocery retailers over the past two years has been extreme.

For example, aggregate grocery spending by users earning R25–R60k pm increased by 22.1% over the two years* between 1Q21 and 1Q23 (i.e. 10.5% pa CAGR). For reference, official food inflation increased by 21.9% over the same period.

However, what is significant is the enormous divergence in growth between retailers. In the same R25–60k pm income bracket, user spending at Checkers increased by 47% over two years, while spending at Pick n Pay declined by 7%.

What explains this discrepancy in growth rates?

To attribute growth, we grouped users based on their loyalty in the 1Q21 starting period. As mentioned earlier, a user is loyal to a retailer if  45% or more of their total grocery spend in a period goes to that retailer. A few further points on loyalty:

• A user can be loyal or have no loyalty.
  
• A user can only be loyal to one retailer – we take the maximum loyalty in the rare cases where a user spends more than 45% of their total grocery budget at more than one retailer.
  
• Loyalty is evaluated across eight predominately grocery-based retailers. Insights further distinguishes expenditure by channel and format. For example, our analysis of Pick n Pay may include spending at Pick n Pay supermarkets, and spending at its Clothing, Liquor, asap! and online channels. We have not included Boxer, which is a separate retailer in this analysis. For Spar, we include Tops, Spar2U and Spar supermarkets. Checkers is made up of Checkers stores and Checkers Sixty60.
  
• Woolworths includes grocery and clothing spend because Insights cannot reliably distinguish from user transactions whether purchases are in the Food or  the Fashion, Beauty & Homeware (FBH) divisions. 70% of Woolworths South Africa’s reported revenue is from its Food division, and our other investigations have suggested that Food accounts for the majority of 22seven user spending at Woolworths. Furthermore, over the period of the analysis, the two years between 1Q21 and 1Q23, we estimate very minor differences in the reported sales growth of Woolworths Food and FBH divisions.

Our analysis finds that there is always some churn in loyal customers shifting between retailers, or customers shifting to ‘no loyalty’. The following chart tries to visualise this dynamic process over the period of the analysis.

Can you guess which retailer is which?

We’re delighted you’ve read this far. But research is only valuable if it’s exclusive. To be fair to the retailers and investment teams who subscribe to Insights, we have to say goodbye before we get to the juicy bits.

Contact us find out how our research can help you.

Simon Anderssen  
simon.a@22seven.com
+27 84 730 0309

‍

This report is about spending in the Home category, but first let’s have a look at general spending trends that emerged during July. Apparel was a standout category: 22seven user spending on clothes and footwear held up well – it was one of the few categories to record stronger growth. Elsewhere, spending growth slowed once again.  

Spending on Prepared Food Delivery decelerated the most of all categories, perhaps signalling an end to the boost that services like Uber Eats and Mr D Food enjoyed as a result of load-shedding, which peaked in April and May and has since moderated.

Once again, the fastest-growing spending category that we track was Sports Betting & Lotto. Although it remains relatively small compared to other categories, it continues to grow rapidly.

We identify similar trends within income segments, shown in the table below. In general, the intensity of decelerating growth is more prevalent in the lower two income brackets: users earning less than R40k pm.

Home truths

This month’s Shop Talk focuses on the Home category, which we broadly segment into three sub-categories: DIY & Garden, Furniture & Homeware, and General Merchandise (GM).

GM is the largest of these and includes Takealot, Makro and Game, as well as electronics-biased retailers like JD Tech (Incredible Connection and HiFi Corp). Naturally, there’s overlap between the sub-categories and there’s an element of judgement required to categorise the retailers.

The three largest retailers in each sub-category are highlighted below. The structure of each market is similar, with one main retailer making up 32-42% of total spending; two retailers of roughly equal share as runners-up; and a large number of other retailers (25-35) making up the remaining 35-45%.

The earlier charts show that user spending in the Home category has been the weakest of all the major categories during the past year. When we look at the last three months, our data suggests that this weakness has been concentrated in the Furniture & Homeware sub-category.

The next chart shows relative spending growth among a sample of major retailers within Furniture & Homeware. Yuppiechef and West Pack Lifestyle outperformed while Hirsch’s lost ground.

To get a snapshot of what middle-income shoppers are spending in the broader Home category, we segmented spending into income brackets. In the R15-25k pm income bracket, users spent R657–R1,077 per month at retailers in the three sub-categories.

The chart below further breaks down this average monthly spend into average transaction value and frequency of transactions. In this example, average transaction values at Furniture & Homeware retailers are comparable to average transaction values at General retailers. The difference is that users transact ~23% more often at General retailers, which leads to a higher average monthly spend per user in that category. DIY & Garden retailers capture the lowest average monthly spend, mainly due to low average transaction values.

That’s it for this month. There are many more insights available from our spending dashboards. Please set up a time to discuss these in more detail to ensure you get the most from this unique data set.

If you want to know more about data-driven consumer trends in South Africa, sign up for Nugget, our bi-weekly newsletter. You can also follow Insights on LinkedIn and Twitter.

‍

Some reports have criticised Shoprite’s newest venture: a basics clothing line that not-so-subtly copies the successful Uniqlo brand, which was originally started in Japan. However, since most of South Africa’s clothing brands are built on foundations that ‘borrow’ fashion trends from the northern hemisphere and replicate them here, the criticism of Shoprite may be unfair.

Instead, given Shoprite’s pedigree, formidable scale and extensive resources, it's arguably more important for investors and competitors to keep a close eye on how this new venture is performing…

It will be many years before UNIQ is material to the Shoprite group; and going by prior precedent, its performance may also never be separately disclosed to shareholders, irrespective of how significant the brand becomes to profitability.

With that in mind, we set out to answer four questions to help analysts evaluate UNIQ’s performance to date, and to use this early evidence to contextualise its potential scale in the future.

1. How is UNIQ performing? We look at the market share it has gained since launch in March 2023.
   
2. Who is shopping at UNIQ? We compare the income and age demographics of UNIQ shoppers, compared to other 22seven users.
   
