Introduction Pricing is a key strategic competitive factor for retailers. Price optimization uses data analytics to formulate the desired pricing level and expand the probability of achieving specific strategic objectives. Advancements in AI and machine learning are helping price optimization strategies become more accurate and effective.   This report — sponsored by global retail strategy, analytics and cloud software company Precima — will first look at the main challenges retailers face when devising pricing strategies. We will look at how AI and machine learning are strengthening retail analytics and empowering advanced price optimization strategies, making it easier to achieve strategic goals, such as profit maximization and operational management.  Finally, the report shows how we expect the role of AI in price optimization to increase in the future, and how the analytics capabilities of AI will be complemented by new disciplines including neuroscience. The Challenge of Setting the Right Price Pricing has become a complex endeavor and a key strategic competitive factor for retailers. A sound pricing strategy requires an in-depth understanding of internal and external dynamics that affect pricing. The challenges pricing strategists face reflect the complexity of the operations and business environment in which the retailers operate: 

• Retailers now have more SKUs to price, and more channels – each with their own pricing dynamics. 
• Retailers face a highly competitive environment in which other players offer attractive price propositions and competition comes from multiple channels.
• Consumers have become very savvy, and shopping around for better offerings has never been easier.

When pricing products, retailers need to consider many factors: 

• Competition: Regularly monitoring how competitors price similar products.
• Market positioning: For example, mid-market grocer operations run on higher costs than no frills discounters. These differences should be considered when formulating pricing. 
• Consumer perception: How much shoppers are willing to pay for specific items can vary dramatically depending on seasonality or demand peaks. 
• Strategic objectives: Pricing should also reflect strategic objectives. For example, the retailer could price a new product competitively during launch to encourage shoppers to try it, then raise prices once demand is established. 

Retailers also need to consider correlations between the price points of different SKUs in their tiered pricing structures, and the impact of a single item’s price on overall sales.   Impact of AI on Price Optimization Digitalization is making pricing more complex and competitive, but also gives retailers the data needed to drive pricing decisions. Strategies such as adjusting pricing according to consumer demand were first introduced by airline operators and the hospitality industry with dynamic pricing. This strategy — often improperly used as a synonym of price optimization — refers to a more specific pricing method that consists of applying variable prices to the sale of products and services. Digital technology made it possible for retailers to implement dynamic pricing strategies with large and wide-ranging inventories, such as grocery. Through AI and machine learning, optimal prices are periodically calculated by pricing algorithms and updated to reflect changing conditions. AI and machine learning can glean patterns from data on its own, instead of being explicitly programmed to do so.    [caption id="attachment_93828" align="aligncenter" width="700"] Source: Coresight Research[/caption]     The algorithms calculate optimal pricing by considering several variables, including:  

• Competitor prices: Competitor pricing is a key variable – especially as e-commerce makes price comparison extremely easy for consumers. 
• Consumer behavior: Tracking consumer behavior reveals information about the intention to buy. For example, if a customer repeatedly returns to a product page, it is likely the customer is interested in buying. This insight could trigger the retailer to raise the price, knowing the prospective customer has already planned to purchase. In fact, raising the price could create a sense of urgency that would encourage immediate purchase prompted by the fear of a further price increase.
• Shopping periods: Retailers can adjust prices when demand for a certain item usually surges, such as before and during holiday seasons. 

Other factors that influence pricing can be more category-specific. For example, weather conditions tend to affect pricing of categories such as grocery, apparel and home improvement more than other categories. There are different pricing strategies that can be considered by a price optimization platform. Examples include: 

• Peak pricing: Charging higher prices when consumer demand is particularly high and the offering is scarce, conditions that make the shopper less price sensitive. 
• Price skimming strategy: Pricing a new product high when the market has few competitors.  
• Penetration pricing: Pricing a product at a low price to encourage shoppers to try it. 

