Defining dynamic pricing is easy. The clue’s in the name — it involves prices that move. Explaining how and why it works is a little more tricky. The long answer requires a grasp of mathematics beyond the reach of mere mortals, but the breakthrough in dynamic pricing in the new digital and consumer age is the result of its ease of use, so let’s keep things as simple as we can.
It doesn’t get any simpler than a fixed price, or a price that remains static until, say, inflation or shifting supply or demand or competition compels a change. An adult ticket to London Zoo has cost £20.45 online since October 2016, or £21.50 at the gate. In February the price goes up to £24.30, or £27.04 at the gate. Quite a jump, but not dynamic — visitors on a sunny afternoon during the October school holiday paid the same as those who turned up on a rainy Tuesday morning the following week.
Variable pricing is a step towards dynamism, but it’s still pretty basic. Take a rough look at shifting demand during a time period — a year, say — and set different prices for different chunks of time, or seasons. You might have just two prices (peak and off-peak) or several. But prices remain fixed within each tier and can’t respond to unforeseen changes or more granular shifts in demand or other factors.
Disney’s recent changes to ticket pricing at its US resorts gained a lot of attention, but they are more variable than dynamic. The parks now have three price categories — peak, regular and value — set ahead of time.
Dynamic pricing is a notch smarter than this, and where the algorithm comes in. In a variable model, humans look at demand and perhaps a couple of other data sources — school holidays, weather from season to season — and crudely define the start and end of each time period. In a dynamic model, multiple data streams can be analyzed, including competitor pricing, with the assistance of computers to adjust prices day by day, or even more often than that.
The algorithms can incorporate mathematical modelling and economic theory that does not exist in the average zoo offices, for example, to present new solutions and movement in prices. The world’s largest retailers including Amazon can change the prices of millions of products minute by minute, something no human could achieve.
But dynamic models still require human participation, and smaller businesses prefer to review proposed price shifts, particularly if they are significant, before they go live. It takes a very smart algorithm to understand a customer as well as a fellow human can, or to predict the potential reaction to a price rise. But when dynamic models are used well, in every case they are shown not only to move prices — but profits, too, and only in one direction.