3. What is UNIQ’s market positioning? We contrast the average transaction values at UNIQ with other clothing retailers.
   
4. What is UNIQ’s potential market share? We compare the spending of 22seven users at UNIQ stores in two malls in Cape Town (Canal Walk, the first in the country, and Table Bay Mall) with other retailers in the same malls, including Mr Price, Ackermans, CottonOn, PnP Clothing, Edgars and others.
   ‍

How is UNIQ performing?

In June 2023, UNIQ accounted for a respectable 30bp of the aggregate spending of all 22seven users at the 140+ clothing and footwear brands that we track. This puts it on a par with a host of other recognisable outlets.

In the chart below, we compare user spending at UNIQ with spending at JAM Clothing and G-Star. These brands were selected because all three had the same share of spending during the period, and JAM and G-Star are useful extreme examples that we can use to highlight differences in each retailer’s market positioning. Insights data allows us to split market share into three factors:

• Number of customers: Is there an increase or decrease in the number of customers transacting?
• Frequency: Are customers shopping more or less frequently compared to previous periods?
• Average transaction value: Are customers spending more or less per transaction?
  ‍

Lets begin with the number of customers. In this example, nearly twice as many 22seven users shopped at JAM Clothing than at UNIQ, and nearly 10 times more users shopped at UNIQ than at G-Star.

The difference in frequency of purchases is less significant: during the same period, the average customer made 1.2 purchases at each of these stores.

‍Average transaction value shows the reverse trend to number of customers: the average purchase at G-Star is nearly 3.6x greater than the average purchase at UNIQ, and ~7x greater than what users spend at JAM Clothing.  

The chart highlights that JAM is a high-volume, low-value retailer, whereas G-Star is a premium outlet. UNIQ seems to sit between these two extremes, and the overall result (so far) is that each of these retailers, despite different positioning strategies, make up a similar proportion of total user spending.
‍

Who is shopping at UNIQ?

Digging deeper, we separate the users who have shopped at UNIQ into age and income brackets and compare this to the distribution of all 22seven users. The results suggest...
‍

We’re delighted you’ve read this far. But research is only valuable if it’s exclusive. To be fair to the retailers and investment teams who subscribe to Insights, we have to say goodbye before we get to the juicy bits. Contact us find out how our research can help you.

Simon Anderssen  
simon.a@22seven.com
+27 84 730 0309

‍

Trend-setting taxonomy

The apparel market encompasses a wide range of products and caters to various customer preferences. While some retailers specialise in specific product categories, others offer a diverse range of clothing options for all demographics. For example, Markham focuses exclusively on menswear, whereas Mr Price operates as an own-label department store, offering formal and casual apparel for women, men, and children. Additionally, related categories such as footwear, accessories, beauty, and homeware are also prevalent in the market. And in today's landscape, nearly every ‘clothing’ store will also sell you a smartphone.

To enhance the dataset of actual consumer transactions, Insights has refined the categorization of spending within the Apparel category. We have introduced three sub-categories for spending: Clothing, Sports & Outdoor, and Footwear. This allows for a more detailed analysis of spending patterns across the 165 individual retailers that comprise the Apparel category.

The chart below displays the composition of the Apparel category. The majority of spending is attributed to the Clothing sub-category, which includes major retailers and department stores such as Mr Price, Edgars, Ackermans, Markham and others.

For analytical purposes, we have grouped retailers like Sportsmans' Warehouse and Decathlon, along with standalone stores of major sports brands like Nike and Adidas, into the Sports & Outdoor sub-category. This grouping proves valuable in our analysis, and the chart demonstrates that the contribution from the Sports & Outdoor sub-category increases with higher incomes.

Finally, the Footwear sub-category, while relatively small in size, distinguishes retailers primarily focused on selling shoes, such as sportscene, Tekkie Town, and Crocs, from other categories within the apparel market. Spending at specialist Footwear retailers consistently represents a notable share of total apparel spending across different income brackets.

Pulse check: a leading indicator of clothing retail sales

Sales growth is the most important driver of corporate value* and Insights’ estimate of users’ clothing spending offers clients an early indicator of near-term category growth. Moreover, this estimation helps identify retailers that are either outperforming or underperforming.

Over the past five years, the 22seven growth index has exhibited a strong correlation (around 77%) with the monthly growth in StatsSA's Textile, Clothing, and Footwear Retail Sales Index. This correlation underscores the increasing reliability of the 22seven growth index as a valuable metric.
‍

In the context of a weakened market, the following chart highlights the retailers that have gained or lost the most market share among 22seven users over three and twelve months.

Among 22seven users, Shein remains the fastest-growing apparel retailer. We recently published a comprehensive analysis of how 22seven users make their first purchase from za.shein.com, including the retailers they engage with during the lead-up and those they reject. Further details can be found in our publication.
‍

DuPontifying changes in market share

Investment analysts are likely well-acquainted with the DuPont analysis, a powerful tool that dissects return on equity (ROE) into three vital components: profit margin, asset turnover, and financial leverage.

Typically, applying such an approach to analyzing market share proves challenging.

However, equipped with transaction data from 22seven users, Insights seeks to emulate this approach to deconstruct the driving forces behind changes in market share. This decomposition process focuses on three key elements:
‍

1.     Customers: Examining whether there is an increase or decrease in the number of customers transacting.

2.     Frequency: Assessing whether customers are shopping more or less often compared to previous periods.

3.     Average Transaction Value: Analysing whether customers are spending more or less per transaction.

The accompanying chart illustrates each of these factors for a selection of major apparel retailers, specifically for 22seven users earning between R15,000 and R25,000 per month. The purpose is to decompose the change in market share between the first quarter (1Q23) and the second quarter (2Q23).
‍

Through this analysis, it becomes evident that attracting customers is the most significant driver of market share growth. All retailers that gained market share in 2Q23 were successful in attracting more customers compared to 1Q23. Conversely, retailers that experienced a decline in market share attracted fewer shoppers.