The result is an optimal price to achieve strategic objectives, which could vary based on the different stages of the product’s life cycle. For example, there could be a lower optimal price for a new product launch, increasing with maturity and reducing again in promotional and markdown stages. [caption id="attachment_93829" align="aligncenter" width="700"] Source: Coresight Research[/caption]   The Holistic Approach: Total Store Optimization With today’s complex retail operations, pricing decisions cannot be taken in isolation but need to be balanced within the context of a retailer’s entire inventory, since how one category is priced has an impact on the sales of other categories. This means when devising and implementing price optimization, retailers need to adopt a holistic approach that gives a broad view of the impact of pricing different SKUs on the business overall. Total store optimization (TSO) defines this approach and consists of a strategy that allows retailers to make pricing decisions across the entire store and to simultaneously plan price, assortment and promotion decisions. TSO uses AI and machine learning advanced analytics to analyze consumer data and simultaneously formulate price, assortment and promotion recommendation decisions. TSO capabilities become especially valuable in e-commerce operations, where the large SKU ranges that make up online vendors’ inventories require pricing strategies informed by advanced analytics. But the approach finds valuable applications in brick-and-mortar stores also. In this context, for example, some items can be priced lower to drive perception and volume, while others — perhaps the ones characterized by lower demand elasticity — can be priced higher to drive more profit without compromising consumer loyalty.  TSO enables retailers to successfully plan simultaneous implementation of different pricing strategies, making it easier to implement promotional campaigns and loss leader strategies — in which a product is marketed at a low price to drive traffic in-store — by enabling the analysis of the overall results of simultaneous and different pricing strategies on KPIs. Moreover, analysis of promotional activities allows retailers to determine which promotion is more effective and should be continued or expanded, and which is not performing and should be discarded.  Price Optimization Platforms Improve Retail Operations The key advantage of price optimization is the ability to offer the optimal price at the right time to the right customer, to maximize sales conversion and margins. The approach is used to achieve different strategic objectives, such as encouraging shoppers to buy a new product, or the simultaneous roll-out of promotional campaigns.  Adopting price optimization solutions can lead to significant operational improvements, thanks to the use of AI and machine learning for advanced analytics that enables better understanding and operations management, including:   

• Enhanced operational efficiency and target achievement: Price optimization enhances operational efficiency thanks to increased automation of the decision-making processes, which makes it faster and more accurate. Inventory planning is improved by advanced analytics that can promptly inform retailers what SKUs need to be stocked for optimal response to consumer demand. Enhanced operations help maximize returns and improve KPIs. 
• Enhanced shopper loyalty: Predictive analytics give retailers the insight needed to formulate a more accurate forecast of what shoppers are likely to buy and inform inventory management decisions. The enhanced ability to meet consumer demand by having in stock what shoppers want to buy improves shopper loyalty, as consumers will not need to turn to other retailers to find what they are looking for. 
• Stronger collaboration within the supply chain: Effective price optimization strategies require closer collaboration between the retailer and supply chain partners, which means they also share insights and more closely integrate operations. 

Given the numerous advantages of price optimization, many large retailers have already embraced the approach and the technology that enables its implementation. Online Players Pioneered Price Optimization in Retail Fluid pricing policies were pioneered by the travel and hospitality sectors. In FMCG retail, online players were among the first to adopt price optimization strategies, thanks to the abundance of granular data on shopper behavior that e-commerce can deliver.  Amazon is a classic example of the use of price optimization in retail. Amazon constantly scans competitor prices and uses dynamic pricing algorithms to make decisions based on competitor dynamics.   Third-party retailers selling on Amazon Marketplace also use price optimization strategies as they compete to appear on the “buy box” for a product. While a vendors’ position on this buy box is determined by factors such as sales volume and reviews, price is also a key determinant.  Figure 3 shows how dynamic pricing works on the Amazon Marketplace.   [caption id="attachment_93830" align="aligncenter" width="700"] The price of a skincare product sold by Amazon ranged from $24.99 to $20.01 over the one year period through June 18, 2019
Source: CamelCamelCamel/Amazon[/caption]   Other e-commerce companies using price optimization include:

