The recent Retail Executive Summit in London featured some fascinating discussions of retail trends and the use of machine learning and Artificial Intelligence (AI) I in retail. For example, we heard from guest speaker Forrester Vice President and Principal Analyst George Lawrie on “Secrets to Meaningful Prices, Personalization and Promotions.”
George summarized key findings from three recent Revionics-commissioned global shopper studies conducted by Forrester Consulting1. The research found that shoppers are very mindful about what promotions they receive and they have clearly preferred channels. The three studies surveyed shoppers in the U.S., Germany, the UK, France and Brazil, and George dove into some of the geographical variations.
On the topic of promotions, for example, relative to shoppers in the UK and France, German shoppers more strongly prefer in-store promotional offers to those in other channels, with 44% of German respondents saying they only pay attention to promotions in the store. Respondents also revealed that they have different preferences for the frequency of promotions depending on the category they are shopping – preferring more frequent offers on, say, groceries versus long-life items like automobiles.
On the whole, George noted that retailers err on the side of misguided, too-frequent promotions, particularly in email. In fact, 69% of respondents said they had received a promotion via email in the past month for a product they would have been willing to pay full price for. Retailers who spray out indiscriminate promotions risk more than lost revenue and margins, however: 37% of shoppers said that receiving irrelevant offers caused them to feel annoyed, shop at that retailer less often or provoked no reaction.
George also noted that the majority EMEA shoppers are tolerant of frequent price changes as long as prices were fair and non-arbitrary. The concept of fairness is critical, since 59% of shoppers would not make a purchase if they perceived a price to be arbitrary.
Utilizing science- and AI-enabled price and promotion optimization can enable retailers to price in accordance with localized preferences, deliver fewer but more focused and meaningful promotions, and better synchronize promotions with customer buying cycles.
This was a great lead-in to the panel discussion, “Revionics Executives Speak Out on AI and Machine Learning in Pricing,” again moderated by Cheryl. Panelists included Founder and EVP Corporate Strategy and Development Jeff Smith, Chief Customer Officer Steve Leven, Chief Science Officer Jeff Moore, and Chief Technology Officer Dave Thompson.
Jeff Smith talked about his vision in founding Revionics, with the goal of an AI-driven, SaaS-based architecture with complete transparency for retail users baked in from the very beginning. Steve echoed the importance of full transparency and discussed the value of a modular adoption approach where the ROI from each phase can fund the next phase. This enables retailers to start with, for example, rules-based price management, followed by more sophisticated price optimization at a later date, earning trust and adoption along the way.
Dave Thompson and Jeff Moore discussed the industry hype around the terms AI and its Machine Learning (ML) currently. Neither is new, with a history of tools and algorithms that date back several decades. The key to applying AI successfully in the real world is to know which tools to apply for which types of business challenges, as well as having deep domain or industry expertise is the business area you are addressing. They pointed out that with algorithms that have been learning for well over a decade, Revionics’ capabilities are very mature and well-proven.
Revionics Principal Optimization Scientist Dan Marthaler capped off the day with “Uncovering the Mystery around AI in Pricing.” To get past the industry hype, Dan provided Revionics’ definition: AI-powered solutions are goal-seeking agents that seek to identify or implement decisions likely to produce beneficial outcomes as informed by the data in their environment. ML, in turn, is a problem set and a set of associated tools that experienced data scientists can utilize to develop approaches to solve these problems. Dan explained the many ways in which ML is applied in Revionics’ products, and noted that an extensive toolkit is required due to the importance of using the right approach for each type of problem.
Dan explored real-world situations in which Revionics science was applied for promotion optimization. In reality the vast majority of promotions do not generate money for the retailers. With AI-powered solutions retailers can identify optimal promotions and (immediately stop!) ineffective offers, and zero in on the balanced offers that drive both positive revenue and profit lift. Based on real data analysis, a typical retailer sees 100% of their profit driven by just 5% of their promotions – while the vast majority of other promotions cancel out each others’ profits. Good analysis can uncover the nonproductive or counterproductive promotions so retailers can stop the bleeding immediately.
Check back soon for a summary of the other engaging discussions at the retail summit!
1 Revionics-commissioned studies conducted by Forrester Consulting: Understanding Retail Customers’ Pricing Expectations And Tolerances, May 2017, Understanding Retail Customers’ Pricing Expectations And Tolerances, November 2017, and Indiscriminate Promotions Cost Retailers, May 2018.