While Frequency and Average Transaction Value occasionally contribute incremental improvements in market share, a consistent trend is not clearly discernible from the data.

Checking out bash.com

Recent media articles indicating that TFG's newly launched online marketplace, bash, has outperformed its established competitor Superbalist in terms of online traffic serve as encouraging signs for TFG shareholders. However, since bash is consolidating traffic from other TFG brands, like sportscene.com and Markham.co.za, establishing the extent of net traffic growth is more difficult than these reports suggest. And traffic doesn’t always translate into sales.

Investigating this issue further, Insights’ identified significant growth in bash transactions during June. Notably, bash ranked either first or second, based on the value of spend, among the 17 TFG brands that Insights monitors across seven income brackets.

Differentiating between existing TFG online domains and bash.com, Insights estimates a net 10% increase in transactions between May and June. As the chart demonstrates, bash.com transaction volumes increased c3.4x between May and June, but mostly reflect redirected payments from existing domains.
‍

For context, 22seven users recorded a comparable number of transactions on za.Shein.com in June 2023 as they did on bash.com, however the value spent on Shein was 75% higher.

If you haven't already, read this report on Shein. We think you'll learn something new..

That’s it for this month. There are many more insights available from our spending dashboards. Please set up a time to discuss these in more detail to ensure you get the most from this unique data set.

If you want to know more about data-driven consumer trends in South Africa, sign up for Nugget, our bi-weekly newsletter. You can also follow Insights on LinkedIn and Twitter.

‍

Shein is a global phenomenon. Consumers are smitten with the Chinese retailer’s heady blend of hot fashion and low prices. From 2019 to 2021, revenue increased by 398% and although growth has slowed since then, the company has ambitiously predicted sales of $60 billion by 2025.  

Shein has a loyal South African fan base and its share of spending among 22seven users has increased significantly – we have shown previously that this market share is composed of a relatively small number of customers spending relatively large amounts, especially when compared to the average transaction values at competitors.

Shopping at Shein is not as simple as one might expect from the world’s largest online-only fashion company. Since all items are manufactured in China and shipped from there, delivery times are slow and customs fees come into play. Similarly, returning items is a headache.

But local consumers seem quite happy to overlook these issues. 22seven users are just as likely to shop at Shein as they are at well-known brands like Cape Union Mart or Zara.We wanted to find out why, by asking this question: How do consumers arrive at the decision to purchase from Shein?

‍The starting point is to assume that every consumer is on a journey. Decisions about where to spend money are informed by the user’s needs, wants and current circumstances. Such decisions are also informed by past transactions and the purchase and product experience of transacting with other retailers.

In the research note Mirror, mirror, we provided quantitative evidence that success in the apparel market is dictated by variety and choice, especially when attracting customers in higher income brackets. If anything Shein offers plenty of choice!

In this report, we seek to understand the purchase decisions that precede a Shein purchase: What experiences did the customer go through to make them decide to shop on a foreign platform with import duties, that takes two weeks to deliver? Also, which purchases did this customer forgo when they moved to Shein?

‍The beauty of Insights data is that we have sight of a user’s entire purchase history over time. For this analysis, we identified all 22seven users that transacted at Shein in 2022. There are 4,450 of them.

Next, we sequenced all of these users’ apparel transactions at more than 125 clothing and footwear retailers, ending up with each user’s purchase journey until the moment they decided to buy from Shein.

‍To make it easier to explain the methodology and findings, we’ll focus on users earning R25 – 40k pm. We’ll highlight where the findings differ in other income groups.

What does a Shein customer look like?

The chart below shows how 22seven users who shopped at Shein allocated their total apparel wallet during 2022, compared to those who didn’t shop at Shein.

The first insight from the data is that Shein shoppers are valuable apparel customers. Controlling for income, these customers spent R1,062 a month on apparel in 2022, more than 2x what non-Shein shoppers spent.

Next, these Shein customers spent 20% of everything they spent on clothing and footwear in 2022 at Shein. That’s a lot!

The split of the other 80% is shown on the top right pie chart, where TFG, MRP, Superbalist, Retailability and PEP make the top five by share.

The spending split by users who didn’t shop at Shein has similar proportions, but the difference (shown in the lowest chart) shows that Shein customers spend more on average at Zara, Superbalist, PnP Clothing and Pep, and less at Sportsman’s Warehouse and Cape Union Mart.

At the risk of stereotyping, this evidence suggests that Shein customers are more likely to be female. (Shein mostly sells women's clothing.)

Which retailer is the gateway?

Now let’s look at the sequence of purchases preceding a customer’s first Shein transaction. The example below depicts a 10-purchase journey for a random user in our sample. Seven apparel transactions before this user spent R1,578 at Shein, they spent R499 at a TFG store. The very last purchase before they clicked ‘buy’ on za.shein.com was R650 at Tekkie Town.

Does a user’s behaviour change in the lead-up to the first Shein purchase? Our hypothesis is that the proportion of spend for all users in the sample should be the same at each point leading up to the Shein order.

To test this, we aggregated all the user journeys and compared the spending share immediately before the first Shein order (the green block above) with transactions earlier in the users’ history (the red block).

We’re delighted you’ve read this far. But research is only valuable if it’s exclusive. To be fair to the retailers and investment teams who subscribe to Insights, we have to say goodbye before we get to the juicy bits. Contact us to find out how our research can help you.‍‍‍

Simon Anderssen | simon.a@22seven.com | +27 84 730 0309

‍

TFG has a portfolio of ~24 retail brands in South Africa, six of which have been acquired over the past three years through the purchase of Jet and Tapestry. We want to understand how the existing 18 brands have performed in recent years, so that we can better understand the ‘organic’ growth track record of the TFG South Africa division.

