By Heinrich Müller, CEO at Aimondo Pricing Solutions
Even if you’re far from the IT world, you hear news about Artificial Intelligence (AI) 24/7 now. And, let’s be honest, not all of the stories are inspiring. For instance, new Adobe products can detect Photoshop use, but does it matter in a world with AI-generated faces? Your kids can’t memorize certain English language rules. But who cares if people submit their PhD thesis written by ChatGPT?!
The only positive thing in this apocalyptic picture is that AI can be extremely helpful in your line of work. How exactly? No one but machines can gather, analyze, and apply insights from huge amounts of customer data. One area where AI has proven to be particularly valuable is in the realm of pricing data analysis. By leveraging AI algorithms to process and analyze large amounts of data, retailers can make more informed pricing decisions, resulting in improved profit margins and reduced costs. Let me give you an example from our recent experience with a Germany-based DIY retailer Kömpf. It has managed to increase its profit margin by 5% using AI-led pricing data analytics (more on this in a detailed case study here). Another customer we worked with showed incredible results in cutting their inventory holding costs by 38%.
Why some genius pricing strategies fail
Here’s why it works. The traditional approach to pricing in retail involved setting a fixed price based on factors such as cost of goods, competition, and desired profit margin. However, over the years, most market leaders have realized that this angle can be overly simplistic, failing to take into account the complex and dynamic factors that influence consumer behaviour and market trends. Remember the textbook example of the epic failure JCpenney went through when they introduced their infamous discount policy in 2012? It was innovative, reasonable, and was supposed to work like a charm because nobody else had done it before. Instead of “complicated and confusing” coupons and sales that only made customers hunt for discounts, they offered an “everyday low prices” strategy, aimed to simplify the shopping experience for customers and reduce the reliance on promotions to drive sales. It never worked. Instead of feeling happy and relieved, customers considered “everyday low prices” to be suspicious. They assumed the company wanted to get rid of some counterfeit goods and thus, offered ridiculously low prices. Sales numbers dropped, and JCPenney lost millions on this experiment. As we can see, in this case, a smart pricing strategy should have also taken into consideration not only product price sensitivity but also customer’s perception of value as well as market price sensitivity.
X10 your brain power
Let’s be honest, nothing is more suited to handle these multidimensional nonlinear predictions based on numerous data sources than AI. It beats an old good Excel sheet (even accompanied by a Solver) in all departments. AI can take all your data and spit out the solution, as well as outline possible outcomes of every step you will take. It will optimize for less risk, higher margin, and whatever personal KPIs and business goals you have in mind for this quarter. It not only enables retailers to make more informed pricing decisions but also builds better demand forecasts and highlights actionable insights that you would never have noticed on your own using the method that, sadly, is still the most popular across the industry: “eyeballing the data”.
Retailers have departments of data analysts, yet they still admit that AI-based pricing tools are extremely helpful in their line of work. For example, Kömpf tasked its data scientists to write custom lines of code within our dashboard. It has taken the impressive possibilities the Aimondo Platform already offers and has taken them to a completely different level, building a 100% customized product that serves its unique goals and needs.
From analysis to action
However, as we all know, ideas (aka analysis) come a dime a dozen. It’s all about how accurately we can implement the results of the complex data analytics AI solutions provide. And here we can look at the industry best practices and market leaders. Amazon has been using AI algorithms for years, not only to get a 360-degree market view and build a competitive context within the platform but also to dynamically adjust prices in response to market trends and competitor pricing. Amazon’s pricing algorithm, known as “Amazon Repricing,” uses a combination of real-time data on competitor pricing, historical pricing trends, and customer behaviour to determine the optimal price for each item. As former Amazon executive Nadia Shouraboura explains: “AI allows us to process huge amounts of data and act on the received insights in real-time.”
Another example of how AI-enabled pricing can be used to fuel long-term growth is fashion retailer Zara. Zara uses AI to analyze all incoming data sources (including POS-collected data and the results of marketing surveys) to maximize profitability while also maintaining a competitive edge in the fast-paced fashion industry. As Zara CEO Pablo Isla explains: “We use AI to help us make informed decisions about pricing, inventory management, and other key areas of our business. By analyzing vast amounts of data, we can identify patterns and trends that allow us to make smarter decisions and stay ahead of the competition.”
What if I’m not a good fit?
Sometimes we at Aimondo receive requests from companies that want to leverage the power of AI to stay ahead of their competition but are objectively not the best fit for the complex solutions we offer. These are companies with a very limited range of goods (most DTC brands) or companies that sell incomparable items (such as artwork). If you fall into this category, the honest answer would be, “No, you don’t need AI to analyze your competitive environment, consumer behaviour, and market trends.” Most of it can be done using very simple code snippets in Python or even good old Excel spreadsheets. AI is not a magic wand. It simply can’t be replaced by anything else when it comes to really big data. When the range of your data sets is very limited, you can skip the complex stuff and use a standard data analytics toolset and formulas. They will get you to the same sound and clear business insights market leaders might get with the help of AI.
In response to these numerous requests, I am writing a book, “Pricing Data — New Equation: A Data Analytics Playbook for Retailers and Manufacturers That Can Still Get Away with Using Excel.” If you believe this is something you can apply to practical use, please feel free to visit the book page and request your free copy. Otherwise, good luck with surviving in today’s AI-led world that, hopefully, seems slightly less apocalyptic to you now.
