Predictive Price Optimization With Artificial Intelligence

Setting prices can be a challenge, in particular in industries like travel and transportation where prices may change almost hourly, constantly adapting to changing market conditions more or less in real-time. Can AI come to help with predictive pricing? As it turns out, yes.

In this blog post, I interview Mr. Alex Shartsis, the cofounder and CEO of Perfect Price, an artificial intelligence company focused on solutions for pricing. He leads product development as the company works with innovative travel and transportation companies ranging from the newest startups to the largest companies in the world.

Now, over to the interview!

Alex Shartsis, CEO of Perfect Price

What is Your Company’s Background?

The founding team experienced the challenge of setting prices first hand, especially at TripIt, a global company in the travel industry. This challenge is growing as commerce becomes more digital and dynamic.

Today, consumers accept and even expect price fluctuations. Frequently they are able to time them better than the companies setting the prices – costing companies billions annually. 

It seemed clear that there should be a simple service to automate the complex data science required to make dynamic pricing possible for any business. We started Perfect Price in 2013 to do just that, and launched our first solution in 2014.

What Problem Does Perfect Price Solve?

Perfect Price artificial intelligence powers next generation dynamic pricing and revenue management.

What Does Perfect Price do, and why is it Important to Your Customers?

Perfect Price saves money for its customers while simultaneously saving them time. It does this by being a systematic way to address dynamic pricing that is more flexible, adaptable and therefore more accurate than either humans or legacy systems.

Key problems Perfect Price solves are:

  • Adapting quickly to rapidly changing market environments
  • Managing utilization and profit to ensure no money is left on the table when demand spikes or drops off
  • Growing the business profitably (or unprofitably) in a predictable and repeatable way, in order to increase market capitalization or shareholder value
  • Enhancing usually understaffed and overwhelmed data science teams with a complete solution to the pricing function

Core platform 

Perfect Price leverages a core AI platform based on a number of the latest AI technologies, including supervised machine learning, reinforcement learning, clustering and more. This platform generates a demand function that is accurate down to the micro segment level. It then can calculate an optimal price based on business rules and other configurable parameters, in order to optimize the customer’s objective function–profit, revenue, etc.

Key technologies

At the core of Perfect Price is micro segments. Micro segments are much more granular than traditional segments, enabling a much more accurate fit between prices and demand. A traditional segmentation approach for a car rental company might look at people picking up in the morning versus the afternoon. It might decide to boost morning prices, because these people are more likely to be business travelers who are assumed to have higher willingness to pay.

The Perfect Price approach will set every car class and location as its own microsegment, and can separate days into as many as 24 different additional microsegments.

Thus it can tell if a certain car class at a certain location on a certain day shows higher demand, thus increasing the price much more accurately, while possibly leaving unaffected car classes alone. It can do this even if this demand pattern is only at 1 location for 1 car class out of tens of thousands of possible combinations.

Perfect Price uses other core AI technologies on its platform to increase the stability of prices, adjust prices as supply becomes more scarce – or demand does not arrive on plan and oversupply becomes a problem.

Core products

The three main products customers use are the dynamic pricing manager, simulator, and predictive analytics. The dynamic pricing manager enables companies to automate their pricing, while retaining as much or as little human oversight and control as they desire.

The simulator enables companies to backtest models, both before they implement Perfect Price and to test new models against existing models to maximize the likelihood that a new model is better than a prior model.  Predictive analytics enable companies to look at “what if” scenarios, such as promotions, price changes, or other tactical or strategic moves.

Business use cases

Customers use Perfect Price to manage dynamic prices. Unlike legacy solutions, which separate modeling and price management, and require very time consuming exception management and rule management, Perfect Price’s AI provides a comprehensive and highly accurate solution that can run with as much or little human intervention as the customer desires. 

Use cases include

  • Global rental car company managing thousands of locations and changing tens to hundreds of millions of prices every day.
  • Global ground transportation company operating at ~100 airports managing pricing on dynamic routes while competing with Uber, Lyft and other disruptive competitors.
  • Global airline managing prices of its ancillary products such as bag fees and seat upgrades across thousands of planes, flights and customer segment combinations.

How Does Perfect Price use AI?

We use a combination of many technologies mentioned above, including;

  • Supervised machine learning
  • Reinforcement learning
  • Clustering
  • Linear regression
  • Classification

And much more. We have patents in process on much of our product so we cannot talk in depth about the technologies in use at this time.

In Your Mind, What is the Future of AI in Business and Marketing?

Marketing gets most of the focus for AI – but business is where AI will have the greatest impact. In 2007, 0% of digital ads were bought by AI. By 2017, according to emarketer, 78% of them were (and we were a big part of that transition, at Drawbridge). So marketers have been using AI – whether or not they realized it – for quite a long time. Its use case in businesses is just emerging.

Amazon has been crushing its competition largely due to both its deeply held cultural belief in focusing on data, and the widespread use of AI technologies. Orders are predicted. Inventory is reordered. Rather than focus on the mistakes, Amazon focuses on improvement. Getting better. Turning everything possible over to AI.

And it works.

Most companies we talk to think about AI’s mistakes, all the things that might go wrong. That is a positive change – because a year or two ago, they didn’t think about AI at all. As they outgrow this “can’t do” phase, and start to see what AI “can do” they will more rapidly adapt it, phasing humans out of rote mundane and error prone tasks–like setting prices–and refocusing them where human intelligence is actually needed.

Do You Think Marketers Will be Replaced by AI Robots?

Our AI doesn’t work for marketers – it runs the business. I believe that humans will always have a major role to play in marketing, because while tactics might be well handled by AI, we are decades away from AI inventing winning strategies.

But where the core problem is a math-based optimization problem – serving digital ads, setting prices, optimizing routes – AI will definitely win.

Think about one area in the real world where AI has already made dramatic change to daily lives. Would you trade Uber or Lyft for a taxi dispatcher with a phone and radio? Did a taxi company ever send you a taxi in less than 2 minutes?

Only AI routing and optimization in real time can be that good – and it makes a massive difference in how you interact with that service. The same will be true in countless ways, that we don’t even know about yet.

In this interview, Mr. Shartsis gave a very fascinating insight into dynamic and predictive pricing, and how artificial intelligence can be used to run and improve business in the future.

In fact, I think it is safe to say, the future is already here!


I am an author, speaker and consultant in marketing automation and artificial intelligence.

Do you need help with marketing automation or AI-based solutions? Contact me and let’s discuss how I can help you!

Magnus Unemyr

Author, speaker and consultant in the aras of marketing automation, artificial intelligence, and the Internet-Of-Things. Contact me if you need help! Learn more.