• JD.com: Chinese e-commerce company JD.com announced at the Rise tech conference in Hong Kong that the company was using AI algorithms to deliver personalized shopping experiences, anticipate consumer demand and optimize pricing in real time on inventory of four million SKUs in each JD warehouse. The company’s price optimization approach has significantly increased the efficiency of its pricing process, according to company reports.  
• Alibaba: The Chinese technology company and e-commerce giant uses AI across its business. In retail, Alibaba uses AI in product search and recommendation, customer support, personalization and supply chain management. Alibaba applies AI to assist vendors with inventory management and price optimization, by forecasting product demand, determining the right product mix and selecting the right pricing strategies, according to Alibaba’s news portal Alizila.  
• eBay: eBay has also been using AI extensively. According to company reports, in 2016 the e-commerce giant bought a technology company that specializes in advanced analytics to expand its capabilities to help vendors optimize pricing. 
• B2W Digital: The Latin American e-commerce company — which operates through Submarino.com, Americanas.com, Shoptime and Sou Barato — has been working with the Massachusetts Institute of Technology (MIT) since 2015 to develop and roll-out an AI-powered price optimization platform, which uses internal and external data such as pricing, discounts, advertisement spending, weather conditions and competition activity to formulate optimal pricing. 
• BuildDirect: The Canada-based home improvement online marketplace has been developing a proprietary AI engine to analyze data such as brand attributes, consumer price preferences, competitor pricing and customer transactions. The resulting insight helps vendors inform pricing strategies, according to company reports. In 2017, the company hired an executive from Amazon to oversee the development of the AI platform, according to company reports. 

Brick-and-Mortar Retailers Embrace Price Optimization  Several brick-and-mortar retailers have also embraced price optimization strategies to keep up with the competition from online juggernauts such as Amazon. The multichannel operations that characterize today’s legacy retailers imply that price optimization applies to store operations as well as e-commerce platforms.  Price optimization in a physical store environment differs from online, as the advanced analytics platform that formulates optimal pricing needs to factor in variables associated with a physical space, such as the way items are displayed and how much space they take on shelves or how shoppers react to a given layout of the products in store. 

• Kroger: The US retailer has been testing a machine learning-powered “smart shelf.” When Kroger’s mobile app is open, sensors detect shoppers in aisles and offer personal pricing and highlight products in which the customer might be interested, Forbes reported in July 2018. 
• Harps Food: The US supermarket chain has been collaborating with a tech vendor since 2017 on pricing and promotion strategy, according to media reports. The partner’s AI-based platform is helping the retailer make strategic decisions such as what products to promote, what price point and how much inventory should be allocated. 
• Kosmo: The Eastern European FMCG retailer has been collaborating with a tech company to roll out an AI-based price optimization platform, to help the retailer gain more control over pricing and promotional strategies. Kosmo ran a successful two-month pilot and said it is now ready to roll out the technology across its operations, according to specialist publication Retail Insider. 
• Bonprix: The German retailer has been collaborating with an external tech provider for the last four years on an AI-powered price optimization platform to improve margins, specialist publication Essential Retail reported in June 2018. The technology uses datapoints on brands, seasonality, buying patterns and national holidays to calculate the optimal price for some 20,000 items every day, enabling the retailer to set prices across its international markets automatically.

