For the past few years, retailers have increasingly experimented with various alternative digital and eCommerce models to evaluate their applicability and profitability for certain markets and customer segments. This has included online ordering for curbside pickup or home delivery, as well as an array of mobile shopping options.
Prior to the recent emergence of COVID-19, adoption of digital commerce technology by traditional brick and mortar retailers has been somewhat underwhelming, and even painfully slow. This is because many of these options have not yet proven themselves to be adequately profitable to the retailer, especially when compared to traditional in-store visits. As a result, in the search for a profitable model, the industry has been more focused on the pros and cons of using proprietary retailer-operated eCommerce solutions, outsourcing to third party delivery companies such as Instacart or Shipt, or a combination of in-house and outsourcing.
In 2017, online grocery sales were predicted to capture 20% of total grocery retail by 2025 to reach $100 billion in sales, according to a Food Marketing Institute (FMI), with a study conducted by Nielsen. These figures were subsequently revised and accelerated to reflect that the industry would achieve that $100 billion sales volume target hit by 2022. It is now likely that this milestone will be achieved much sooner due to the extraordinary eCommerce volume attributed to COVID-19.
With the recent virus-related stay-at-home orders forcing so many people to be socially distanced, this has unexpectedly accelerated, and stress-tested the technologies and operational constraints of these eCommerce models. Some retailers have consumed all of their existing eCommerce capacity and are unable to satisfy new shopper demand for these services. At the same time, virus-related increases in traditional in-store shopper traffic has likely obscured an undeniable point that these eCommerce models have been artificially turbo-charged and shopper comfort with these digital experiences has changed forever.
While some retailers are focused on the fundamental enablement of these digital commerce models, others who were earlier adopters have realized that they now need to shift their focus. While they have been focused on enablement, they have underinvested in the necessary analytics to run these new channels in a sustained and profitable fashion. During the nascent state of eCommerce initiatives, most channel-related decisions have been made instinctively, rather than in a data-driven fashion. The timely emergence and affordability of AI-powered analytics for retailers now seeks to help remedy the online blind spots that currently exist.
What is AI and why is it an important enabler for eCommerce?
Artificial intelligence (AI) is the ability of a computer program or a machine to think and autonomously learn on its own. AI encapsulates a broad set of technical capabilities designed to provide better outcomes, augmenting what humans would otherwise need to do. More clearly, AI is software that mimics and automates the tasks that humans previously had to do exclusively. These tasks include learning, reasoning, problem-solving and even understanding language. Common AI terms that you may have heard include machine learning, deep learning, robotics, natural language processing (NLP), and more.
Below are some examples of how AI-powered analytics for retailers can help bolster eCommerce performance:
Online Product Assortment
Just because a retailer may have 20,000 to 50,000 items sold in a given store, does not mean that the entire online product assortment must match what is physically sold in that same store. To the contrary, a retailer’s online assortment should be tuned to provide product coverage on those items that are most in demand, and most commonly available-to-promise with inventory stock on hand to fulfill orders.
AI-powered analytics for retailers can assist in curating what the ideal online assortment should be. This can be tuned location by location to ensure alignment with unique demographics that may exist in those locations. Smart substitution recommendations can also be optimized in the off chance that the product ordered by the shopper is not available. AI can also help to evaluate the viability of extended aisle products that add non-stocked items usually fulfilled by complementary third-party companies.
Online Intelligent Pricing
Just because a retailer has optimized in-store shelf pricing does not mean that the online pricing must be – or even should be – the same. In fact, all experts acknowledge that online fulfillment costs including order picking costs and delivery costs are much higher, no matter how efficient a given retailer may be. In addition, the shopper demand signal, i.e. a shopper’s sensitivity to price or price changes on specific items, in an online channel will be materially different than the behavior of shoppers buying in-store.
AI-powered analytics for retailers can help them understand what the optimum pricing should be for items purchased in online orders, as well as their respective pick-up fees and/or delivery fees. Since online shopper behavior is generally less-price sensitive even on Known Value Items (KVIs), AI can help retailers maximize profitable revenue, drive repeat orders, and match competition where it makes sense.
Online Intelligent Promotions
Most grocery retailers have three primary promotional vehicles that are used to shape shopper behavior; an ad/circular which may be print-based, digital, or both, in-store temporary price reductions (TPRs), and manufacturer-sponsored coupons. Some retailers may also elect to offer two-tier promotional pricing models, featuring deeper discounts for the most loyal frequent shoppers, and less attractive promotions for less loyal shoppers.
AI-powered analytics for retailers can help them figure out how to optimize promotional offers to achieve the desired shopper-specific behaviors including securing bigger baskets, increased profitability of those baskets. Additionally, retailers can avoid counter-productive behavior such as over-discounting, cherry-picking, pantry loading and unwanted offer-stacking.
Online shoppers generally have very negative reactions when retailers make non-relevant, distracting offers to them that interrupts the flow of the online ordering experience. After all, does it make sense to promote meat products to a vegetarian shopper? How does it help your business grow if you promote baby products to empty nesters?
Online Demand Forecasting
Nothing is more frustrating to an online shopper than not getting what they originally ordered. Some shoppers experience as much as 15% of their original order either not fulfilled as intended or substituted with another item that wasn’t originally ordered. When customers complain, retailers often just refund the amount without getting the product picked up. Or worse, they do pick up the product, which creates an even bigger impact on profits. These hidden losses can amount to tens of thousands or hundreds of thousands of dollars.
