With pricing being a touchy topic across nearly all industries, automated pricing in ecommerce becomes even more sensitive. Companies often dismiss re-pricing: why apply AI when something more familiar works just fine? Well, for starters, because your e-shop can perform on an entirely different level. What if you’ve been using only a fragment of your selling potential all this time? Like many things in life, you won’t know unless you try.
Many terms, similar meaning:
- Automated pricing
- Optimized pricing
- Dynamic pricing
- Price intelligence
- Machine learning
- Deep learning
Generally speaking, should you encounter any of these terms in ecommerce, note that they all refer to a similar strategy based on which retailers change the price of the product based on the current situation on the market. Although dynamic pricing is often understood in correlation with machine learning and AI, retailers use a wide spectrum of pricing strategies: from having their buyers track the prices daily, to more or less sophisticated Excel sheets or other in-house invented systems, to deep learning pricing tools usually provided by niche companies to e-shops with a monthly fee.
Will comparison kill us?
When asked how they design their pricing strategy, most ecommerce leaders are fast with an answer: We have it designed already. It works just fine. Sure, but what about re-evaluating this strategy and re-pricing it based on a variety of crucial factors, from what season it is to who the competitors are? “Many companies have three or four price ranges with a certain margin added to them,” Petr Blaha of Dataweps says. “For instance, anything with a price tag from 2 EUR to 6 EUR has a 20% margin, products from 6 to 12 EUR have 15% and anything above that would have 10% added to the price. This is of course an example to showcase a frequent behavior, not a real functioning model.”
For some, this works just fine. For others, it does not. But regardless of an online business‘s perception of its own performance, dynamic pricing has forever stirred up waters by imposing a rhetorical question: How can I say I am successful when I have no idea what the potential of my business really is?
“Ecommerce companies have no anchor. What would happen if the price went up by 1 EUR for this particular product and went down by 10% for these other products? Would you sell more or less? Would your competitors change their price as a result of your behaviour? What would that mean for you? What would happen if your competitors ran out of a certain product and you still had it in stock? There are countless scenarios and no one has the time and capacity to answer them all day every day,” Blaha explains.“Pricing strategy is never really finished. It requires constant evaluation and re-design, something that is way easier for robots to do and much harder for humans to implement. But harder does not mean impossible.”
Who is it for, really? “If you have a unique brand or unique products (e.g. fashion), you definitely can (and should) use dynamic pricing. It’s even better to use dynamic pricing when you don’t have competition because your visitors don’t have any price anchor. So it’s easier to increase the price, as visitors compare it only to their emotional perception of the product’s value,” Blaha explains. “Dynamic pricing isn’t viable in some segments, though. For example, sometimes distributors don’t allow it, or clients are extra sensitive to price changes (e.g., with some pet food). Or you’re a wholesaler and you don’t want to undercut your products.”
Where to start and what to expect
“While optimizing marketing strategies can bring you up to 20% more sales, optimizing pricing can put you on a completely new level.”
Blaha, consultant in Dataweps, bases his argument on a recent experience with AI increasing e-shop’s revenue by 240%. “Try achieving that with marketing only,” he challenges. From his experience, while everybody is impressed by the potential, many hesitate with implementation. Why? “Fear of the unknown, simple as that. Most founders, managers or buyers are increasingly familiar with the importance of dynamic pricing, yet they are afraid they won’t be able to control the prices now – robots will,” Blaha adds. Many ecommerce businesses, including market leaders, haven’t really touched their pricing strategies for quite some time, so questions like: What if it doesn’t work? Or, what if it does? How do we cope with the growth? are valid.
Starting with a segment is usually the answer. “Pick a segment out of your portfolio where you can implement automated pricing and do it in a three month span. Three is best, because it takes some time to gather data, re-evaluate and adjust again. After a designated amount of time, look through and compare the results of the whole segment to balance out all external factors like season etc.,” the pricing expert claims.
Implementing dynamic pricing step by step #bestpractices
Let’s take, for example, an outdoor fashion store. What would some of the best practices be?
- Apparel you produce and distribute through other stores. You don’t want to re-price this segment aggressively to prevent the prices from being pushed too far down as well as to be able to guarantee refunds.
- Apparel you produce without distributing it further is a great segment to start re-pricing using machine learning. When optimizing prices, you don’t need to take into account your competitors since your apparel is strong in this segment.
- Latest collection (seasonal clothing) not subjected to strict pricing by the manufacturer is another opportunity to use dynamic pricing. Opt for a business model that maximizes the margin in combination with selling out your inventory.
- Introducing dynamic pricing to older collections is a smart move, too. Pressure to sell out is lower, which creates space for maximizing the margin without trying too hard.
- Once the inventory is full or needs more space, try slightly different tactics in a certain segment. Try pushing revenue increase while keeping the original margin.
