The importance of using AI responsibly

At Datasparq, we regularly help businesses transform their operations using AI tools such as reinforcement learning—a technique that can be used very effectively to dynamically vary prices.

We take the responsible and fair implementation of AI projects for our clients extremely seriously and will often make modifications to a technical approach or use-case in order to keep them and their customers safe. These tools are immensely powerful so they have to be implemented with appropriate thought towards safeguards and protective mechanisms.

Given the recent—and entirely justified—controversy surrounding live concert ticket pricing in the US and UK, we wanted to highlight three things that are often broken in instances of dynamic pricing. Not only are these practices frustrating for the consumer—they’re plainly unfair on a human level. Read on to find out how we believe dynamic pricing should be implemented responsibly.

First up, what is dynamic pricing?

Dynamic pricing, as the name suggests, is the practice of adjusting prices based on various factors, such as demand or time constraints. We're all quite familiar with dynamic pricing in one guise or another and have been for some time. For instance, train companies offer peak and off-peak fares, which are “time-of-day” dynamically priced. Advance purchase train tickets in the UK are “limited-quantity” dynamically priced—when one band of tickets runs out, a new price is set as a new, more expensive band of tickets becomes available. Hotels frequently adjust their rates based on a combination of demand and time-to-stay; even food items in supermarkets are discounted near the end of the day to clear inventory and mitigate food waste.

Three ways dynamic pricing is broken

When implemented poorly, dynamic pricing can absolutely lead to poor outcomes and unfair consumer practices. This can happen when prices are excessively inflated—regardless of whether the pricing is static or dynamic.

Now, let’s get into the three ways that dynamic pricing is broken…

1. A company or service is operating as a monopoly

When a company holds a market monopoly, such as being the sole seller of event tickets, this can lead to exploitative pricing practices, especially when dynamic pricing is implemented. In such situations, consumers have no alternative market to find competitive prices.

While monopolies can be easily identified in some cases, others are harder to spot. For example, in the UK, care needs to be taken with monopolies of partial markets, such as a single supermarket operating in a remote region or an online delivery service with no local competitors. 

Extreme care should be taken when using dynamic pricing in these situations. It is essential to implement anti-bias measures to ensure that pricing practices do not inadvertently discriminate against certain demographic groups. For example, detecting whether higher prices might affect one socioeconomic group more than others. Even then, there should be a good justification for whether a dynamic pricing approach is appropriate at all in this situation.

2. A product or service being priced is a basic necessity

For luxury goods like art, watches, or high-end hotel rooms, dynamic pricing may be less controversial. However, for essential products like food or holidays, the need for dynamic pricing should be justified. Discounting food near its expiration date is a reasonable use of dynamic pricing that is not driven by demand but by the product's decreasing value.

Holidays are amongst the biggest challenges when it comes to dynamic pricing. While holidays are considered a necessity, they experience increased demand during school holidays, which can lead to unfair pricing practices if providers exploit this demand. Ultimately, this leads to families being penalised or, arguably, being treated unfairly. Again, ensuring that consumers have multiple market options is crucial to preventing monopolies and maintaining competitive pricing.

3. Consumers are not getting sufficient time to consider their purchase decision

One of the most frustrating aspects of the recent live music ticket scandal was the practice of keeping fans waiting in virtual queues for hours, only to be given a very short time window to choose whether to make an expensive purchase. It’s arguably the oldest trick in the book: rush the buyer, make them believe that it’s now or never and prevent them from checking with others whether the price is reasonable. Combined with a lack of transparency, this leads to potential consumer exploitation and explains why the current level of upset with concert ticket sellers seems very justified.

So, how can companies using dynamic pricing protect themselves from claims of anti-consumer behaviour?

The solution: Greater transparency and guardrails

Increased transparency regarding pricing policies is crucial for improving consumer experiences and creating a more efficient marketplace. Unconstrained dynamic pricing can lead to excessive price fluctuations, challenging consumers. To address this, companies can implement "guardrail limits" and transparently disclose what these limits are. From a technical standpoint, it’s perfectly possible to design dynamic pricing algorithms which have a maximum and minimum price—and these thresholds can be published at the outset of a customer journey in the same way as a static price is advertised. 

By enforcing constrained dynamic pricing and exposing maximum and minimum prices, consumers can understand the potential pricing range and make more informed decisions. This could even be extended to offer the consumer insights into the last price at which the product was sold, or the price being offered to the next consumer. Importantly, competitors would be able to view these prices as well and react appropriately as they do now with static pricing; enabling the efficient market which is just not possible with hidden pricing.

We expect this transparency on supermarket shelves and, as such, it seems more than reasonable to expect it of our AI-driven pricing algorithms, too.

-----

Want to explore how we could help you implement AI dynamic pricing responsibly? Get in touch with us today. Or send us a message to chat about the many ways that you can use AI to make your business more efficent.