James Taylor
Today, we'll be looking at data from consumers in previous recessions, in the area called consumers in crisis, and then we'll move on to companies in crisis. We'll look at some data from companies in previous recessions. We'll then move on to how to thrive, we'll look at the different mindsets that exist in retail, and specifically the opportunity that exists online. And finally, we'll close with three rules: how to adapt and how to execute progressive change. Retail economics. Spending is down. If you look at this data, we can see that over the past four years, purchasing groceries online has grown and grown and grown. Habits form and stick during periods of uncertainty, smart businesses are creating or deepening footholds now. Companies in crisis, this is where we look at data from companies in previous recessions. We see what's worked and what hasn't. Harvard Business Review put out some research that looked at data from 4,700 companies across three recessions, recessions of 1990 and 2000, and the results are something that we really need to look at and listen to. 17% of companies do not survive. 80% of companies do not regain pre-recession growth rates for sales and profits in the three years following recession. 40% of companies do not regain their absolute pre-recession sales and profits within three years. That means they don't get back to where they were before the recession within the three years following the recession. Finally, just 9% of companies flourish after slide outs by outperforming their rivals. So, how does marketing and recessions impact profits during the recovery? Well, again, we probably will have seen lots of articles that have said spending more on marketing lead to better results for companies. And during the GFC we saw that companies that invested in marketing did see a 4.3% increase in profits during the recovery as compared to companies that did not, who saw a contraction of 0.8%. And it's not just limited to the GFC, this data. We're also gonna look here at the 1920 recession, the Great Depression, and it's clear here that even companies that increased advertising spend had a significant outperformance in sales during the recovery. Those that decreased advertising saw a drop off, but is spending more on marketing a solution? Is cutting costs a solution? Well, from the Harvard Business Review data we know that 79% of firms that cut costs faster and deeper than rivals do not perform better coming out of a recession. And we know that 74% of businesses that invest more than their rivals do not become market leaders coming out of recession. Finally, 85% of market leading companies before the recession will not stay at the top. So Harvard Business Review splits out the companies in the dataset into all types. They look at prevention focused companies who simply cut costs. They look at promotion focused companies who increase spending and investment in marketing and R&D. Then they look at pragmatic companies who do a mix of both. They increase spending on advertising and they cut costs. And finally they look at a cohort of companies that they call progressive. Progressive companies find the optimal or intelligent balance of optimising between increasing marketing spend in the right areas, increasing R&D spending in the right areas, and reducing costs in the right areas. And just to give you an idea about the difference, promotion focused companies do perform better coming out of recession than companies that don't increase spending. Promotion focused companies see an 8% growth in post-recession sales and a 6% growth in post-recession earnings, but it's not as good as it could be. We look at the progressive cohort of companies that do an optimal mix of increasing spending and decreasing spending, and we actually see a 13% increase in post-recession sales growth and a 12% growth in post-recession earnings. So let's take a look at a real world example. During the 2000 recession, Office Depot and Staples both behaved slightly differently. Office Depot cut workforce by 6%. Staples increased their workforce by 10%. Office Depot couldn't reduce op-ex. Staples contained op-ex and closed down underperforming facilities. Office Depot released an incentive plan to boost sales. Staples adapted, they introduced new high-tech product categories and services. So slightly different behaviour, but what were the results? Well, Office Depot's sales growth fell from 19% before the recession to just 8% afterwards. Staples sales growth of 13% after the recession was a whole 5% higher than Office Depot's. Office Depot saw an overall revenue increase of 50% between 1997 and 2003, as compared to Staples who saw a 100% revenue increase between 1997 and 2003. And to boot, Staples was 30% more profitable than Office Depot in the three years following recession. This is more data from Harvard Business Review, back to McKinsey, how to win in a recession while focused on customer experience. So address customer needs to prepare for the future. What can you do to address new customer needs? You can optimise your stock on hand. You can get your inventory allocation right and you can ace digital product discovery. You need to reimagine the world. You do need to enact some cost cutting. You need to migrate consumers to digital