Price-setting used to be an art, a skill derived from years of industry experience, intuition, and hard-bought competitive business intelligence. In 2022, it has become a science. There are whole consultancies set up to master the art of dynamic pricing. A cursory Google search throws up dozens of specialist firms.
Take Capital Pricing Consultants, for instance, whose CEO Lydia di Liello leveraged a successful career in procurement to build a reputation for pricing expertise. Di Liello believes the tendency to think about pricing last is “completely backwards,” as she told Industrial Sage. “We need to be starting with pricing from day one, to say, what’s the market willing to pay for this?”
Clearly, pricing is something at the forefront of executives’ minds, in part due to the vast expansion of dynamic online pricing – a process whereby prices are altered seasonally, monthly, or even hour-by-hour, tracking patterns of supply and demand. Firms such as Amazon have become masters at leveraging AI to ensure their prices are maximally competitive.
Pricing in a Volatile Market
The COVID-19 pandemic helped accelerate a boom for online shopping. Ecommerce sales were up by 6% year-on-year by December 2021 according to Statista, with 84% of consumers reporting a shift in shopping behavior towards more online purchasing.
This surge towards shopping online created a more volatile sales environment for many businesses who suddenly found themselves competing in a world where supply and demand fluctuations created panic-buying patterns, opportunities for new products, and challenges for supply chains and fulfillment.
As a result, for any sales forecasting to be accurate, real-time, moment-to-moment data became vital, with prices having to change accordingly. The 2021 Global McKinsey Report on AI stated that over 74% of respondents saw corporate revenue increase as a result of AI adoption for marketing and sales, with 11% seeing an increase of more than ten percent.
Meanwhile, an increased focus on the bottom line in a highly competitive marketplace has led companies to dig deeper into their data and use AI to set prices to improve margins, increase market share and optimize the customer experience.
Pricing remains a delicate procedure. A 2003 study by Andreas Hinterhuber showed that, on average, a price hike of just 5% lead to a 22% increase in operating profits for Fortune 500 companies.
However, in the 19+ years since Hinterhuber’s study, schools of thought have changed, with some recommending making cost reductions to keep prices level in an inflationary market. With contradictory strategies available, you need data and lots of it. To interpret data at scale, you need the power of AI.
The Challenges of Price-Setting
Price is very much a function of perceived value. Customers only care about how a product will benefit them. “How much the customer is willing to pay for the product has very little to do with cost and has very much to do with how much they value the product or service they’re buying,” says Eric Dolansky, Associate Professor of Marketing at Brock University, Ontario.
Today’s inflation rates and supply chain issues make pricing even more challenging since prices must alter alongside increases in taxes, manufacturing, and shipping costs. Businesses are forced into the position of taking a reactive, rather than proactive, approach.
And while AI-driven solutions are popular and effective, they demand huge volumes of business intelligence data to derive meaningful actions. Getting hold of this at scale and securely mining it for insight presents a technological challenge.
Finally, being fully responsive to consumer buying patterns means your pricing strategies must operate 24/7, 365 days a year. The only way to make this financially feasible is to automate as much of the process as possible.
What were the Traditional Methods of Price-Setting?
Before the AI revolution, pricing was a highly subjective and manual process. The existence of so many pricing consultancies demonstrates that many businesses are still leveraging human insight as their primary tool. What do these pricing consultants use for data?
They operate on a mixture of market trends, anecdotal evidence from phone calls, an analysis of competitor pricing and a long look at the company’s go-to-market strategy. Sales executives “feel their way” into a fair price for their products, effectively.
Because everyone is looking over their shoulder at their competitors, this tends to flatten out pricing for broadly similar products. Within their own lines, pricing will be largely standardized (a medium-sized t-shirt will cost the same, regardless of its color scheme. Discounts will typically be offered at the same rate across product lines, in an online equivalent of brick-and-mortar discount sales.
Some Traditional Price-Setting Strategies
Let’s break down the most common options for traditional product pricing:
- COST-PLUS. You simply add up the cost of producing, marketing, and distributing a product and then add an acceptable mark-up.
- COMPETITIVE. Look at the direct competition and price accordingly. Generic “own brand” products will typically reduce the price from the average, while luxury brands add on a premium.
- VALUE-BASED. Using market research, pricing is set at what consumers believe your products are worth.
- SKIMMING. Setting a high price to begin with, then progressively lowering it until you hit the right balance of scale and margins.
- PENETRATION. Starting deliberately low to saturate the market and beat off competition, then raising the price once demand is established.
All these strategies are based upon specific objectives such as prioritizing margins, dominating the market, or establishing a luxury brand. It’s quite possible for companies to follow multiple strategies for different product offerings in the same year.
The Problem with Traditional Price-Setting Methods
Since many of the above methods derive from a particular philosophical premise or belief (“ours is a luxury brand” / “we need to destroy the competition”) they may not reflect what’s really going on in the market.
Put simply, you don’t have the data to make sensible decisions, and may be losing revenue that would only be available to you with a more agile approach. These methods suit a pre-digital age, not one in which trends can shift within hours and customer behavior can morph overnight, at the speed of an influencer’s tweet.
There are many reasons to ditch these traditional methodologies in 2022:
- Consumer behavior is rapidly changing, so prices must change quickly too.
- Rules-based methods are too static and inflexible.
- The vast resource of sales data is there to be tapped.
- Your rivals are already doing it.
- Data-driven insights are the only way to remain competitive.
