To maximize profits, companies have always adjusted their prices depending on demand and timing. But traditional price optimization methods are working less and less well online. A new and better alternative is machine learning-based pricing.
Setting optimal prices and maximizing profits while avoiding discouraging customers from buying – this has always been a key challenge for companies. Until now, retailers have mainly used traditional methods for price optimization: employees analyze customer and market data and use simple mathematical models (e.g. linear regression) to calculate how price changes affect willingness to pay and profits. They then store rules in price optimization tools, on the basis of which prices are automatically adjusted. To ensure that no sales are “given away,” employees must repeatedly check the rules against current market data. This costs a lot of time and money – and is usually only worthwhile for large companies.
However, digitization has dramatically changed price optimization in recent years. On the one hand, it is a help because it provides companies with more and more precise data on which to base their decisions. On the other hand, the sheer volume of data makes it increasingly challenging to evaluate it correctly and continuously.
Machine learning (ML) comes into play at precisely this point: machine learning based algorithms are able to analyze much larger data sets and take into account significantly more variables. While in traditional pricing, employees work out the pricing rules in a manual data analysis, the machine learning model is given a sales target. After a short training period, it determines the best prices fully automatically, much more precisely, and at a fraction of the effort.
The market conditions in which companies must compete are becoming increasingly complex. Traditional price optimization can no longer adequately evaluate all of the diverse data points. Companies therefore have only two options: retreat to cost-based pricing and give away enormous revenue potential, or adopt machine-learning-based pricing to let smart algorithms calculate the best prices.
Machine learning is one of the key technologies for companies to leverage their growing data sets to create value. Price optimization is an effective use case in this regard.
Machine learning-based pricing uses complex algorithms that measure price elasticity to determine changes in demand at different price points. One strength of the machine learning approach is that the model recognizes patterns in historical data and can evaluate many data sources independently, without being explicitly programmed to recognize specific patterns.
Examples of data that can be included in the evaluation are:
Combined with forecasting algorithms, machine learning models can predict the effects of price adjustments on profit and revenue. Machine learning-based price optimization can respond to changes in data sources automatically and much more accurately than static models of traditional price optimization.
What predictions can machine learning-based price optimization make?
In retail, it is mainly e-commerce companies that have been using machine learning-based pricing so far. It is not possible to say with certainty how widespread machine learning based pricing currently is. However, according to a study by Business Insider, 72 percent of retail companies plan to invest in AI and machine learning in 2021 – presumably also in price optimization.
Corporations in particular are already harnessing the power of machine learning. These include well-known brands such as fashion retailer Bonprix, U.S. electronics company Monoprice and British supermarket chain Morrisons.
Some other international brands which are known to use machine learning-based pricing include:
The fashion company determines its entry-level prices via AI, and lets prices react to trends in an automated way. As a result, Zara only has to sell 15 to 20 percent of its products at discounted prices, according to Ghemawat and Nueno, as opposed to 30 to 40 percent at other European retailers.
Ralph Lauren – and Michael Kors – are using machine learning to sell fewer garments via markdown prices, manage inventory more efficiently and increase sales.
Fashion discounters are known to use machine learning to achieve their business goals despite low entry prices. For concrete statistics and potential successes through machine learning-based pricing, read our Case Studies.
First it’s the multinationals and innovative e-commerce startups that are adopting new technologies, then the rest of the industry. Machine learning-based pricing is well on its way to becoming mainstream in retail. According to an IBM study, 73 percent of companies surveyed plan to optimize their pricing and promotions through smart automation before the end of 2021. But how exactly does it work now?
To develop a machine learning model, different types of data, structured and unstructured, are needed. In the context of price optimization, the database might look like this:
Transactional data: List of products sold at different prices and buyer information. Product descriptions: Data on each cataloged product (category, brand, size, color, etc.), product photos, and manufacturing or sourcing costs
Customer testimonials: Customer reviews and customer opinions about the products sold.
Competitor data: Competitor prices for comparable products.
Inventory and delivery data: Data on inventory utilization, orders, and deliveries.
Not all information is necessary or available for every industry or business. If a retailer is new to the market, for example, it may not have access to customer testimonials. In addition, companies are rightly very cautious about using personal data. However, since the machine learning model is always adapted to the individual case, it can deliver good results even without these data points. For example, personal data is not even necessary for price optimizations at the product level.