From the retailer’s perspective, avoiding lost eCommerce sales attributed to out-of-stock merchandise is a key benefit of highly accurate AI-powered forecasting specifically for online orders, ideally with simultaneous awareness of in-store demand which competes for that same merchandise. If these lost sales attributed to out-of-stock merchandise could be reduced by 50% or 75%, it would significantly help retailers grow their eCommerce top line revenues.
Being able to do AI-powered demand forecasting specifically for online shopping activities can help retailers optimize and manage constrained pickup and delivery time slots / windows, schedule labor for pickers and drivers, and it can maximize availability of prepared foods by time of day and day of week which will result in less overproduction and spoilage.
All online ordering takes place with full individual shopper identity. In other words, we always know who placed eCommerce orders, what they searched for, what they bought, and when they bought. This means that even in non-loyalty card environments, retailers are able to and should personalize the online shopper experience to reduce friction or frustrations that can easily be avoided. A highly personalized experience will also drive better shopper experiences resulting in more frequent and profitable online orders.
AI-powered personalization also allows retailers to segment their shoppers by frequency, recency, category level participation, total spend and product attribute preferences, and then tailor the many offers in their “offer banks” into highly relevant, basket-building experiences.
In traditional loyalty shops, for those with some in-store shopping history, AI-powered basket analytics for retailers can help curate online order guides based on that prior shopper history. Doing this can simplify helping the shopper to find the products that they are most likely to buy. This is especially important for first time online orders.
Set high benchmarks with AI and monitor KPIs across functions
Beyond Advanced Planning for online Merchandising and Marketing, there is an assortment of general and operational Key Performance Indicators (KPIs) and analytics that are highly relevant for eCommerce retailers who seek to run a tight ship. Here are examples.
Sales & order volume by channel (curb-side, delivery)
- Order size in revenue and items per order by time of day and day of week.
- Category diversity analytics
- Customer profitability analysis
- Identify causal relationships or attribution on what factors can drive sales outcomes
- Customer purchase patterns by channel
- Picking efficiency and accuracy
- Picker and Driver productivity analysis
- Out-of-stocks and substitutions by order
- Total labor per order (projected vs actual)
Shopper feedback and satisfaction
- Shopper performance trends related to frequency, recency, spend and category participation
- eCommerce Order refunds and adjustments by reason
- Abandoned online baskets / carts
- Store order picker and driver satisfaction ratings
- Support ticket analytics and response times
Legacy BI Analytics for Retailers are Not Sufficient in the Digital Era
Over the past two decades, in-store-centric decision making has been supported mostly by legacy tools including data warehouses, report generators, business intelligence (BI) solutions and even Microsoft Excel. Regrettably, these tools weren’t built for the agility required in the digital era. They are too expensive, too difficult and complex to use, and they don’t provide actionable real-time insights. The inadequacy of traditional, legacy reporting and business intelligence tools results in three common challenges.
- Decision making based on Legacy systems is too slow, laborious, and inaccurate. This results in decision makers having a dependency on data analysts and technology professionals to extract clean and useful information and usually this ends up being stale by the time it is received. eCommerce data-driven decision making requires democratization of data without a dependency on analysts.
- Source data required for eCommerce optimization is derived from many legacy systems and consequently is siloed and not accessible for timely decision making. Decision makers focused on eCommerce require synthesized data from multiple systems which today are not well-integrated and are provided by different software vendors. AI solutions solve this by “stitching together” disparate data into a unified data fabric leveraging Auto Machine Learning (AutoML) which is required for agile, data-driven decision making.
- Spreadsheets and traditional reporting tools are not sufficient, insightful or actionable enough for digital operations. These traditional tools typically only provide a complicated look back on past business results that can’t be changed. Spreadsheets and transitional reporting tools fail to explain why results are good or bad and what should be done to improve results. AI-based solutions can also look backwards, however, they are really designed to provide predictive and prescriptive insights automatically, a huge sea change from the old approach.
eCommerce is now mainstream; retailers require analytical solutions to optimize their results
While it is clear that in-store shopping is not going away, it should also be clear that eCommerce has become a critical-mass component of retailer revenues. This will only grow for the foreseeable future.
Perhaps you have heard the phrase “cobbler’s children have no shoes” meaning that the shoemaker hasn’t gone to the expense or invested the time in providing shoes for his own children. eCommerce businesses can no longer be the “cobbler’s children” of retail enterprises. They must be provided the tools to make better, faster, smarter decisions that are data driven and forward looking. AI-powered analytics for retailers can play an important role in making eCommerce channels a highly profitable and attractive part of the retail enterprise. Online shoppers are depending upon their retailers to go all in on making the online shopping experience truly great.
Meet Hypersonix – Real AI, Real Results for eCommerce
Hypersonix is a fast growing, venture-backed, San Jose, California based company that offers a comprehensive cloud-based, AI-powered analytics solution enabling decision-makers to analyze disparate data sources and derive actionable insights, simply and quickly, without needing to depend on IT or analysts being involved. The company serves Consumer Commerce enterprises like retailers, restaurants and hospitality. Hypersonix helps clients drive profitable revenue growth, save money, and deliver exceptional customer experiences.
Harnessing the power of AI innovations like automated machine learning (AutoML) and natural language processing (NLP), Hypersonix offers its customers an exceptionally simple and fast “Google-like” text and voice search experience for their business augmented by “Jarvix” an embedded virtual decision assistant. It offers predictive and prescriptive analytics for retailers designed to help users measure results, understand root cause on why results are positive or negative, and what to do next.
✎ by Todd P. Michaud, President & Chief Customer Officer at Hypersonix, Inc.