What’s in store for retail analytics?

It’s back to school in September, and we’re all looking ahead to 2020.

At OC&C we’re increasingly deploying our big data analytics capability in Retail. As we support retailers with analytics, we’re looking at what’s new and what’s to expect in the next 12 months.

Here’s the 10 things on my mind for Retail analytics in 2020.

1. Right-sized estates, formats and locations

OC&C’s detailed analysis of drive time maps in our store location tool indicates that you can cover 90% of the UK’s population with just 70 stores in two basic formats – small and large.

With the significant decline in UK’s high streets (and US malls), this is nothing short of a revelation. For some specific product categories (e.g. shoes, kitchenware, crafts, etc.), there is no need to have more than this number of stores.

Using spatial analytics, drive time data, TfL’s London commuter patterns by tube station, competitor locations, e-commerce’s hot postcodes, and “points of interest” databases, we can effectively optimise store formats and locations in any product category.

Store location and format decisions now need to be accurately informed by e-commerce purchase data now. Some stores now are showrooms for e-comm, others are transaction stores with specific SKU assortments. We’ve also supported retail clients with calculating the percentage of catalogue, e-comm, and store customers that overlap.

In tough times, retailers are not so much shrinking as right-sizing their estate. Let’s think of UK supermarket Sainsburys in the news this week for planning to shut up to 15 supermarkets in the wrong locations, along with 40 of its convenience stores, such as Sainsbury Locals, and 70 branches of Argos. However, Sainsbury simultaneously plan to launch 10 new supermarkets, 110 new convenience stores and place 80 Argos branches in other Sainsbury stores as SiSs. So, a complex mosaic of right formats in the right locations, with as many openings as closures.

We have been working with many retail clients exploring the wider mix of channels and how they work together; such as wholesale, store-in-a-store, kiosks, concessions in other stores and e-commerce. Analytics is incredibly useful looking at 2020 in this context to design effective store formats and location strategies, taking account of stores, e-comm, and catalogues working together.

2. Integrate, but differentiate the focus of, Bricks vs Online

I have a personal dislike of the word “omnichannel” as it implies all channels are the same. Actually, the opposite is self-evidently true. As is so often true in life, it’s better to embrace differences.

An effective (what most people call) omnichannel strategy requires seamlessness across the customer journey and a differentiated experience in each channel, that makes the most of that channel’s attributes. It also requires retailers to create and transition superior experiences across channels, which is a challenging task.

Additionally, each channel should be able to positively influence the customer journey – for example, how easy it is to click and collect in-store or trial items in a store and then finding them easily online for ordering.

Retail analytics is increasingly going to play a critical role in designing and delivering superior omnichannel experiences. It will encompass the whole customer journey and will help retailers in managing its fragmented nature.

Right from optimising CRM initiatives to digital marketing campaigns, the areas of application are wide and diverse.

Analytical 2020 prowess will also enable retailers to identify strengths and weaknesses across the journey and as it shifts between online and offline channels.

3. Higher prices are coming

Many threats point to higher prices in 2020, but let’s use analytics to harness this as an opportunity.

Brexit, shortage of labour, marketing costs all point to rising costs. Truth is that so much Retail demand has been propped up by massive PPC and brand marketing spend and simultaneous price cutting! None of this is sustainable, some increases in value have now got be intrinsic, i.e.: realise price increases, but from what consumers will pay more for.

Amazon’s biggest impact on traditional retailing goes beyond the convenience factor and is dictated by its threat of lower prices. The recent discovery(?) of its algorithm supposedly “favouring” Amazon own-brand products over that of branded competitors in search queries and recommendation results is a serious concerns for sellers on the platform. The challenges for retailers in product categories that are easily “commoditised” are even bigger.

High price points in categories that don’t have the premium or luxury equity is non sustainable, regardless of how much PPC you pump in.

More PPC spend, pumped up brand marketing, and price cutting initiatives have created a lot of artificial demand.

Consequently, higher prices where the price-value equation is not clear to consumers, is a barrier to growth. As this is not sustainable, some of the increases in value now needs to be from taking price, based on actually figuring out what consumers will pay more for.

