Dynamic pricing: How consumer internet companies change prices in real time with sophisticated algorithms

Complex though that it might sound, the concept of dynamic pricing has been around for years in the offline market

Ever alighted from a swanky new car to buy fruits from the vendor at a market and discovered that you are paying a premium compared to someone who arrived on foot? If so, you were given a glimpse of what is known as 'dynamic pricing,' a craft when the street-vendor practices it and nearly an art form when it is determined by complex computer algorithms.

Let us suppose friends are reacting to a group message that says "Let's plan a trip to Ladakh in July." They all log in to their digital devices to check airfares, only to realise that rates fluctuated by a few thousands rupees each time a new search was made for flights. What was happening to the excited friends was that an algorithm detected a spike in incoming requests for a particular flight, causing air fares to rise. In industry parlance, this is 'price optimisation' or the determination of fare based on real-time demand.

Almost every product or service you buy online is subject to dynamic pricing, and understanding how it works is important to make smart spending decisions. For companies, too, it is smart to price dynamically because it helps them adjust to the vagaries of demand and supply.

"There is a misperception that dynamic pricing is an exploitative pricing mechanism," says Kartik Hosanagar, a professor of technology and digital business at The Wharton School, University of Pennsylvania, explaining that the fact that supply and demand are not always perfectly balanced justifies resort to dynamic pricing.

Demand and supply, actually, are just two of the variables that go into determining price, as we saw in the case of the eager friends who were headed to Ladakh. In the case of travel industry, apart from demand and supply, events such as festivals or holidays, time of day and occupancy are also factors that can cause a change in fares, says Sanjay Mohan, the chief technology officer of travel site Makemytrip, which is listed on the Nasdaq.

While occupancy is a major factor determining fare in travel, the number of buses or flights operating on the route are also considered. For instance, if there is only one flight scheduled at 8 am across airlines, it is priced higher. But if another airline schedules a flight at, say, 8.30 am, capacity is doubled and competitive pricing kicks in, says Ankur Bhatia, a director at travel software provider Amadeus India.

Also in the travel industry, it's imperative to have a vigilant dynamic pricing model in place since it deals with "perishable inventory," meaning if a plane flies an empty seat there is no way to recover the cost of the seat.

Complex though that it might sound, the concept of dynamic pricing has been around for years in the offline market. The auto driver you approach to take you to your home at 6 pm asks for an "extra Rs 20," blaming it on traffic. The fruit vendor round the corner who sells strawberries throughout the year raises the price of the fruit by a few hundred rupees in off-season. You also pay a premium for passport- size pictures that you get printed urgently.

With services migrating to online platforms, pricing them dynamically is now done using sophisticated algorithms that list out an optimum price by considering more variables that an individual can possibly include. For instance, online classifieds portal Quikr uses one such algorithm to determine, say, the ideal price at which to list that old car for sale. It helps sellers decide what it calls the 'maximum selling price' using an algorithm that takes into account active listings as well as historical data and comparing these with attributes of the product, such as the model's current price and year of manufacture.

"Delivery time, weather, market conditions are some factors that trigger an alert to change prices in ecommerce," says Madhusudhan Rao, the head of India operations at Boomerang Commerce, a startup that helps retailers integrate dynamic pricing solutions.

For cab-hailing apps such as Ola and Uber that work on an on-demand model, ensuring that there is always a cab available on request requires dynamic pricing to fulfil demand by providing supply. To do so, they resort to 'surge pricing,' which has two effects: people who can wait for a ride often decide to wait until the price falls; and drivers who are nearby go to that neighbourhood to make an extra buck. As supply increases, or demand falls, prices head back to the normal equilibrium.

"For the duration that surge price is active, the algorithm reads demand data every five minutes and updates the multiplier effect that determines fare," an Uber representative said.

A product that is a runaway hit can also command prices, as in the case of the hard-boiled candy, Pulse, that Noidabased DS Group launched last year. It became so popular that local kiosks were asking consumers to pay a 50% premium to buy it.

While this might be a classic example of skewed demand and supply spiking product price, there are also sly tricks by some online commerce platforms which are not shy of discriminatory pricing. In what is termed by industry experts as a shrewd way of targeting the well-off, some of them quote prices depending on the devices from which the requests emanate.

So, if you are checking from the latest iPhone, you could end up paying more than a buyer who uses an old desktop.

It's not all bad for the customer. Have you ever noticed price variations in the same products on an online shopping portal?

An algorithm makes it possible. It cross-references features of products, matches common attributes, compares price and triggers an alert if a similar product is charged at a different price on a competitor's platform. A rule set in the algorithm decides whether the price of the product needs to be changed to match the competitor's price. This means that if a seller lowers the price of a popular product online, its competitor is highly likely to do so, too.

So while the friends planning their trip to Ladakh learned their lesson and booked their tickets through one device in a single attempt, they also put their learning to use by buying their winter wear at lower price in summer.