This research was inspired by the findings in Mirror, mirror, which analysed MRP’s lofty growth aspirations by unpacking consumer apparel spending by brand within income brackets. Two things were clear: the degree of fragmentation in the apparel market among higher-income earners (the largest market segment by value); and the importance of strong local brands to compete against the greater participation of foreign or global brands in this segment.

By better understanding TFG’s organic growth performance by brand, we believe analysts are better equipped to evaluate the retailers’ strategies and growth prospects.

To perform this analysis, we evaluated the spending of 22seven users by brand within four income brackets over the past five years. The charts below show the contribution of each brand to total 22seven user spending at all TFG’s existing brands over the period. The brands are organised in groupings that have similar scale.

A word on credit: Credit sales are important for TFG and account for c30% of annual sales. Fortunately, since Covid, most of TFG’s credit payments have shifted online, making it easier for us to distinguish credit payments from other purchases. That said, in-store account payments, EFTs and other credit idiosyncrasies do introduce a higher forecast error into such an analysis…

Still, based on the large samples of transactions for most brands, we believe the trends shown below are significant.

What you need to know

• @home has lost spending share within the TFG stable, without a corresponding adjustment to store numbers.
• TFG’s growth brands may need to grow ~20% pa for TFG Africa to achieve 5% store growth. Store numbers for many large brands appear to be saturated.
• Over the past five years, Archive, Sneaker Factory, The FIX, Exact, sportscene and Totalsports have gained share of total spending at TFG.
• There are signs of better store density for Foschini over the period.

‍

Observations:

• ‍‍Due South and Donna have closed their doors over the past two years. Donna has been absorbed into Foschini stores and Due South has shut down completely. In 2018, these brands together accounted for ~4.5% of user spending at TFG.
• Footwear brands Archive and Sneaker Factory have grown consistently in recent years, each increasing their contribution to total TFG spend by ~1%.
• Chart 2 highlights brands that are supported by different income segments. Note how higher-earning 22seven users spend more at Fabiani and G-Star Raw than lower-earning users – the green line is at the top and the red line is at the bottom in the first two panels. It’s the opposite for Exact and The FIX, where lower-earning users make up much more of total spend than higher-earning users.
• Fabiani’s share of user spending has increased modestly over the past five years.  G-Star Raw, Exact and The FIX have performed even better – all increased their spending share.
  ‍

• It’s interesting to compare TFG’s sports-focused brands, sportscene and Totalsports. Panels 1 and 2 in Chart 3 show how both have increased their share of total TFG spending. And, while both brands have a similar average share of spending, sportscene has a much higher dispersion from different income segments. Notably, lower-income users contribute more than higher-income uses. For Totalsports, each income segment accounts for a similar share of total spending, and the growth trends across income brackets are similar.
• Markham is a significant brand for TFG, but Chart 3 shows how its share of spending has trended lower over the last five years. This trend is consistent across income segments.
• Homeware brand @home is shown in Chart 4. It’s the brand that attracts the greatest share of spend, mostly from higher-income users. @home gained share in 2020 and 2021, but that share has since declined in each income segment in 2022. This likely reflects a normalisation of post-Covid nesting.
• According to our findings, Foschini is the largest standalone clothing brand in the TFG stable. Its share of total user spending has increased in the R25-70kpm income bracket.
• Of TFG’s two jewellery brands, users have allocated more of their total TFG spend over the last five years to the smaller Sterns brand than to larger American Swiss.

‍
Now, to store numbers…

TFG has a long history of financial information that allows us to chart store numbers across brands. This information, used in conjunction with the 22seven spending analysis, provides a good picture of overall brand-building performance.

The following charts show TFG’s store numbers by brand in South Africa over the last 17 years. Again, we’ve separated them into four sections for ease of reference.
‍

Observations:

• It’s very difficult for a brand to grow bigger than 350 stores in South Africa. Of TFG’s long-standing brands, only Markham achieved this – temporarily – in 2018.
• Chart 6 shows how store numbers for TFG’s largest brands increased rapidly in the decade leading up to 2018. Thereafter, store numbers for most brands declined and have since stabilised, with little growth in numbers in recent years.
• Totalsports and sportscene are the stand-out growth performers over the period. From ~100 stores 17 years ago, both brands had 310+ stores in 2022. sportscene’s growth seems likely to continue.
• Jet is big. With more than 400 stores, it has a much larger store base than any existing TFG brand. As we’ve suggested previously, the opportunity for TFG here is to drive sales density at existing Jet stores, rather than significantly expanding the brand’s store footprint.
• The mid-sized brands are shown in Chart 7. For most of them, store numbers peaked before 2020. Based on the trajectory of these brands over the last eight years, it’s likely that the opportunity for these brands in South Africa is saturated.
• TFG’s growth brands are featured in Chart 8. Sharp increases in store numbers for Relay Jeans and Sneaker Factory show management’s confidence in these brands becoming the future engines for growth. Other brands with increasing store numbers include RFO, Fabiani and @homelivingspace.
• Finally, Chart 8 is a reminder of brands from the past. 26 brands have traded under the TFG banner over the past decade (excluding Jet), and on our count there are six that are no longer trading today – a survival rate of 74%. Closures are to be expected, of course, but it’s a reminder that even seemingly established brands die. Due South reached 77 stores and Donna reached 100+ before failing. This information curbs our enthusiasm for the success of rapidly growing brands like Sneaker Factory and Relay Jeans…
  ‍

What does this mean for future growth? If the brands in Chart 6 and Chart 7 are relatively mature and their growth prospects are limited (with the exception of sportscene) Insights estimates that the remaining ‘growth’ brands need to expand their store numbers by 19-23% pa, for the overall store base to increase by 5% pa.

What have we learned?

The following chart summarises both data sets: 22seven user spending data and TFG’s reported store numbers.

