Grocery: Building an economically sustainable e-commerce model

Opportunities & pitfalls for retailers, CPGs, tech platforms and logistics partners

Summary:

• Grocery e-commerce is growing, but models that are dominant today will continue to have challenging economics and will not scale beyond a point
• There is no single silver bullet to addressing this. Not gig economy. Not robotics
• To build a sustainable e-commerce business, retailers should build an integrated operating model across stores, last mile, merchandizing, supply chain and digital
• Commercially available technology solutions to enable this are inadequate, creating an advantage for tech-savvy retailers and an opportunity for tech firms

Note: This is written with a primary focus on grocery retail, though the framework and several concepts are applicable more broadly to other forms of retail

Introduction:

Customers are expected to continue doing more of their grocery shopping online. Retailers are focused on e-commerce growth to keep market share. Yet, as e-commerce becomes a larger proportion of their revenue, its profitability is coming into increased focus. CPG manufacturers are investing to win in e-commerce but are often finding the traditional approaches to shopper marketing inadequate. Many technology vendors have emerged as solutions to these. But these are often point-solutions. While helpful in the short run, this can lead to increased complexity and poor customer experience in the long run.

To build a sustainable e-commerce offering, retailers should build an integrated operating model across five core areas: stores, last mile, merchandizing, supply chain and digital.

Exhibit 1: Building a sustainable e-commerce model

In this post, I have highlighted some observations, strategic questions and design choices along these areas. I have also included some clear opportunities for technology solution providers.

1. Store

Stores are increasingly becoming hubs for e-commerce. Stores are already located near where people live and carry the products suitable for that demographics. However, picking from shop-floor for e-commerce is subject to limitations of the store such as in-stock conditions and operating hours. At higher volumes, it can crowd out customers during busy times. In addition, it is inefficient as stores are not designed for customer experience and not picking efficiency. The higher cost needs to be borne by the grocer, the customer and the CPG manufacturer. While picking on shop floor is the dominant picking model today, I expect some of the volume to migrate to other models, particularly in large markets. Yet, I do not believe picking on the shop floor is going away any time soon. There will always be markets and use cases where this is the most logical choice. 

Of late, there has been a lot of interest in automated Micro Fulfillment Centers within stores. I am bullish on this model for the long run, have piloted it (see WSJ article). MFCs can leverage the real estate footprint, truck deliveries to the store, store labor, and long tail of SKUs. They can do all this while driving down pick cost for the most common items. But retailers should still be aware of the following: First, many of the robotics platform vendors are young and unproven. Second, it involves high capital expenditures. Making space for automation machinery involves contracts, permits, and on occasion, structural work, making it even more expensive than the machine itself. Third – it is not a silver bullet to solve all the retailer e-commerce problems. The front end, merchandizing, order management, marketing, delivery orchestration, etc. still need to work well and the MFC system needs to plug well into these. Finally, like all automation, it is designed for best efficiency given one set of parameters and could prove to be somewhat rigid. For example, once installed, it might be challenging to change the number of SKUs or throughput capacity significantly.

Retailers should pilot different models. However, some retailers may find running a “manual fast pick room” in a few locations a good first step before evaluating automation.

Exhibit 2: Summary of alternatives to picking and packing:

The cost noted here is indicative and will vary based on design, operating efficiency and number of SKUs offered.

2. Last mile & service layer

Delivery and pick up are not new to retail (think milk route). Yet, most retailers today do not have a robust last mile capability necessary to leverage their store footprint for e-commerce. This capability is more complex than it appears. Customer preferences vary (pick up at store, home delivery on kitchen counter/pantry, home delivery to customer at the door, home delivery left at the door, etc). There are cold chain considerations (refrigerate the truck, place grocery bags in a cooler, or insulate each bag). These choices have an impact on cost, cross docking and limitations on some models (e.g., how long groceries can be left unattended at customers door). Further, large orders (e.g., multiple cases of water), alcohol or prescription medication, etc, all require special handling.

For many retailers, particularly smaller ones, it might never make sense to build their own last mile fleet. Instead, they should partner with one of the many last mile logistics providers. For others, owning the last mile experience might be a core part of their customer value proposition. Below are some broad models and considerations.

Exhibit 3: Summary of last mile delivery models

3. MERCHANDIZING 

Merchandizing is central to retail. Yet, most of e-commerce merchandizing decisions are small tweaks of decisions made for the store.

The merchandizing decision that can impact e-commerce business most is how many and which specific SKUs are offered on e-commerce. The optimal operational model to offer 4,000 SKUs will be different than one needed to offer 20,000 SKUs. A customer-back SKU-rationalization exercise with e-commerce pick efficiency in mind can be very instructive in identifying the right operating model or models.

In addition, there is opportunity for many retailers to offer items that are popular on e-commerce but not in the stores (e.g., large pack sizes). There is also an opportunity to build an affiliate model for traffic generation to leverage CPG advertising to drive traffic to retailer e-commerce. In addition, there might be opportunities to address the risk to impulse-buy categories on e-commerce – for example, by branding and up-selling at customer pick up points.

Furthermore, digital and e-commerce can influence merchandizing for the overall business. For example, digital signals (e.g., search and product page visits) can provide information on store assortment decisions. Integrated shopper marketing is another big opportunity. Customers shop multiple channels. Shopper and brand strategies should follow. Evolving a model for coordinated in-store and online tactics to achieve a sales goal for a CPG partner has shown promise.  

