How to Use Big Data and Artificial Intelligence for Demand-Based Pricing in Retail

You can call yourself a guru of retail pricing if you can make the right pricing decisions for every one of your products, separately and combined, based on their demand elasticity at any given moment.

Each of your pricing decisions has to help you reach all of your current business goals and ensure the best shopping experience at the same time. In other words: to find a balance between your profits and traffic. 

All of this is possible thanks to demand-based pricing, which is the peak of retail pricing evolution. But to reach it, retailers need to enhance their pricing teams with the blend of big data and artificial intelligence. 

And here’s why.

From business strategy to profitable pricing decisions 

Let’s take a step back to where everything begins – your business strategy. The company’s strategy is supposed to be converted into its pricing strategy and subsequently into pricing tactics. That’s how a typical business strategy may sound like: “I want my company to grow fast and be profitable.” 

In this case, the pricing strategy has to ensure a price-enabled competitive edge in the market. Such a strategy is obviously market-driven and can be presented like: “I want to offer the lowest prices in the premium segment, OR I want to set by 10% higher prices than mass-market retailers.” 

But at the same time, this strategy may have the following specifications: “I am ready to sell my key-value items (KVIs) at mass-market prices, while offering up to 15% discounts for certain groups of products. I’ll also invest in high-level customer services and product quality to support my pricing offer.”

Once the pricing strategy is more or less clear, the most challenging part starts. It’s about the exact steps you need to take for every product in your portfolio and all the factors you have to consider when taking these steps. 

So, how can you get from “I want my company to grow fast and be profitable" to every pricing decision contributing to this goal?

Usually, the road to success looks like this.

Step 1. True competitors.

You should start with your price positioning in the market and knowing your true competitors. From my experience, some retailers tend to mistakenly consider certain companies their rivals, when in reality their sales are in no way impacted by these retailers. This way they waste their time and money monitoring wrong companies and eventually making wrong decisions. 

But even if you are lucky to know your real competitors, you are probably bombarded with the following questions: What is my optimal price positioning against these companies? Should I set prices that are lower by 15% or higher by 10%? Or it’s better to offer the same prices? How will it affect demand? Along with hundreds of other questions. In particular, if you manage thousands of products which need to be repriced regularly. 

Step 2. True KVIs.

Meanwhile, the most important question you need to answer here is what products you and your competitors have in common. That’s where you lock horns price-wise. That’s where your price is your weapon.  

You should start by defining your key-value items (KVIs), which drive your brand perception and demand. In fact, according to Nielsen, KVIs “drive more than 50% of shoppers’ perception of a retailer or brand’s overall value.” Here you set lower prices to lure customers from your competitors to your stores. 

But you also need to understand which of your products are profit-drivers (for example, exclusive items) to compensate for the margin losses of traffic-generating items (KVIs). What is the price premium you're entitled to charge for such products? How to calculate it? 

You need to have clear data-driven answers to all of these questions to succeed. Quickly.

Success = Big Data + AI + Team

To be able to do that retailers require three things:

• massive amounts of relevant data;
• advanced pricing software that can analyze this data and provide actionable insights with the necessary speed;
• a ready team.

Let’s move step by step.

Step 1. Big Data. 

Large retail companies have tens of thousands of products in their portfolios. What’s more, very often retailers operate in different price zones and reprice their products weekly, daily or even several times per day like Amazon. All of this translates into massive amounts of data pricing teams need to analyze before setting ever single price. 

But there has to be something to analyze in the first place. The whole process needs to start a little bit earlier — with data collection. As retailers are just beginning to use big data to optimize various areas of operations, from pricing to logistics, they usually have trouble with data gathering and storage. The data is often distributed among different departments, is insufficient or ill-structured, or has errors. 

So, retailers need to prepare data before using it to their advantage. They have to collect, structure, clean and store it in a single format and place.

The process may take several weeks or even months. Some companies prefer hiring a specialized data management provider to do everything for them. The good news is that you need to do it just once and use the data for as long as you need.

Big data is the first step of the digital transformation of your company. 

Step 2. Artificial Intelligence.

Let’s imagine all the necessary data is ready. But how can you use it? What’s the best way to analyze it? Some companies would hire a team of data analysts, but even they can fail. The speed of pricing decisions needs to be extremely high, while the decisions themselves need to be extremely accurate, which goes way beyond human capabilities.

That’s where artificial intelligence jumps in. AI algorithms can analyze enormous amounts of data the way you need and provide you with recommendations which will subsequently raise the probability of making the right pricing decision. 

The algorithms help you answer such questions as:

• What are my true competitors?
• What products do I share with them?
• What will happen if I offer this toothpaste by 10% cheaper than my rivals?
• What prices should I set for my exclusive products?
• How deep a discount should I set for these products to sell them before a deadline while saving my margins?
• Should I get involved in this price war?
• What should be my repricing steps during flash sales?

As well as thousands of other questions.

In this scenario, managers have the final say in defining the pricing strategy. What machines do is they offer a "sandbox" to test various pricing scenarios, as well as a set of steps to implement the best strategy and reach all the set goals. 

Step 3. AI supervisors.

But machines, even intelligent machines, still remain a tool. It is extremely sophisticated and powerful, but it is nothing without humans. And yet, AI is changing the role of pricing managers completely turning them into superintelligent humans.

Not so long ago, a typical day of a pricing manager would mostly consist of managing data and negotiating with vendors. Today, when machines are getting responsible for boring and repetitive tasks and data analytics, humans are given a chance of reaching the next level of their professionalism.

What’s expected from pricing managers today is their ability to work with AI-powered software, interpret its insights, course-correct it, as well as set and manage high-level goals. Thus, managers have time and opportunity to upgrade their business analytics and strategy-creating skills.

I think I may use the relationship between Tony Stark (Iron Man) and Jarvis (his AI-based software) as an example of the relationship between pricing managers and AI-led pricing software. Jarvis brings all the necessary information, while Tony makes the final decision.

Truth be told, transitioning to the new role may be somehow difficult for pricing managers used to a certain status quo, to things being done a certain way for decades. That’s why, as AI adoption is usually a top-bottom process, they need constant support from their top managers, as well as training and the ability to test the software and make mistakes fearlessly. 

Having rolled out AI-based pricing software at numerous retail companies, I’ve noticed that the period of adjustment is not that long. Pricing managers start appreciating the innovation once they see the first results, which is usually within 4-6 weeks. AI price optimization brings a double-digit revenue uplift and an unprecedented boost in productivity. 

Ultimately, demand-based pricing which takes into account the demand elasticity of every product to help set the right prices at any given moment is what every retailer is starting to aspire to. To benefit from it, retailers need three essential components:

• Sufficient and errorless data which is stored in a single place;
• Powerful AI algorithms to process the data quickly and provide actionable insights;
• Pricing managers who set relevant goals, supervise and course-correct AI-based pricing software and implement AI price and promo recommendations.