It's long been known that who you appear to be on the web can influence what companies try to charge you. Most famously, travel agency Orbitz was found to be offering Mac users more costly hotels than their Windows-using counterparts. But how widespread is such price tinkering?
A group of researchers from the mobile operator Telefonica and a colleague at the Polytechnic University of Catalonia in Barcelona set out to investigate, studying 200 online retailers, including the likes of Amazon, bestbuy.com and Gap.com.
The researchers reasoned that retailers were likely to consider two forms of discrimination: search discrimination, where users were pointed to different ranges of products, depending on how they had been categorised; and price discrimination, where users were offered the same products at different prices.
“Detecting price or search discrimination online is not trivial. First, we need to decide which information vectors are relevant and can cause or trigger discrimination, if it exists,” they wrote in their research paper.
Their study of the 200 different vendors, conducted over a 20-day period in July 2012, showed that Orbitz was unusual – none of the retailers it studied segmented customers on the basis of their operating system or browser.
But users in different locations, or apparently affluent ones were treated very differently.
Checking how location affected retailers' offers was relatively easy: the team used proxy servers at six locations across the globe, using six separate, identical virtual machines that did not accept cookies and blocked tracking. With this configuration, the team were able to ensure that the only way for websites to distinguish between the 'bot buyers' was through IP address.
Separately, they collected price comparisons for a group of bot buyers that could accept tracking cookies. They then created personas by visiting different sites. An affluent persona was created by visiting luxury brand websites; visiting sites associated with budget-conscious buyers was used to create an opposite persona.
The team found numerous examples of both price and search discrimination. For example, a simple Google search for headphones resulted in suggestions for the affluent persona which were four times as costly as those presented to the budget persona. Similarly, a search for hotels on Cheaptickets.com produced more expensive options for the affluent buyer.
Also, the team discovered that mostly, Amazon doesn't distinguish between where a buyer originates, when it comes to its prices – except in the case of e-books. Here, at its most extreme, prices differences could be as big as 166 percent.
They didn't, however, find any evidence of websites discriminating on the basis of operating system or browser.
The team are cautious about the implications of their study. “Our measurements suggest that both price and search discrimination might be taking place in today’s internet,” they concluded.
But the results will no doubt be seized upon by supporters of systems such as Do Not Track, as it appears, the websites you visit might affect the prices you're offered elsewhere.
The team's work is being presented at the HotNets Workshop in Redmond on 29-30 October.
Using photocatalysts to convert carbon dioxide into usable energy such as methane or ethane
Trained on curated data from Moorfields Eye Hospital, the neural network also shows clinicians how it reached its judgement
Yokohama National University demonstrate technology that could lead to a fault-tolerant universal quantum computer
Top-of-the-range Threadripper 2990WX now available from Scan, Ebuyer, Overclockers, Novatech and Amazon