The Wall Street Journal has a long profile on how Carnival Cruise lines use big data to maximize the revenue and profits extracted from each customer. It's a good example of how companies use data to find the maximum prices customers are likely to pay and push them towards those goods and services at that mximum price.
Carnival owns 100 ships and nine brands and can dynamically price cruises on each ship and figure out who to market those seats to in order to apply its big data analytics:
Every day at Carnival, a data science teams sets in motion several algorithms that crunch information about items such as passenger behavior, vacation trends and queries from travel agents and potential customers online and by phone, Mr. Leibowitz said. The analytics systems run for eight hours overnight, devising thousands of recommendations for ticket price tweaks to Carnival’s slate of scheduled cruises worldwide.
The company also use "psychographics" to figure out who they want on the ships in the first place, since there's "a difference between someone who will pay $12 for a pair of Carnival logo flip-flops and one who will shell out a few hundred dollars for a champagne catamaran side trip on the high-end Holland America line."
One way to understand why companies increasingly manage multiple brands is to better implement this kind of price optimization/discrimination since slotting people into particular brands de facto slots them into different prices often for goods or services that are similar except for the brand name. The WSJ cites to the example of Hilton recently purchasing two new hotel chains to "add rooms priced in the low-end and mid-tier rungs to its lineup."