• In the top-right quadrant, it shows that growth in user spending at brands like sportscene and Sneaker Factory is consistent with growth in store numbers.
• In the lower left quadrant, the decline in user spending is consistent with the brands that have shut down over the period.
• Spending at Foschini has increased, while its store numbers have decreased, suggesting positive increases in store density (sales per store).
• Of concern, user spending at @home has decline substantially, while the store base (@home plus @homelivingspaces) has declined only modestly. This indicates a shift in spending that is not yet reflected in the reallocation of TFG’s real estate.

Mr Price wants to become the most valuable retailer in Africa, which is ambitious. Today, it’s maybe the fifth biggest retailer in Africa with about one third of the market capitalisation of the largest. Achieving its objective will require some doing – and a lot of growth!

Having retreated from off-shore escapades in Australia and Poland, and showing tentative appetite for expansion in the rest of Africa, the implication of this statement is that growth must come from South Africa. But how?

Our understanding is that while value is most often determined by a company’s shareholders, management does have control over earnings. To this end, Mr Price articulates multiple actions that it is undertaking for organic and inorganic growth. Recent organic growth has been disappointing, however – something that we predicted as early as November – and total growth has benefited from three sizeable acquisitions over recent years.

A matrix of future opportunities has been a feature of Mr Price’s analyst presentations for a few years now. This matrix stratifies different product categories (Apparel, Homeware, Financial Services & Telecoms and E-commerce) into different product segments (Value, Aspirational, Premium and Luxury).  

In the most recent results presentation, Mr Price also showed a breakdown of consumer spending by LSM (Living Standard Measure – a widely used but flawed measure of income, now replaced by the SEM) that showed the significance of higher LSM groups to total expenditure in South Africa. Customers in LSM groups 8-10, earning more than cR15k pm, make up just 25% of households but 56% of expenditure.

Management commented that Mr Price had never ‘tackled’ the upper-income end of the market. The more subtle point is that future endeavours into this end of the market promise rich margins, which management expects might offset the potential margin dilution from recent acquisitions in the Value segment.

In other words, expanding from its current stronghold as a Value retailer into the higher-income end of the market is important, both for sales growth – given the absolute size of this market – and for sustaining Mr Price’s margin into the future.

How should investors evaluate this strategy? Is it reasonable?

What follows is, we believe, the most detailed retailer-level decomposition of the South African Apparel market, across income and product segments, available to investors or retailers.

We build on Mr Price’s segmentation of the apparel market into Value, Aspirational and Premium, based on 660,000 actual transactions by 45,000 22seven users at 140+ clothing and footwear retailers, split between four income brackets. The income bracket segmentation is important because South Africa’s notoriously wide income distribution means that consumers are not homogenous: What is Value to one consumer is Premium to another.

The objective of this report is to assist Mr Price shareholders in better understanding the context of the group’s starting position within the Apparel market; to better understand organic growth prospects; and to expose the opportunity set for inorganic growth.

In doing so, we highlight some hard truths about the structure of the South African Apparel market and conclude with a very short-list of potential acquisition targets in the upper-income segment. We identify these based on size, investability, and the likelihood of being a good-fit for Mr Price, based on preferences in their customers’ transaction behaviour.

We don’t know what Mr Price is incubating internally to catapult the growth of its existing brands, or what it might pay for an acquisition, but when the next deal is announced, don’t be surprised if it’s on this list…

The focus of this report is the Apparel market. Based on your feedback and interest, we can replicate this analysis for Home and other categories. Contact us and share your views.

Value to whom?

Every retailer positions its brand to influence consumer perception and to differentiate itself from competitors.

When retailers analyse a market, they often use a two-by-two matrix to visually represent how a retailer (subjectively) believes its brand offers something unique. Quality and Price are usually the variables, or in the clothing market, the level or type of ‘Fashion’ and Price.

Value, Aspirational and Premium are terms used to group retailer positioning on such a matrix. In a clothing market, for example, with Fashion increasing from left to right and Price increasing from bottom to top, Value retailers would be positioned in the lower left area. In that sense, Value retailers offer basic clothing with a few fashion features, at a reasonable price.

Aspirational retailers cluster closer to the middle of the matrix and sell fashion that most consumers want to wear but can’t afford to.

Premium retailers are at the top-right, selling high-fashion items that are only attainable to the few consumers who are prepared to pay a significant sum.

In a market where consumers are relatively homogenous, this type of analysis can be useful. But it requires more nuance in a market like South Africa, where we have a wide income spectrum and where price is a key variable in how a consumer perceives a brand. In other words, a consumer's income level will determine how they perceive a brand, which results in situations where – for example – one customer perceives a retailer as Value and another customer considers the same retailer as Premium.

Our goal is to estimate the market share of each retailer within four income brackets, and to stratify that share by segment (i.e. Value, Aspirational and Premium). We believe that such a granular understanding of where a retailer currently attracts its trade will reveal opportunities for expansion into new segments.

We start by decomposing the clothing and footwear market based on the contribution of different income brackets to total expenditure. We use the data provided by Mr Price for this.  

Next, we categorise 22seven users into income brackets aligned to the LSM categories in the Mr Price report, and we identify transactions at more than 140 clothing and footwear retailers.

We can now consider market share for each product segment. But how to segment the market into Value, Aspirational and Premium? These are subjective marketing terms and they largely have to do with price. In other words, the price that customers are willing to pay to shop at a certain retailer is an indication of that retailer’s positioning to those customers. Using that logic, we can analyse transaction values to determine which segment a retailer fits into.

To go into more detail about our method… We ranked each retailer by the median transaction value of all user transactions within each income segment, then we segmented the retailers within each income group using the following rule: The transaction values of Value retailers rank in the bottom 55%, Aspirational retailers rank between 55-80%, and Premium retailers rank in the top 20%.

If that’s a lot to take in, maybe the following chart will make it easier to visualise. Each dot is a retailer and corresponds to the median transaction value at that retailer by 22seven users in the specific income bracket. The retailers in the bottom half are seen as Value retailers to the users in that income bracket, the next 25% are Aspirational retailers, and the top 20% are Premium retailers.