- Distinctive segments such as price anchors or sales triggers do not need dynamic pricing. Try re-pricing these according to your competitor’s prices. It is actually beneficial to keep the price of some items down so you can sell other goods or make your product portfolio look more affordable.
CEE customers: Driven by price?
Customers everywhere, and in Central and Eastern Europe more so, are price-sensitive. In the United States, Amazon is used as the number two search engine after Google to look for goods and compare prices in one click. The only difference between Americans and Europeans is that the latter use local marketplaces or price comparison sites. We all compare. According to a PWC survey, 61% of customers visit an e-shop’s website to compare pricing. “An average online shopper will visit at least 3 websites before making their purchase, and also 86% of first time online shoppers say it’s important to be able to see and compare prices from different sellers,” Oberlo informs.
This is especially true in regions such as Hungary, the Czech Republic, Poland and even Romania and Bulgaria. With that said, it’s no wonder that bidding strategies have gotten quite popular. Dynamic pricing is expected to influence online retail positively, both on macroeconomic (clearing the market from phenomena such as underpricing) and microeconomic levels (each company is well aware of its strengths and where to focus). “The CEE region is learning that lowering prices below the competitor’s price is creating a downward spiral that no one truly benefits from. Not the market, not the customer and certainly not the seller,” the pricing consultant explains.
The 2 most common pricing strategies in Central Europe
- A REACTIVE APPROACH where you mostly change prices in reaction to your competitors.
- A PROACTIVE APPROACH where you use AI to adjust your pricing and business strategy regularly according to a number of variables, competitors being one of them. Your aim is marketability in a certain price range. For instance, imagine you sell a certain type of tea cup for 3 EUR. The dynamic pricing tool raises the price to 3,5 EUR for a week to collect a critical amount of data. After the system evaluates whether 3 or 3.5 is better, it then starts calculating a new optimal price. It is also constantly checking for marketing anomalies to bring the best result in terms of profit.”
Both strategies are valid; which one you opt for depends on your product strategy. “Some products, especially electronics, perform best when compared with competitors,” Blaha explains.
The future of pricing in Europe
With more and more retailers implementing machine learning in their re-pricing process, it is less likely you would measure up unless you innovate your pricing strategy, too. “Companies like Amazon have embraced dynamic pricing. Not only does the company alter its prices more than 2.5 million times a day, but it also changes the price of around 20% of its total inventory every day. Wal-Mart can, by comparison, currently change the price of only 50,000 products per month, Future Agenda informs. “There is no going back, and the industry will have to adjust. There are several interesting facts and trends about optimized pricing and machine learning in the CEE region in particular, namely:
1. Product segmentation
Most online stores do this already; however, the trend is going to rise with AI powered dynamic pricing, offering more sophisticated mechanisms to customize the inventory by segmenting the customers on the basis of product choice and thus proffering different prices to them – this is dynamic pricing. In CEE countries, major local actors are implementing machine learning mechanisms to segment products by dynamic pricing, or developing their own solutions. In our talk with Blaha of Dataweps, we found one more interesting bit: “The make-or-break situation of machine learning is to have a sufficient amount of data to work with. We are talking about tens of thousands of sales, which is not achievable for certain exclusive brands. When you sell premium cell phones for 500 EUR each, there is a high chance your data won’t be sufficient enough to test dynamic pricing,” claims Blaha, adding that product segmentation goes both ways: “It would of course depend on many factors, but you can overall cluster segments together to form a bigger segment.”
2. Customer segmentation
Have you noticed that airplane ticket prices differ significantly when searching from an iPhone compared to, say, a Samsung device? “Customer optimization is big in many industries and it is only logical that online stores should use it further,” Blaha explains. The area where most of us have experienced dynamic pricing most directly has been in booking airline tickets, Future Agenda declares. As well as, allegedly, showing higher prices to Mac users than PC users, Amazon monitors customers’ behaviours to determine the best time to raise or lower prices to get the sale, and amend prices by what else is already in a customer’s basket. “Again, when faced with insufficient data, stores can always cluster segments. When you, for instance, have a female customer over 30, living in Prague and shopping from a Mac, you can – and again, it very much depends on the situation – first erase “female” as a variable. Then, when your sample is still not sufficient enough, try cutting out the age and the city. The sole data point that he or she shopped from a Mac is enough of an indicator for you,” the pricing expert sums up.
Will there be a stronger tendency towards product or customer segmentation? That always depends on the unique situation of the company. Typically, food delivery businesses will lean towards segmenting customers whereas marketplaces will tend to the side of product segmentation. We should nevertheless expect companies to adopt both.
With these exciting trends dominating Europe’s ecommerce scene in the next couple of years, let’s still bear in mind that however innovative, dynamic pricing tools are just that – tools. Who you are as a company, how strong and trustworthy your brand is and how you care about your customers will, even in the era of growing marketplaces, play a crucial role in your success. Machine learning or business intelligence is here to see opportunities and make decisions.
Make sure that once you walk through these newly opened doors, you walk prepared.