prudsys, Spezialist für KI im Handel: Dynamic Pricing, Personalisierung & Recommendations

Since the dawn of business, supply and demand have determined the price of goods. This principle has not changed in the age of digitization. What has changed, is the speed at which prices rise and fall. Intelligent algorithms support retailers in determining the optimal price. They determine the many pricing factors in real time and calculate the current fair market price for each individual product. In October, etailment magazine interviewed our CEO Jens Scholz about topics including intelligent pricing in retail. Read his responses here:

It is common knowledge that price is the main determinant in the marketplace. That is why there are a number of repricing solutions. How are you different in this respect?

Many of the pricing solutions in use today follow rigid pricing rules which are frequently based solely on the price behavior of the competition. The use of so-called repricing tools runs the risk of disastrous price distortions if retailers constantly try to outbid one another. Companies looking to get the most out of their pricing strategies are better off using AI-based methods to automatically calculate prices.

Our AI calculates the fair market price for thousands of products at any given time and in real time. The solution ranges from regular individual price optimization to markdown pricing right down to intelligent couponing. Intelligent algorithms automatically adapt prices. They continuously analyze huge amounts of data and include all relevant pricing factors.

When it gets cold, fuels are more expensive, when it rains, umbrellas are – AI is not that trivial. But how does AI determine the optimal price?

AI for intelligent pricing should be able to take into account many different influencing factors when pricing, e.g. inventory, current demand, product life cycle, day of the week, sales promotions, competitor prices and even the weather (which, by the way, happens implicitly via the demand). The sheer volume of factors makes it clear: In the age of big data and highly volatile market environments, it is no longer possible to effectively control prices manually. This is where artificial intelligence (AI) in the form of intelligent algorithms comes into play. AI automatically calculates the optimal price for each item at any given time.

Prices are then always optimized while taking into account the retailer’s pricing strategy. It is up to retailers to determine how much they want to intervene in the functioning of the AI. Generally, the retailer will at least set the upper and lower price limits for each product. The retailer will then also specify which parameter the AI should base the price optimization on – in other words, sales, turnover or gross profit. The algorithms take care of the rest.

The great advantage of dynamic price optimization is that pricing is not cost-driven. Instead, customer price acceptance becomes the focal point with this approach. The key issue is this: How much is a product or service worth to a customer at a given time in a given place?

[For more on demand-oriented pricing see this blog article.]

Our AI can track certain objectives using its own dynamics and can draw real-time conclusions as to the effectiveness of the pricing based on customer response. This way, AI maximizes the results potential of the retailer.

[Interested readers can find out more in our Whitepaper about dynamic pricing.]

Speaking of dynamic pricing: Do customers actually notice when the system is in use?

If the customer observes the price development of a product, he will naturally notice the price trend. Fluctuations in online trade based on supply and demand are normal. If the customer waits for the right moment there are real deals to be had.

Another example that customers probably do not associate with dynamic pricing: Couponing.
To retain loyal customers who purchase large amounts frequently, retailers like to give out coupons for their next purchase. When retailers combine price discounts (measured based on customer value) with products that the customer is really interested in (which brings us back to the topic of personalization), the retailer creates real added value for the customer and an incentive to purchase.

[For more on this please see our blog article entitled Instead of “Everything 20 % off”: Use Personalized Coupons to Increase Customer Value.]

Can multi and omnichannel retailers use such an AI solution even at a bricks and mortar POS? What would such a solution actually look like?

Yes. Here, too, there are various possible scenarios. The determining factor is the retailer’s strategy. Does the retailer want to increase turnover, carry out markdown pricing, increase customer value through intelligent couponing or do everything all at once if possible? An AI solution can optimize prices at the POS based on these objectives, e.g. using Electronic Shelf Labels (ESLs).

The AI recognizes, among other things, the price effects between various items. One example is the demand relationship between strawberries, canned whipped cream and sponge cake. If retailers want to stimulate the sale of strawberries in order to sell all of them before they go bad, they adjust the price accordingly. An AI solution can take into account both positive and negative demand relationships (cross-price elasticities) when optimizing the price and after adjusting the price of the strawberries also indicates the products, in this case canned whipped cream and sponge cake, affected by that price adjustment. The AI solution then suggests the optimal price for canned whipped cream and sponge cake.

[If you would like to know more about this, have a look at the example in our video or in the blog article.]

The topic of markdown pricing is particularly important, especially in grocery retailing: The AI designs the prices so that grocery inventory is at zero by a certain point in time. This means that considerably less food goes to waste.

[For more on markdown pricing in general and particularly in the fashion industry read this blog article.]

Order today, delivered by tomorrow? How long does such a project take on average, from the time it is commissioned until the retailer uses it for the first time? What difficulties do retailers realistically have to contend with?

The majority of our pricing customers opt for a PoC (Proof of Concept). How do turnover, margins and sales develop when using AI? Here we look closely at what the AI contributes in terms of added value. This phase generally lasts 2-3 months. The integration of the pricing solution itself takes place very quickly, generally in under ten days. Ultimately it depends on the circumstances and requirements of the customer.

It is important to keep in mind: Dynamic pricing brings with it an enormous increase in efficiency as the AI takes the repetitive and monotonous activities like pricing off people’s hands. This gives category management room to manoeuver, which can then be used for strategic pricing and product range decisions.

The possibilities of dynamic pricing are compelling. But isn’t it just for the really big players in the industry?

When it comes to intelligent price optimization, we are experiencing a booming demand in the market. Amazon changes its prices dozens of times a day. If you can’t keep up you are at a clear competitive disadvantage. The majority of our dynamic pricing customers are in the top 100 online shops in Germany in terms of trade volume. That does not necessarily mean, however, that using dynamic pricing is not worth it for retailers that are not in this category. In such cases we are happy to consult with individual retailers.