(Note that that median transaction values appear to cluster at specific values – many dots at the same horizontal level. Our rule to segment retailers into Value, Aspirational and Premium might be subjective but it takes guidance from these thresholds or price levels.)

Some observations from this chart:

• The upper transaction values for each segment increases with income. In other words, what customers pay at some Value retailers in the highest income bracket is what customers earning R7.5–15k pm pay at a Premium retailer.
• The number of dots increase with income, meaning that higher-income customers shop at more retailers than lower-income customers. For example, customers earning R25k+ pm shop at 75 Value retailers, whereas customers earning R7–15k pm shop at only 26 Value retailers. This is a significant observation that will become more apparent when we estimate market value in the next step.
• The samples are large and the data is amazing, but it isn’t perfect. Insights tracks transactions at more than 140 clothing and footwear retailers, but there are likely to be more. We also ignore instances where there are fewer than 30 transactions at a retailer in an income bracket. Finally, we do not include Woolworths in this analysis because we’re unable to distinguish clothing transactions from food and the majority of its revenue is food.

Having categorised every retailer into a product segment for each income bracket, we then added up the value of spending by customers at each retailer. This allows us to estimate the contribution of each segment to spending in each bracket, and to disaggregate each segment by retailer. The graphic below shows how spending in each income bracket is allocated to Value, Aspirational and Premium retailers.

Some observations:

• The doughnuts are sized to mirror the contribution of each income bracket to total expentiture. This proportion is shown in the centre of the donut. For example customers earning more than R25k pm account for 31% of total spending, customers earning R7.5-15k pm account for 24% of spending and so on.
• In the two lower-income brackets, market share by product segment aligns with the rule used to categorise retailers: Value retailers c55%, Aspirational retailers c25% and Premium retailers 20%. This implies that the higher frequency of transactions at Value retailers offsets the lower average transaction value.
• In the two higher-income brackets, spending at Value retailers accounts for more than 62% of total spending. This has to do with the distribution of transaction values in the higher-income segments: Even after controlling for outliers, there are some retailers where users spend very large amounts, but the low frequency of purchases does not compensate for this high transaction value.

The next graphic shows the top five retailers by market segment and income bracket.

We’re delighted you’ve read this far. But investment research is only valuable if it’s exclusive. To be fair to the retailers and investment teams who subscribe to Insights, we have to say goodbye before we get to the juicy bits.

Get in touch to find out how our research can help you. simon.a@22seven.com

Black Friday is a big deal. Half of the 1,400 22seven users surveyed by Insights at the beginning of November indicated they were deliberately holding back purchases in the hope of a better price during a Black Friday promotion. A third of respondents knew what they wanted and were keeping an eye out for a better price, and another third were keeping financial powder dry for an irresistible deal.

Respondents spanned the income spectrum and were most likely to buy online. (See here for further details.) Respondents who were only planning to shop in-store were in the minority, which further highlights the significance of online/omni-channel options for most consumers. (The likelihood of respondents to an online survey being more digitally inclined is not lost on us, but the skew in the answer distribution to this question was still significant: only 10% of respondents across a wide range of incomes expected to only shop in-store.)

BankServ tells us that total spending through its network increased by 16.24% on Friday 25 November, compared to the year before. Since this is only a measure of spending on the day and does not include spending when the retailer’s point-of-sale and customer’s card issuer are the same, it’s not necessarily the most accurate reflection of retailer performance. What about spending on pre-Black Friday promotions, for example?

Yes, Black Friday is more than one day for retailers – it’s a month-long push for sales. Depending on the category, Black Friday festivities mean that spending during November accounts for 9-13% of total annual spend. (On average, a month should account for 8%.) November is even more significant for online retailers, where it counts for 12-13% of total annual spend by 22seven users. (Spending peaks for online retailers in November; December is the peak spending month for most other retailers.)

Considering that the whole of November is a crucial trading month for retailers, who won Black Friday this year? Which retailer outperformed its competitors, and which improved on its own previous performance?

To answer these questions, we contrasted user spending at a wide range of retailers from June to November, in 2020, 2021 and 2022. We therefore judged a retailer’s Black Friday performance in terms of the monthly change in spending over the six months leading up to November. Doing so reveals the trajectory of growth over the period.

The chart below shows the category-level results. Spending on furniture increased the most in November – as it has done in previous years – and by more than it increased in 2021. Grocery spending declined marginally in November this year compared to October. (Last year, spending increased in November compared to October.)

Spending on apparel and general merchandise increased in November 2022, but at a slightly slower rate than in 2021. Spending at major DIY retailers declined in November 2022 – a significantly weaker performance than the year before.

Category growth

So, if shoppers were mostly after homeware, apparel, and general goods on Black Friday (the latter category includes major electronics and online retailers like Takealot) at the expense of groceries and DIY products, which retailers outperformed and which underperformed?

The answers are in the charts below. Let’s begin with the grocers…

Grocery

The change in spending in November 2022 at the main grocery retailers was lower than in 2021. Spending at PnP increased at a faster rate in November 2022 compared to October, whereas the rate of growth declined at Checkers, Spar and Woolworths. Woolworths delivered the strongest cumulative trading performance over the past two months. Spar experienced the weakest growth in spending during November.

We note that sales growth dipped sharply for all grocery retailers in September 2022 when loadshedding intensified.

Apparel

We plot these over two charts with MRP on the left of both, to help calibrate performance between the two.

CottonOn, TFG and Superbalist outperformed this group of retailers during November, each with a similar shape of monthly growth in 2022 as in 2021. For the first time in three years, spending at MRP declined in November compared to October. Given the size of MRP, this suggests that it underperformed in November and over Black Friday.

Albeit at a lower absolute rate of change, the data suggests that Truworths performed well in November 2022 compared to its own history, with spending increasing at the highest rate in November 2022 compared to the two years before.

Spending at Retailability – mostly Edgars – increased for two consecutive months, whereas spending at H&M slowed. Zara appears to have had a challenging six months after a strong start to the year.