channels. You can save money with this and boost satisfaction within your consumers. Reassess the utility of physical assets. There's long been a narrative about the changing role of physical assets and retail. Now is the time to enact these changes. Finally, build agile capabilities, acquire market share, adapt business models to accelerate growth after the crisis. And invest in machine learning and automation. So how to thrive, how to acquire market share? There are two mindsets in retail today. One mindset is one of scarcity, the other mindset is one of abundance. Scarcity mindsets are reflected by the likes of Sears and those victims of the retail apocalypse pre-COVID. The winners of the abundance mindset are the likes of Amazon, Netflix, Shopify, Spotify. Scarcity mindsets look at the rental cost of shelf space. Abundance mindsets look at the low inventory cost through warehousing and third parties, digital distribution. Scarcity mindsets are hindered by limited space. Abundance mindsets may as well have near infinite space. Scarcity retailing has to focus on best sellers. Abundance retailers can sell everything. Scarcity mindsets follow long hierarchical decision-making and ROI forecasting. Abundance mindsets permit experimentation and learning. Scarcity mindsets are top down hierarchies. Abundance mindsets are bottom up. I don't mean from the most junior people in the marketing team, I mean from the customer. They are data driven. They react and they adapt based on changing consumer behaviours. Physical retail channels were far from perfect. They were dependent on often opinion led inventory. They had to stock lowest common denominator items. There were assumptions around popular taste, and typically individual stores, at least, would often be subject to poor supply and demand matching. Physical retail is also relatively high cost for relatively low return. Store rent costs money in high footfall areas, and the cost of serving walk-ins is high. Stocked items have to earn their keep, they're occupying shelf real estate that costs money. And finally, finite shelf real estate means limited ranges. There's only so much space that you can actually fill with items. So this has created a dependency in physical retail on reliance on best sellers. And there is nothing wrong with best sellers. Everyone's taste departs from the mainstream somewhere. The more we explore alternatives, the more we are drawn to them. UBS published some analysis last week, they expect that 100,000 stores will close across the US in the next five years. They expect consolidation and the bigger getting bigger. Walmart, Costco and Target could be among the last left standing. The pandemic will accelerate a trend that was already in motion. The fundamental reason for stores closing down in the future will be the rise of eCommerce. UBS expects that many people will not go back to stores, or at least limit their visits. We think people will go back to stores, but we do think that they will limit their visits. Now, this isn't only because of new habits put in place during the pandemic. So what are the benefits of selling online? There are cost savings, you get your out of town warehouses, your distribution centres, you don't have cost of walk-ins. Having an unlimited catalogue improves your odds of converting someone. You're able to serve more niches, and you're also able to provide a better customer experience. Now, I realise that's a sweeping statement, and of course there is still a role for physical retail stores to provide a great customer experience. But the benefits of online retailing are search, reviews, recommendations and consistent and intuitive convenience, being able to buy items anywhere, anytime, and to find those items. The long tail profit opportunity is also greater. The profit threshold for physical retail is high. You have stores that cost money in expensive places that are limited in the amount of space they have for products. For retailers that have a hybrid model that sell online and in store, that profit threshold drops. And for retailers that are digital only, that profit threshold is even lower. So tiered pricing strategies, that is being able to price items differently as you go further down the long tail, and recommendation engines can drive customers to products they may not have found otherwise. So let's look at this in more detail. How to adapt: three rules for executing progressive change for futureproof commerce. Your long term market potential is twice the size of your best seller market. The three rules are: number one, make everything available. Number two, get smart on pricing. And number three, help me find it. Make everything available. You can have more style variance, but you should still be selective. You can use data to inform what you should be retailing. That's trend analysis. It's identifying what people are looking at down to an attribute level. You can use data to inform stock depth and assortment, size depth. On smart pricing, you need to forget channel conflict. Online should be cheaper. People buy more at lower prices and long tail items actually cost less, but it doesn't have to be channel conflict because of course many of the items you stock online don't actually exist in store. Finally, help me find it. Do not become a