A large European supermarket chain worked with Precima to leverage consumer data gathered from the retailer’s network of over 750 stores. The objective was to inform strategic decision-making including pricing, assortment and promotion, managing omnichannel operations, enabling personalization and targeted marketing. The supermarket chain saw a 1% to 3% year-over-year sales increase in the four-year period to 2018, according to the retailer. Price optimization is a key part of the retailer’s strategy, and was gradually implemented from an experimental stage to a comprehensive application of the approach across the business. [caption id="attachment_93831" align="aligncenter" width="708"] Illustration of how the large European retailer that worked with Precima implemented price optimization, starting from an experimental phase in year one and extending the strategy to 80% of its inventory and 100% of its distribution network by year three.
Source: Company reports[/caption]     Partnering with Specialist Data Analytics Providers  As we have seen from the examples above, retailers can either develop a price optimization platform in-house or turn to an external vendor, such as a specialist data analytics or/and software company. Developing an in-house platform is a viable option for companies that have the resources and advanced technical skills within the company, which tends to be the case for digitally native companies such as Alibaba and Amazon, which over time have evolved from e-commerce companies to tech giants.  But for most retailers, and particularly legacy retailers, developing solutions in-house can be an expensive and resource-draining endeavor. The best option, in the long run, could be to turn to specialist data analytic vendors to develop a price optimization platform that can be integrated into their operations. The Unconscious Future of Price Optimization We expect the number of retailers using AI-powered platforms for price optimization to increase in the coming years, as pricing strategy will continue to play a key role in competitiveness. Figure 4 shows how the global market for AI software — of which price optimization platforms are part — is expected to grow at a 43% CAGR in the seven-year period through 2025.  [caption id="attachment_93834" align="aligncenter" width="700"] Source: Tractica[/caption]     We expect further expansion of individualized and hyper-personalized pricing strategies in the years to come. Retailers will find more advanced ways to capitalize on the data shoppers share through membership and loyalty programs, or by using retailer apps on mobile devices.  Location marketing will increase in importance, and retailers will have different product offerings and pricing based on location. AI will find advanced applications to determine how much an individual consumer would be willing to pay for given products through different channels and in different locations. Algorithm-generated prices and promotions will specifically target consumers based on location. Beyond AI, we expect neuroscience to emerge in price optimization in the future and to complement AI-powered analytics. Neuroscience is instrumental in understanding human conscious and unconscious buying patterns and to identify the ideal price. Shopper buying behavior is informed by conscious decisions based on factors including price, quality, convenience and availability, but part of the decision-making process is unconscious. Applying neuroscientific methods, such as electroencephalography and eye tracking, will help retailers gain additional insight into the unconscious decision-making process, adding further information that will complement the analysis already tracked by AI-powered analytics platforms.  Key Insights Price optimization uses data analytics to formulate desired pricing levels and maximize the probability of achieving specific strategic objectives. Digitalization has made price optimization processes possible and, thanks to AI and machine learning, much more effective. Factors such as competitor prices, consumer behavior, weather and seasonality are computed by algorithms to come up with optimal pricing.   Retailers apply price optimization to achieve different strategic objectives, including profit maximization, shopper loyalty, expanding the consumer base to new segments and managing product launches and promotional activities. Through TSO, retailers can organize pricing decisions across the entire inventory and simultaneously plan assortment and promotion decisions.  Price optimization can lead to significant improvement in retail operations, by strengthening inventory management and producing faster responses to market dynamics. Price optimization is currently part of a growing number of retailers’ operations, from e-commerce giants Amazon and Alibaba to legacy grocery retailers such as Kroger and Harp Foods.  Rather than developing price optimization software in-house, the most efficient option for most retailers could be to turn to specialist data analytics vendors to develop a customized AI-powered price optimization platform, which enables retailers to integrate the platform into their operations right away and to save on costs and resources in the long run.  In the future, we expect the number of retailers using AI-powered platforms for price optimization to increase, in line with the dynamic growth forecast for the global AI software market. New technologies and disciplines — such as neuroscience — will be used in combination with AI to expand the consumer analytics capabilities available to retailers. About Precima  Precima is a global retail strategy and analytics company that provides tailored, data-driven solutions to drive sales, boost profitability, and build customer loyalty. Leveraging deep analytics expertise, Precima helps organizations improve their competitive position across all facets of planning and operations from pricing, promotional planning, assortment, personalized marketing, category and shopper insights, and supplier collaboration.   Precima's head office is located in Toronto, with global offices in Den Bosch (the Netherlands), Chicago and Boston area (US), and London (UK).  www.precima.com.