The data suggests that spending at Pep & Ackermans did not increase meaningfully in November in prior years, and this trend continued in 2022.

General Merchandise

This category includes many products that are sought after in Black Friday promotions, like home appliances and electronics. Among the retailers we track, spending at Incredible Connection increased the most in November, and at a faster rate than in 2021. We note a much weaker October 2022 than October 2021, however, which may dilute November’s performance.

As to be expected, spending at Takealot increased sharply in November and at a similar rate and trajectory of monthly growth as in previous years.

We include OneDayOnly in this group, too, because November accounts for its highest proportion of total spending than any other retailer included in this analysis: c14%. It had a much stronger performance in November than in previous years.

Spending at Makro increased in November, but not to the same degree as in previous years.

Furniture

The furniture segment delivered strong growth in spending in November and a better 2022 result than in 2021. Insights data shows that @home delivered the strongest growth in November and better performance than in 2022. Tapestry Brands followed close behind, suggesting a good overall performance for TFG’s furniture business.

JD Group brands – mostly Rochester, Russells and Bradlows – performed well but at a slower pace in November than in previous years.

Spending at MRP Home stores performed better than at MRP brand stores.

And the winner is…

This analysis has evaluated retailers’ Black Friday performance based on the month-on-month change in spending in November – compared to competitors’ growth and compared to 2021 and 2020 growth for the retailer in question.

Overall, it seems that Black Friday triggers the highest increase in spending in the Apparel, Furniture and General Merchandise sectors. That said, monthly growth in spending in November 2022 was lower for all categories than in 2021, with spending declining among grocery and DIY retailers.

Of the major listed retailers analysed, the data indicates that PnP, Woolworths, TFG and Truworths outperformed in November 2022.

In ‘No stores necessary’, 22seven Insights showed how spending in the apparel sector is shifting towards online-only retailers like Shein and Superbalist, threatening the investable, mostly store-based part of the market. The customers leading the online-only change are young and they’ll carry their preferences forward as their spending power increases, and they’re high-value customers. After controlling for differences in income, we showed how customers who shop online spend more on clothing than their contemporaries who only shop in-store.

But what does this mean for profitability? After all, don’t online shoppers just return everything they order?

We’re going to attempt to answer this question by contrasting the transaction economics of in-store and online purchases, to quantify which channel is more profitable for retailers.

One of the truisms in retail is that every year a retailer must start from scratch. Sales – the revenue kind – are not guaranteed to recur in the next period. The base unit of sales is the transaction. Gross profit is earned on each transaction. Aggregated over many transactions, gross profit is used to pay for the cost of operations, people and taxes. Anything left over is available to the providers of capital.

Analysing profitability at the transaction level therefore seems like a good approach to contrast overall profitability between online and in-store sales. In most cases, the data to do this comparison simply doesn’t exist, but with the 22seven user base at our disposal, we were able to aggregate 280,000 actual transactions from real consumers to provide a quantitative assessment.

For anyone concerned about the future earnings of clothing retailers, this analysis provides a robust foundation to evaluate the assumptions in your forecasts. For a retailer, this note might give you an idea of how you stack up against your competitors.

Let’s explore the data…

Profit per transaction

This analysis seeks to contrast the gross profit of the marginal transaction. That is, when the store’s doors or website ‘open’ at the start of the year, what is the profit earned from the first online or in-store purchase.

In South Africa, most store-based clothing retailers distribute items that are purchased online from a store that is close to the customer. Similarly, sales from online-only retailers are fulfilled by staffed distribution centres.

Therefore, for the marginal order, the cost of the store or distribution infrastructure, and any personnel to fulfil the order, is fixed.

The key difference between the marginal gross profit of an online purchase versus an in-store purchase comes down to the transaction value, the gross profit margin, any delivery costs and the cost of returns. We’ll unpack each of these.‍

Transaction values and return frequencies

The value of online purchases is generally higher than in-store purchases because customers buy additional items and return clothes that don’t fit or clothes that aren’t as flattering in real life as they are worn by the model on the website.

We wanted to find out how much more people are spending online than in-store, and how often online orders are returned. To do so, we identified purchases and returns from 15 retailers:

• Five omnichannel retailers where Insights could distinguish online and in-store purchases: Mr Price, CottonOn, PnP Clothing, Sportsmans Warehouse and Cape Union Mart.
• Six retailers where Insights was not able to separate store purchases from online purchases, either because online transactions are channelled through a central payment gateway (Markhams and Sportscene) or because online is not available or is too small to be representative (Ackermans, Edgars, Zara and H&M). Although there may be a small contribution from online purchases in this group, we think it’s minor, so we refer to this category as in-store only.
• Three online-only retailers: Superbalist, Shein and Zando.
• For the sake of comparison, we also included the largest general merchandise online retailer, Takealot.

To estimate returns for each retailer, we compared the number and the value of returns (credits) to the number and value of purchases (debits) over six months.

Returns for omnichannel retailers are more difficult because items ordered online are often returned in-store. For these retailers, we identified all online purchases and restricted credits to the users who had previously purchased online. We used a similar method for in-store purchases: excluding users who had recently ordered online.

The average purchase and return value for each retailer is shown in the charts below. As an example, the average order value for Mr Price online is R584, and the average return value is R428. The average order value for Shein is R1,566, with a return value of R979. ‍

Average transaction values
‍

• The average online transaction value across all retailers (R1,107) is 55% higher than the average in-store transaction value (R713).
• It follows that the return value is also higher for orders that originate online (22%).
• For omnichannel retailers, online transaction values are 20-30% higher than in-store transaction values.
• The online order and return values for Cape Union Mart and Sportsmans Warehouse are significantly higher than their in-store values. This may be because they sell a wider range of goods, such as sports watches and technical equipment, than their peers.
• The average transaction values at online-only apparel retailers comfortably exceed the transaction values of their store-based competitors. Shein has the highest transaction value, on average.

Next question: How often are items returned?