mess of bad products. Lead with best sellers from which further exploration can begin. Guide consumers based on their likes and dislikes. But why do people shop online during a crisis? During economic recession, the number one reason is purchase confidence and comparison. Number two reason is best deal assurance. Number three reason is shopping for discount codes. Overall, people want to save money. So let's dive into these rules in greater detail. Rule number one, make everything available. Item availability, stock availability as well. 87% of shopping journeys now start online. This statistic is pre-COVID. It's of course going to be higher right now, but what are you doing with all that demand data? Consumers want to buy, receive, and return goods anywhere. If you are not using page views, searches, as well as purchases online to indicate demand for different items in different parts of the country, to inform your order management systems, your inventory allocation systems, your stock planning systems, and any in-season reallocation decisions, then you are missing an opportunity and you are adding further under-performance to your physical assets. You should be using your online data to optimise in-store. Just to summarise, you can use that online demand data to optimise online journeys, but you can also use it to optimise in-store journeys. You can use it, which I think is very relevant post-COVID, for online to in-store journeys. In fact, buy online pickup in-store has increased 62% in the US since COVID. You're seeing even things like curbside pickup, drive throughs. Retailers are adapting to utilise their physical assets in a way that complies with consumer demands today. But buy online and pickup in store is not new. In the US, 67% of shoppers have used it in the past six months, that was six months prior to COVID. 88% of retailers are actually able to track incremental in-store purchases from BOPIS customers. Plus retailers can save on shipping costs and cut down on fulfillment times through inventory allocation optimisation. Whether you are allowing customers to pick up in store or whether you are shipping from store, you need to make sure you have the right items available in that store. What happens if you don't have the items in the store? Well, again, this is pre-COVID data, but 24 to 30 billion dollars a year of Amazon's retail sales in the USA are attributed to consumers who first tried to buy items in store, but the retailer was out of stock. That's 21 to 24% of Amazon sales stemming from challenges in inventory management. You have a wealth of online data, more so now than ever. You're able to localise that demand data. You're able to break down to a postcode level who is looking at what, and you can use that for inventory allocation optimisation. You can use it to move items from distribution centres to different stores, between stores, from stores back to distribution centres. Rule number two, get smart on pricing. How to increase sales, cash and earnings with intelligent and dynamic pricing. So this isn't all about discounting, this is actually really starting to think on a SKU level about how pricing can help you sell more items. Pricing, of course, you have your recommended retail price at the top, at the bottom you have your margin that you're able to shed on any given item to turn a profit. But you also have what competitors are pricing at, and this can give you an operating range to be competitive. And prices can decline the further down the tail you go. That is to say that you can pull customers down the long tail with lower prices. Since long tail items actually cost less, you can charge less for them. Remember, even if they come to the site looking for a best seller, do not underestimate the down-sell opportunity. And to reiterate, since most long tail items aren't actually available in store, your channel conflict actually diminishes. So what do we do? We can develop an AI-based dynamic pricing model. This can start with a long tail module, which can see which items are less popular and which items might be competitive with your best sellers. We can have a competitive response module, which is where price adjustments can either be automated or recommended based on competitor prices through competitor website price scraping and price crawling. And finally, your elasticity module, which calculates how much a product's price affects demand and price perception. So you can have dynamic prices on qualifying items, or you can use SKU level incentives, not blanket discounts. Blanket discounts are a blunt object. Moving things en masse into a 25 to 70% markdown in clearance is not the end goal of optimising your pricing. You can get down to a SKU level. Finally, number three, help me find it. Great long tail businesses can guide consumers following their likes and dislikes. If you like that, then you'll like this. Just as lower prices can entice customers down the long tail, recommendation engines drive them to products they may not have found otherwise. And this is key. The more items that you have, the less items will be found. Discovery becomes a problem that can be solved through recommendation systems. Recommendation engines are capable of finding the items in that long tail that are relevant to each customer. Just