The collected data must then be cleaned of errors and prepared for further processing. This step is challenging because data of different formats from different sources must be merged. The task should therefore be performed by experienced data scientists who ensure that the data is correctly and completely transformed into an algorithm.
The next step is to train the machine learning model. First, the model analyzes the variables and determines possible effects of price changes on sales. In doing so, the machine learning model independently detects correlations and patterns that human analysts easily overlook. These are incorporated into the algorithm for calculating optimal prices and form the basis for forecasts of sales and profits.
The initial model is tested in practice and can be manually optimized on a regular basis. With each correction, the algorithm learns and also improves its results independently. Additional data sets can be added to further optimize the accuracy of the algorithm. Over time, the training effort decreases while the effectiveness of the tool continuously increases.
However, the development of such machine learning-based pricing tools is costly. In-house development is therefore not recommended for most companies – especially since it is difficult to find the right personnel with the required combination of pricing and data science know-how. Outsourcing pricing optimization leads to a better business case in most cases.
A machine learning model can define optimal prices for specific business objectives and determine price elasticity for thousands of products within minutes.
Marketing and product teams can use these calculations to experiment more boldly with entry-level prices and discounts because they can better assess the potential impact on sales and demand. Instead of relying on gut instinct and experience, they can reason based on the machine learning algorithm. This gives them room to maneuver, which usually translates into increased sales and profits.
Price optimization is not new: many companies maintain traditional, static models to optimize their margins. But that approach has weaknesses. With the help of machine learning, these inaccuracies and errors can be avoided.
Traditional price optimization works on the basis of simple mathematical formulas that are no longer appropriate for today’s complex market environments. Moreover, results can only be as good as analyst-defined price adjustment rules.
Time and again, employees overlook critical developments or misjudge the significance of variables – with the result that price optimization falls far short of its potential.
Not so with machine learning: machine learning models recognize even non-obvious correlations, can evaluate data more accurately, and their susceptibility to error is drastically lower than that of human pricing managers.
Especially in e-commerce, companies usually manage large assortments, with products of different categories. Traditional price optimization tools cannot control adjustments finely enough, so that automatic changes negatively affect sales of individual products, or price automation triggers costly manual readjustments. Machine learning tools are able to control prices individually and make changes that are not only assortment-wide or per category, but at the product level.
Price reductions are a popular way to increase sales and rid warehouses of older products. However, those who work with flat markdown prices à la “25 percent off the entire assortment” reduce many products that could still have been sold at the regular price – wasted profit. Traditional price optimization works on the principle of bulldozing, whereas machine learning intervenes with surgical precision.
Traditional pricing evaluates predetermined factors according to certain rules. If mistakes are made with individual factors, if they are evaluated too strongly or too weakly, this can trigger unfavorable chain reactions. The result: loss of sales.
Specifically, this can lead to an approach based on price rules not reacting aggressively enough for highly price-sensitive products. The rule-based approach also fails to recognize this potential because it does not deviate from static rules to protect margins on individual products. Another example is that the traditional application lowers the price because it detects price reductions at the competitor. This is logical, but not all products are sensitive to competitor price changes. As a result, revenue potential is wasted here. Machine learning tools work based on business goals and “know” how price changes affect them.
Correctly assessing customers’ willingness to buy is necessary for a competitive pricing strategy. Traditional pricing tools are not able to make reliable predictions for this. Machine learning-based price optimization can accurately infer how price elasticity will evolve from big data variables and thus maximize profits.
The price elasticity of a product is affected by many factors that are not static. This is a weakness of traditional optimization tools: The database and pricing rules must be manually adjusted on a regular basis to reflect current market and competitive developments. Changes in corporate strategy must also be manually incorporated into price optimization.
Machine learning-based pricing works much more autonomously and updates its own rules based on identified pattern changes in the database.
Yes or no? According to a PwC-Studie , for 60 percent of consumers, price is the deciding factor in their purchase decision. Machine learning-based pricing makes it many times easier for retailers to define optimal prices and also offers other benefits.