Frequency of returns
‍

• On average, 15% of orders that originate online are returned, compared to only 3% of in-store purchases.
• 26% of Superbalist orders were returned during the six months ending August 2022. Since returns are typically of a lower value than the original purchase, returns accounted for 19% of total purchase values. Similarly, 24% of Zando orders were returned, with an overall value of 17% of all purchases.
• The frequency of returns for online orders at Cape Union Mart and Sportsmans Warehouse are higher than at other store-based retailers. We think this may be attributed to a variation in products ordered online via these retailers. See the comment above regarding transaction values.
• Albeit not as high as Sportsmans Warehouse and Cape Union Mart, online return frequencies are significantly higher for Mr Price and CottonOn compared to in-store return frequencies.
• Takealot is not a clothing retailer, but it’s the largest online retailer among 22seven users. c5% of Takealot purchases are returned.
• The frequency of Shein returns is low – only 3%. This is arguably because of the challenge of returning items. Although refunds are offered within a 30-day period, customers must package and ship their returns at their own expense.

‍

Free delivery: common but not standard

To estimate the delivery cost for the average online order, we compared the fees that retailers charge for delivery. In nearly all instances, delivery is free if the value of the order exceeds a certain threshold.

For most retailers, the cost of delivery is within a narrow range of R50-60 per order for orders less than R500-600. Shein has the highest delivery fee and the highest minimum order value, which most likely reflects the higher cost of distributing each order from China via air freight. (Note that this delivery fee excludes duties and other fees paid separately to third-party couriers on Shein orders because these do not impact the retailer’s profitability.)

We find it interesting that Mr Price is the only retailer that Insights could identify that does not offer free delivery. Customers are charged at least R50 for any online purchase, irrespective of the purchase value. As we show below, the recovery of this delivery cost from the customer is crucial to the economics of Mr Price’s online offer.

A further observation: In most examples, the average transaction value is 50%+ above the minimum required to qualify for free delivery. The outlier in our sample is PnP Clothing, where the average transaction value is only slightly higher (2%) than the R700 threshold to qualify for free delivery. This suggests that for most retailers, the free delivery threshold is not a meaningful barrier for most orders.

Delivery fees and minimum order thresholds
‍‍‍

Rands or margins?‍

Combining these findings, we are nearly able to estimate the gross profit (in rands) and the gross margin (in percent) for orders through each channel for each retailer.

The only assumption we make in this analysis is the average gross margin earned on the merchandise sold. Gross margin is influenced by the type of merchandise sold (branded vs in-house, footwear vs clothing, fitness watches vs accessories etc.) and it will vary for each retailer. To enable our comparison, we assumed a standard gross margin of 42% for all retailers. This level is informed by the last reported gross margins for Mr Price (42.1%) and TFG Africa (43.1%). Readers are welcome to request the spreadsheet to express their own views in the calculations.

In some cases, we have good reason to expect gross margins to be higher. For instance, H&M and Zara (parent company: Inditex) report significantly higher gross margins internationally, but high import duties in South Africa may mean these margins are lower locally. For this reason, we show a range for H&M and Zara.

In addition, although it’s a private company, some argue that the gross margin on Shein’s products could be even higher than other fast fashion retailers. For the purpose of this analysis, we show a range for Shein based on a low estimate (42% like Mr Price) and a high estimate (55% like H&M and Inditex).

The gross profit on an online order is calculated as follows:

(Purchase value * GP% - delivery cost) – %purchased returned*(Return value*GP% - delivery cost)

Gross margin = gross profit / transaction value

Let’s start with gross margin. For the five omnichannel retailers where we can compare in-store and online purchases, the cost of delivery and returns reduces gross margin by 10% percentage points on average.

For the three online-only retailers, the impact of delivery costs and returns is a reduction of c14% on the assumed gross margin of 42%.

Gross profit, shown in the chart below, is a different story. The three most striking observations are as follows:

The gross profit earned (in rands) on online purchases at the four omnichannel retailers is not meaningfully lower than the gross profit earned on the average in-store purchase.

Superbalist earns more gross profit per transaction than store-based clothing retailers like Edgars, Mr Price, CottonOn and Ackermans.

Shein generates the highest profit per transaction – as much as 2.8-3.7 times the average gross profit of competitors like Ackermans, Mr Price, CottonOn, Edgars and PnP Clothing.

Gross profit
‍

Why is this important?

Executives in charge of store-based retail chains often talk about getting the economics of online to ‘work’. Our findings across hundreds of thousands of transactions show that the gross profit earned for an online transaction is comparable to that earned for an in-store transaction. However, the margin is typically lower, sometimes significantly.

For that reason, online doesn’t complement the income statement of an established store-based retailer because a growing online contribution will dilute the group’s overall gross margin. (Here’s a tip for analysts assessing these businesses: If you expect online revenue share to grow, account for the diluting impact on gross margins.)

We have previously shown that Superbalist is a large retailer by South African standards. We estimate cR2.5bn in annual GMV. In this analysis, we’ve provided quantitative evidence that its gross profit is higher, on average, than many mainstream store-based clothing retailers. Since Superbalist is part of the Takealot group, it’s impossible to definitively assess whether the business is profitable. Takealot itself is not (yet) but when you compare Superbalist to its clothing competitors – many of them highly profitable listed businesses – the evidence suggests that it might well be.

Insights has alerted readers to the rapid gains that Shein has made in South Africa since its launch two years ago. We have shown that these gains are a function of enormous growth in the number of users experimenting with the service, multiplied by outsized transaction baskets – as much as 4x the average transaction value than store-based competitors. Put another way, users shopping at Shein are effectively shopping once, whereas they might have shopped at four other local retailers. In this note, we’ve provided additional evidence that Shein earns 2.5-3.8 times more per transaction than many large, listed retailers. Shein’s impact on the South African market’s revenue share has been sudden and disruptive, and its impact on profit share may be even more so…

Please get in touch if you would like any further info.