The role of retail analytics becomes important, as pricing becomes a fluid dynamic dependent on interactions with consumers. What do consumers really value, we’ll identify how much they’ll pay, and set assortment and price ladders accordingly. This isn’t a trick, it’s creating intrinsic value for customers and so real price gains for the retailer.

As value perceptions shift, pricing will become highly personalised. The emergence of “subscription models” is an example of how retailers are customising product offerings and pricing, compared to having static price tiers in the past.

4. E-commerce will reach an equilibrium point

I may have called this out too soon in the past but: 1. the rate of E-comm share of retail is slowing down and so will at some point stabilise, and 2. Any individual category has a natural equilibrium level that makes sense (depending on factors like the importance of touching the product, in-store assisted sales etc). Books, for example, reached an equilibrium early.

According to the ONS, e-commerce had a share of 18.2% of total UK retail sales in August 2019. For the first eight months of 2019, this share has been an average of 18% and stable. Surely not peak E-comm already?

As retailers strive towards increasing the efficiency of their operations, analytics will be a key enabler in driving this change. A combination of decline in high street retail operations and the stabilisation of e-commerce growth will drive retailers to adopt new and lean business models.

Analytics will play a huge role in this and its impact will encompass store location to price optimisation strategies.

5. Recognising the importance of your ABCs

In an age of intrinsic value creation and shifting dynamics of consumer value, retailers need to pay close attention to realmargin calculations. By attention I don’t mean doing it consistently but doing it accurately.

Taking into account Activity Based Costs (ABCs) for margin calculations is critical, because for example low margin SKUs can cannibalise high margin SKUs (a factor for consideration in SKU optimisation).As a general rule of thumb, the worst 33% of SKUs add no incremental sales and the worst 25% of SKUs destroy margin.

Or for that matter true margins calculated net of wastage, supplier promo funding vs over-riders and bonuses, distribution costs, payment terms and working capital implications, etc…

When we work with retail clients, we look at SKU mix, pricing and promotions as part of a single holistic system that needs to be optimised, including supplier funding, your labour costs, and so on… The objective is to improve true margins across multiple categories, which is achieved by a combination of three aspects – optimised / reduced SKU mix, shift from liquidation driven to strategic promotion and pricing that reflects the right value proposition.

Analytics has a big role to play in measuring ABCs (and optimising store labour), wastage, supplier funding, and optimising SKU mix and pricing based on these real margins net of ABCs, and other bridges to true margin.

6. Automation and robotics

An entirely new and next generation of retail analytics is required to measure and optimise the impact of automation and robotics on retail performance. Depending on the speed of adoption, this next generation of analytics may become visible as early as 2020.

We now have US’s Lowe’s LoweBot, Tally at Target, Softbank’s Pepper and Best Buy’s Chloe as some of the examples of robotic applications in retail environments.

When you have retail robots, you have continuous, live and on-demand collection of data, which when combined with analytical prowess can be a serious game changer.

Currently, the application of robotics is primarily to reduce poor inventory, which has been estimated to cost the global retail industry to the tune of $1 trillion a year.

But as the applications of robotics increase, more would be the areas of retail management where they will start having impact. But this ‘impact’ can only be measured accurately using high-end analytics, which are also able to connect robotics-driven efficiencies to the wider data ecosystem.

7. Analytics and Artificial Intelligence

Just like robotics, artificial intelligence (AI) is another domain, which is significantly impacting retail operations (and it has a head start over robotics). According to research conducted by IBM, there are six domains where the retail industry foresees high application of AI – supply chain planning, demand forecasting, customer intelligence, marketing management, store operations and pricing & promotions.

Marks & Spencer, still one of the UK’s best known retailers, has teamed up with Microsoft in a high profile collaboration around AI. M&S is starting to integrate machine learning and AI, both in its stores and behind the curtain. Every M&S store will be able to track, manage and replenish stock levels in real time – and best deal with unexpected events such as freak weather.