4.SUPPLY CHAIN

May retailers have built efficient supply chains that are optimized to get things to the store. As e-commerce proliferates, we should expect more customer direct capabilities embedded in parts of the supply chain. For example, picking in ‘eaches’ for direct shipment in lower cost warehouses. This will broaden the traditional role of supply chain and make it even more central to the business economics and customer experience.

Logistics analysts, working with the merchants, might first solve for the footprint – ie., Stores, dark stores, local warehouses and capacity, central warehouses. Then run an analysis on which SKU goes where factoring in customer needs, SKU velocity, split shipments, etc. Finally, they might work on pricing items by service levels – for example, an item available for delivery within 2 hours might cost more than the same item if it comes in a container with other things in one week.

A note on dark stores:

A “dark store” is a small box, potentially a former live store, located within the city which is optimized for picking and staging and does not entertain customers. For many grocery retailers, retail operations (including running the store, utilities, store labor, etc.) accounts for between 20 and 30% of gross merchandise value sold. Using the dark store will free up a large part of this cost that can be repurposed for e-commerce operations and last mile delivery. If done right, this can provide much lower operating cost and delivery groceries to the home at same price as in store, while still making a healthy margin. It must be noted that many firms, both startups and established retailers, that have tried this model have failed. Of these, some were too early to the market and had insufficient demand density or were sub-scale to have attractive procurement economics. Others didn’t quite get the ordering and inventory management right leading to high shrink and dump waste. There are a few firms making this model work well, and I expect this model gaining more traction due to its obvious advantages.

Evolution of grocery operating model over time

I believe that store pick and gig delivery will not go away. We will also see a much larger uptake of other models, particularly dark stores, MFC and delivery through van routes with multi-drop. In the Exhibit 4 below, the rows are different ways the retailer can pick and pack. The columns are how it is delivered. Implied in each model is its service levels and costs (not shown on this graph).

Exhibit 4: Evolution of e-commerce models

It might be a worthwhile exercise for each medium or large retailer to build a model similar to this one on how they see their long term e-commerce operations evolve - given their footprint, customer base, merchandizing strategy, etc.

5.DIGITAL

Most retailers will operate their e-commerce in multiple models as described in Exhibit 4. However, most technology offerings tend to be closely tied in with one specific model. Customers cannot be expected to order differently for home delivery vs. store pickup, or engage differently with different last mile providers.

Many of the grocery-specific technology solutions available are limited in scope and do not handle this and other complexity as the industry evolves. Some of the more mature, industry-neutral technology platforms are do not work for the grocery industry without much customization and add-ons. There are five areas of opportunities for technology service providers:

a. Up-funnel traffic, customer, coupon

The up-funnel customer behavior (i.e., discovery, list making, coupon clipping, loyalty programs) is virtually the same for customers wanting to purchase in the store or through e-commerce. Most of the customers that choose to purchase in the stores also use digital channels for most of their up-funnel activities. However, commercially available e-commerce platforms do not integrate well into ways to enable and influence in-store shopping. There is a parallel universe of loyalty/couponing solutions that focus on enabling in-store purchases only, but do not plug well into e-commerce solutions. Most customers are omni-channel customers buying both from store and online and expect a seamless experience. Hence, some retailers have invested in integrating these flows, at least in parts, and have invested in Customer Data Platforms. Such integration is harder if large parts of the e-commerce revenues comes from a multitude of third-party platforms - something that continues to challenge many retailers.

b. Grocery commerce platform

Most e-commerce offerings come with search, order, banner ads, checkout. But these are provided in a “one size fits all” model that will not work with the complexity of larger retailers, and do not plug directly into other parts of the digital journey. Basket building continues to be a chore for many customers that is not fully solved by anyone. Search features of most out of the box offerings are inadequate for grocery. There is also an opportunity for many of these platforms to link better with sources of up funnel traffic and connect seamlessly with different modes of customer engagement (i.e., web, app, email, social, push).

c. Catalog (including product description, price)

Most traditional grocers carry similar items. Yet, getting the catalog right with good quality images, product description and other content is difficult for many. Stores of a retailer even in the same city often carry different SKUs at different price. Items do not always come in and out of stores in a coordinated way. Same fresh items sourced from different vendors are sometimes tagged differently. Items such as wine can have valuable content associated with them (e.g., sommelier comments) that needs to be licensed and integrated. Clever solutions addressing these and integrating with the rest of the systems will be immensely valuable. 

d. Pick systems & local e-commerce inventory mgmt.

There are many pick software vendors with comparable features. The more mature ones can do multi-mode picking (single pick or batch pick), provide central labor management metrics to store manager, link with store planogram, substitution recommendations and customer communication features. In scenarios with fast pick warehouse, MFC or dark stores, it is critical to get inventory management right to reduce shrink and reduce out of stock scenarios. There are some interesting startups with AI based capabilities that are addressing this well. This will continue to be a space to watch.

e. Delivery orchestration

Multi-vendor delivery orchestration is a new area that retailers have not had to focus on so far. Over time, most retailers will have many different ways of getting e-commerce purchases to the customer. This includes store pick-up, gig-economy based delivery services, contracted vans or owned trucks. Customers will expect a unified interface for all these for messaging, routing information, exception handling, etc. There are some good emerging platforms for this. However, there is much more ground to cover in this space.

In conclusion ...

With ongoing customer demand, new entrants and changes in the overall food eco system, the space of grocery e-commerce will only get more exciting. Retailers who integrate e-commerce and digital into all aspects of the business will have an advantage over those who try to strive for e-commerce profitability through point solutions. There is scope for a more robust and comprehensive grocery-tech player to support in this journey.