In effect, companies are in an arms race with one another, except the weapons are increasingly sophisticated AI and mountains of business intelligence and sales data.
Leverage AI for Competitive Pricing
Having looked at the traditional methods of price-setting, let’s get to the 21st century and see how AI can help adjust pricing strategies for maximum return.
AI-driven systems research competitor pricing, sales figures, and market trends, then derive fact-based conclusions at scale, and at lightning speed. Here are just some of the data points they can look at:
- Historical pricing patterns.
- Production costs and trends.
- Seasonal Fluctuations.
- Customer Purchase History
- Competitor Pricing and Stocks
Some of these sources are historic in origin, while others are pulled real-time by IPDA (intelligent process documentation automation) bots and data scrapers.
The best AI systems will look both internally, into a company’s own data, and externally, via web crawlers and bots. To get a bit more granular about those sources, here are just a few of the data points that might be brought to bear on pricing decisions:
- Product Attributes: cost, margin ceiling, base price, MAP
- Inventory Levels
- Units sold and sales conversions
- Click-throughs and page visits
- Consumer Reviews, customer feedback
- Competitor Prices
- Supplier Costs
- Trends across Time (days of the week, holidays)
- Trends by Region
- Weather / Seasonal Fluctuations
Dynamic Pricing and AI
AI leverages large volumes of customer behavior data to discern their preferences, the products they are most interested in, the time of day they are likely to make a purchase, the maximum price they’ll pay, and more.
For example, an e-commerce site might use AI to track search preferences, social media posts, and activity on a competitor’s website. It might then trigger notifications to potential customers, at the right time of day and at the maximum price that each customer will be willing to pay.
The Advantages of Machine Learning
Rule-based algorithms are explicitly programmed, with actions based on simple if/then rules. They determine prices based on a limited set of influencing factors. They are set up in advance, then are let loose to set prices accordingly.
By contrast, machine learning algorithms are “trained” on an initial dataset which helps determine how they make decisions in the marketplace. They can make predictions of how various prices will affect sales, revenue, and profit margins. Their logic is not explicitly programmed in advance, so the outcome cannot be foreseen, and they become better and better predictors the more data they receive.
The benefits of moving to an AI and machine learning approach are many:
- An increase in sales volume and improved profits.
- Reduced customer churn with more appropriate offers.
- An improved customer experience overall.
- Customer loyalty with more repeat business.
How to Implement AI-Based Dynamic Pricing
First, you need access to larger volumes of data, to train better and better algorithms. Secondly, you employ domain-specific AIs which scour only key sites for pricing data. These AI know the difference between relevant and irrelevant data. They know how to process and analyze it for the specific purpose of price-setting. They will deliver their findings in an optimized format that is clearly presented and easy to interpret.
There are important caveats. For AI systems to scour trends, competitor sites, or buyer behavior they need to be trained on data sets that closely match each use case. AIs also take a little time to get up to speed, since they require an initial training phase within the relevant domain to learn properly.
Finally, data quality is highly relevant, since this is all your AI can base its recommendations upon. This is particularly true if you plug in automated pricing based upon your AI’s insights.
AI-Derived Price-Setting Examples
Here are three use cases from some of the leading adopters of dynamic pricing powered by AI:
Incredibly, Amazon changes its product prices on average every ten minutes (based upon 2.5 million product price changes per day). It can do this because it has billions of data points available to it, with incredibly well-honed AIs able to trigger rapid price changes according to fluctuations in market dynamics.
Among the factors Amazon analyzes are customer shopping patterns, competitor prices, inventory levels, profit margins, and untold other factors at its disposal.
Sometimes it will raise the price of less common varieties of products, making them more available versions seem like a bargain by comparison. This may be why you’ll often see different color variants of some apparel items available at varying prices.
Fortunately, you don’t need 200 million users and 1.5 billion items for sale to leverage effective AI pricing strategies.
Although highly popular in its sector, Uber is a minnow compared with Jeff Bezos’ behemoth. Nevertheless, Uber has enough customer journey data to adjust prices according to time of day, trip distance, rider to driver ratio, geography, and more.
Using AI to monitor and change prices has helped Uber keep up with supply and demand issues, incentivize its drivers to maximize revenue, and keep customers loyal to the brand. It has done so in a highly competitive and at times rancorous industry, where rivals’ jealousy at Uber’s profits has merely proved the success of its marketing and sales efforts.
The digital lettings agency which began life as a portal for renting out holiday homes has grown to an international brand with the assistance of AI-derived commercial insights.
Airbnb digs deep into their large pool of customer data and competitor lettings pricing to adjust their prices accordingly. Among the internal factors that AirBnB’s AI considers are:
- The days of the week the booking covers.
- Urgency (distance between intended booking dates and dates the user is booking).
- Proximity to special events/seasons.
- Historical performance of the venue.
- Number and quality of venue reviews.
By adopting a scientific approach, Airbnb has struck the right balance between pleasing customers by undercutting hotel prices, while remaining competitive for their lettings clients, and being highly profitable as a company.
All three of these examples show how a company can leverage the hidden resources that are sales and competitor data. The dynamic pricing strategies they adopt have made these three companies the brand leaders in their specific sectors.
Dynamic Pricing is the way forward for companies wanting to maintain a competitive edge and a truly 21st-century product pricing strategy. AI and machine learning, such as that provided by Evalueserve is by far the best way to utilize the data you already have at your fingertips, and price your way into market dominance.
Check out our data analytic services and AI products here.