Companies that use machine learning to calculate their prices based on historical data and evaluations see a quick return on investment. That’s because ML algorithms can identify short-term market opportunities at the product level and execute price adjustments quickly, helping retailers increase profits.
Traditional price optimization tools are more cumbersome: employees have to manually check market data at regular intervals and convert it into pricing rules. To keep the effort economical, only longer-term and seasonal developments can be taken into account here. Identifying short-term market changes for individual products is not possible in the classic approach.
Machine learning-based price optimization enables very precise prediction of the price effect on profit and revenue, so that companies can adjust their prices according to their business goals – without manual intervention. A one-time definition of business goals is all that is required, and the algorithm uses historical customer data to determine how to adjust prices to maintain the best possible margin in line with business goals.
If companies are planning special promotions or considering adjusting their stored targets, they can have the machine learning-based price optimization simulate the impact in advance: how does a price increase change the propensity to buy? What is the impact of a 25 or 15 percent discount? What happens if we always keep prices at the level of a given competitor (price matching)? Valid forecasts significantly reduce the risk of making the wrong strategic decisions.
To achieve optimal prices, companies have to analyze millions of data points and take various factors into account. Today’s product ranges are so extensive that manual evaluations can no longer be accurate. At the same time, markets are changing faster and faster. Analysts can no longer keep up with this pace, which is human, but costs valuable revenue.
Companies that use machine learning in price optimization will always be three steps ahead of their competitors because their analytics are more accurate and free of human error. The MIT Sloan Management Review points out that automated pricing allows price optimization for significantly more products than is possible with traditional pricing tools. If you don’t want to lose sales to the competition, you can’t avoid automation sooner or later.
Who offers the product at the lowest price? According to a Forbes-Studie, 60 percent of consumers buy from stores with the best prices. However, depending on trends and seasons, willingness to pay differs significantly. With machine learning-based price optimization, companies quickly identify when changes in competition or demand are emerging.
The application acts autonomously based on the company’s pricing strategy and is thus always faster than human market observation can be. With machine learning, companies are no longer surprised by price cuts from the competition, but can actively exploit market opportunities themselves.
It is difficult for humans to predict which prices are marketable, and even static models regularly fail. As with the weather, things can change seemingly out of the blue. Machine learning in price optimization ensures that companies are less likely to be surprised when “the weather changes.” With its forecasts, the application provides reliable future prospects so that companies gain the greatest possible control over prices and sales.
Particularly in times of economic crisis, machine learning-based pricing helps to avoid excessive price adjustments and to adjust prices in such a way that sales can be maintained or sales opportunities exploited despite adverse market conditions. In this way, automated pricing tools contribute significantly to the resilience of companies.
Traditional price optimization works according to rules that have to be checked and updated by employees at intervals. Whether they work better afterwards than before, however, is not guaranteed.
Machine learning determines sales trends and automatically adjusts prices in line with business goals. In the process, the model learns from its actions, so that it delivers better and better results over time in every case.
Traditional price optimization tries to define prices that pay into business goals through market observations and simple derivations. Intuition and experience of pricing managers play a role that should not be underestimated. However, pricing rules are often too schematic and imprecise.
Machine learning models take intuition out of the equation and determine optimal prices for changing market conditions based on data from various sources. In doing so, they take strategic goals into account. Should certain products be excluded from price discounts? Should prices in individual categories be regulated at certain thresholds? Machine learning pricing works goal-based and optimizes for each product.
Machine learning is a new technology, but it is rapidly establishing itself in the enterprise world. Price optimization is one use case where machine learning has already proven its worth. After the global players, smaller retailers are now following suit and introducing machine learning-based price optimization. This is because in the increasingly complex and fast-paced market conditions, manual pricing is reaching its limits. Previous mathematical models are too simplistic, and human intervention makes predictions prone to error.
Companies that forgo the support of machine learning in price optimization will feel competitive disadvantages in the foreseeable future. This is because the new technology works much more reliably and significantly faster.
Why is it worth investing in smart price optimization tools? The applications are an easy way to increase sales and profits without having to question fundamental pillars of strategy and offering. They deliver maximum results with minimum investment. Meanwhile, it is no longer necessary to develop machine learning models from scratch. Modern tools like 7Learnings’ pricing solution make it affordable for any company to get started with machine learning-based pricing.