These six domains identified by IBM encompass the whole retail business model, which means that the application of AI in retail is going to be broad. OC&C predicts that retailers will spend $8 billion a year globally on AI by 2022 (Let’s Get Real About AI). Regardless of whether this is an accurate or inaccurate (the fortune-teller’s dead on the floor?), the applications of AI in retail are evolving and becoming real in 2020.

For example, beauty retailer Sephora has developed ColourIQ, its machine learning driven in-store product that scans the surface of a customer’s skin to provide a personalised foundation and concealer shade recommendation.

Tech companies are buying up AI startups. Should have been my 2014 start-up plan! It’s clear that Retailers like Amazon are going to be increasingly dependent on AI https://interactives.cbinsights.com/artificial-intelligence-acquisitions-by-famga/

The machine learning foundation of any AI application is nothing but a next step from predictive analytics, wherein the machine is able to form hypotheses, test and learn autonomously. It’s the next step for advanced analytics teams for retailers.

8. Building integrated analytics capabilities

Impactful analytics does not work in silos, and this is equally true for retail too. In order to use analytics as a strategic enabler, retailers need to integrate it into the organisation’s structure and business model. In 2020 (and going forward), a key driver of change will be deeper understanding of data available for each individual customer (or prospect).

We have worked with many retail clients helping them to integrate analytics capabilities into their central planning function. To achieve this, some fundamental questions need to be asked, which includes “what should we be using analytics for”, “do we have access to the right data”, “do we have the right system and tools” and “how can we build a data-driven culture”.

In the majority of organisations, analytics capabilities are inefficiently organised wherein pockets of analytics activity are scattered throughout leading to redundant costs and suboptimal adoption. The ultimate objective is to unlock the power of analytics to drive growth for retailers.

9. Predicting macro trends

A mix of qualitative insights and creative thinking has always driven the domain of identifying and predicting trends. I am now happy to see that analytics has started to play an important role in trends work. We’re able to process many signals – for example scraping product reviews (“I wish these jeans had a higher rise”).

This is going to gather momentum in the retail space as retailers and industry experts deep dive into consumer behaviour, demographic, economic and social data to predict what the future has in store for them.

When we work with retail clients, an important component of our engagement is casting a wide lens on the problems at hand. This invariably makes us look outside the organisation to assess the impact of socio-economic, demographic, cultural, economic and political factors.

I still don’t discount the importance of qualitative thinking, but combining it with data-driven hypotheses building and fact finding makes it more impactful.

Analytics plays a key role in minimising the incidence of ‘gut-driven’ decision-making and making them more grounded in facts. In the next 12 months, more retailers would foresee the future through a clearer data-driven lens than ever before.

This will be critical as the costs of failure are significant and the ability to experiment and tinker continuously at a strategic level become riskier.

10. Building the store of the future

It was always a consulting cliche – how to make your store work harder? Sacrificed on the altar of omnichannel, it was seen as a slightly naff thing to say. But of course, for legacy retailers probably still nothing matters more.

My last prediction for the next 12 months is on retail analytics playing a pivotal role in building the ‘store of the future’. The word ‘experiences’ is a much-maligned one but when it comes to building an excellent store of the future, it is the key driver. Consumers are not going to walk into stores to only ‘buy’ things but they would to create their own personal customer journey. They will walk in to explore, evaluate, compare and trial and may walk out empty-handed.

Stores need of course to be smaller, carry fewer SKUs, and exhibit multiple formats so that the economics of any given store makes sense and the estate be integrated with e-comm hotspots. Analytics is great at this.

Use of space within the store needs to be optimised (Analytics is also great at this). And the right rationalised mix of SKUs, with an optimised price ladder.

Retail analytics prowess will enable retailers to create stores that act as touchpoints and not only as a point-of-purchase. Stock optimisation, design and display formats, store layouts, technology-enabled trial mechanisms, in-store ordering, efficient returns and swaps, personalised guides and advisors (both human and non-human) are areas where analytics will increasingly be both at the forefront and backend.

I can’t wait to see what 2020 has in store (ha!) for the Retail industry and how we can enable retailers to seize their future with the full power of their data.