to throw some data at you from one of the Particular Audience clients that we're working with. We've just done some long tail analysis on their SKU base. In the control group, so customers that don't see recommendations, just 16.8% of SKUs make up 80% of sales. But customers that do see Particular Audience recommendations, actually 33.6% of SKUs make up 80% of sales. And to boot, we've seen a 23.4% increase in sales overall. So the opportunity in your long tail of items is huge. Help people to find them, price them effectively and make them available. The overlap where customers looked at both items in a session is something that we can put a score on, it's something that we can calculate. So we're able to calculate, from the wisdom of all of your customers, how similar two items are to one another, or how complementary they are, using purchase data. But what if you don't have any behavioural data on an item? Well, this is where we can use things like computer vision or natural language processing to understand the similarity between items. Men's trekking gore-tex boot, men's boot hiking leather sole support, men's walking ankle support boot. There's some similarity in there as well. And again, we can put a value on that, and then we can look at each different customer. It's not just the item they're looking at now and what's similar. It's understanding what have they looked at recently, multiple items, and what are the relevant or related items to those. And that creates almost a universe for each individual customer. Coming back to that point I made earlier, if you have a million customers, you need a million websites. And this sort of machine learning technology is capable of merchandising your site for every different customer that comes to it. Finally, you've got this amazing relevance that you can have, but you can also meet specific business goals. You want the journey to be relevant, but you might want to prioritise high margin items. You might want to boost overstocked items. If you've got too much of something, of course you want to merchandise it more often. And finally, you may even be charging brands to increase the exposure of their items on your website in the form of sponsored slots. Key takeaways: the opportunity is online, so don't worry. We can still make good use of physical retail assets using online data, and to boot, we can improve the experience online. So number one, make everything available. Number two, get smart on pricing. And number three, help me find it. Remember these three rules. Increase your chances of conversion and you will acquire market share. Tiered pricing strategies, recommendation engines and inventory optimisation are going to be necessary to thrive in a post-COVID world. Meet customer needs to filter your investment decisions, reduce spending on underperforming assets, increase spending on the right R&D and marketing. And I think to mention on that R&D piece, it's not a bunch of siloed digital tools that work. You need to start thinking about all of your data in a loop. You need to start thinking of your data as an ecosystem where technologies such as ours can actually work together using the data to achieve each of the different things in the three rules. So, quick view on Particular Audience. We're obviously specialists in search and merchandising, specifically in collaborative filtering, which is the behavioural led recommendations, computer vision and natural language processing. We also permit dynamic pricing. This can be within margin and related to your actual inventory of different items. We localise data, liquidate unwanted stock and reallocate stock as it's required in different places around the country. And using computer vision and natural language processing, we can conduct trend analysis. We can help guide your design teams and your buying teams on not just which items are selling, but specifically the attributes and trends on those items that are popular. And it doesn't have to be a big piece of work. It might sound like this huge overhaul of your business, but in fact it's actually very simple. Step one, you just need to track behaviour. You need to get a tracking tag on your website that is objectively picking up the right information. You need to process item attributes, this can be done with the tag, but also with product feeds. Your Google Shopping feed, as it stands, is already good enough to start this process. You need to ingest inventory data, localise it if it's applicable, if you've got lots of stores or lots of distribution centres. Then you can start layering contextual information on your website: similar items, dynamic pricing, SKU level incentives. Then you can meet business goals. You can boost sponsored items, you can boost overstocked items, you can boost high margin items. And finally, you can really architect a leading inventory intelligence solution. You can set up key reporting and action notifications to make sure that you have the right items in the right place at the right time. Cool, that's it from me. Thanks everybody for your time, and feel free to reach out to me directly at james@particularaudience.com if you want to pick up on any of the themes that we've gone through today and figure out how they might apply to your business specifically.