Netflix knows what its subscribers want to watch next, airlines know when their passengers arrive at the airport and Facebook knows, well, everything. But are customers — who don't seem at all bothered by Amazon suggesting their next purchase — similarly comfortable with banks using their data the same way?

Banks know a tremendous amount of personal information about their customers — what better insight is there than how people spend their money? — but given the amount of trust that is assumed in a banking relationship, they have to be especially careful about showing their customers they know them without creeping them out.

Essentially, banks are held to a higher standard, so in looking to use data analytics to help their cross-selling efforts, better time pitches and show customers they know them, they have to do it in a way that feels convenient and comfortable, not like they are overstepping.

"We have a premium on trust," said Theresa McLaughlin, global chief marketing officer of TD Bank Group. "That trust gives us an advantage, but it is pretty sacred, so whenever we look at using that data we ask ourselves 'Why did our customers give it to us?'"

Research shows a majority of customers don't mind if their bank uses their data in a way that adds value to their lives. But the challenge is finding the right way to do it.

Customers know banks have personal data and "as long as we are using it for their benefit, they are comfortable with that," said Sandra Nudelman, the head of customer and marketing analytics on the retail side of JPMorgan Chase.

"There are a lot of things that we could do with the data. But we have a strong set of rules and governance around how we use it, and we don't ever cross that line. We've done customer research to identify where that line is, and believe me, there is a lot of space between how we can improve the relevance and timeliness of what we say to our customers and that line. …

"There is so much more we can be doing before we hit the realm of the 'creepy.'"

There is certainly a demand for bespoke experiences.

For instance, a recent Accenture consumer banking study found that 63% of the 4,000 respondents said they were willing to give their banks direct access to various personal information, such as credit card, mortgage and student loan data, in exchange for products and services tailored to them.

That willingness to share data is perhaps even more noteworthy given that the same survey found a quarter of respondents said their financial data had been compromised online in the past two years.

"Banks have more access to data than just about anyone else," said David Edmondson, head of Accenture's North American banking practice. Even though people expect their banks to be more stringent than other companies when it comes to security, "banks that can figure out how to effectively use data to provide better service have a massive competitive advantage."

Banks should use their customer data as a way to build a relationship, not as a way to juice revenue, said Christian Nelissen, head of data and analytics for Royal Bank of Scotland in the United Kingdom.

"Our first layer of protection from crossing the line is intent," and that intent should always be helping customers, he said.

In line with that thinking, his bank uses its data much differently than in the past. "When I first started here six years ago, our use of data was very much focused on selling — meeting targets for whatever campaign was that month," Nelissen said. "We've completely changed the mindset."

Examples include wishing customers a happy birthday or texting them when they walk away from an ATM, for instance.

The bank also has used analytics to spot customers who are drowning in debt. It sent them messages that didn't call out their financial troubles, but merely said, "We think we can help you" and invited them into a branch.

"That cost us money," Nelissen said, noting cases where people were in products with higher interest rates. "But the feedback was awesome. There were people who said, 'We didn't know how we were going to make it to the end of the month.' "

Properly using data analytics can also perhaps make messages feel less intrusive, or at least better timed, said Chris Nichols, chief strategy officer at the $4 billion-asset CenterState Banks in Davenport, Fla.

A few years ago, CenterState's approach to abandoned applications was to call the potential borrower within a day to follow up — it was able to get about 10% of those applicants to borrow. It considered the data and decided to take a more "contextual" approach.

For instance, small-business customers who abandon applications at the point when it asks for three years of financial data likely don't want to hear from the bank the next day — they are busy looking for the statements, Nichols said. "Instead of hitting you right away, our data says to wait three to five days to allow you to gather the statements," Nichols said.

Since it switched to this new approach, the conversion rate rose to 30%, he said.

Nichols said the bank is also using analytics to do more "anticipatory banking" — watching growth rates to determine when a small business might need a line of credit or tracking life events to see when a retail customer might be looking to buy a house.

He said the user experience is key to making sure the message does not feel intrusive. "If presented right, it is not creepy. If it presented wrong, it can be," he said. "No one gets creeped out when Amazon suggests something based on what others have bought."

But banks are decidedly different from Amazon and many of the other companies lauded for their use of things like predictive analytics.

Although banks are increasingly judged across all digital experiences, Chase's Nudelman says they have much more complexity to deal with than other types of businesses when using data.

"Our job is substantively harder than, say, Netflix when it comes to analytics," Nudelman said.

"They have one channel and a massive inventory and use an algorithm to solve the simple problem of what you should watch next. We've got multiple channels, multiple objectives, multiple stakeholders and all of that has to work together and get solved simultaneously."

Also, there are certain hot areas of data analytics that banks are avoiding — at least for now.

Geolocation is an area where banks are interested, but wary.

Some banks are experimenting with beacons or are considering ways to reward customers. It is also viewed as a likely fraud mitigation tool before one used for marketing purposes.

Using Facebook data is a good example of an area where CenterState sees opportunity that it is avoiding for now, Nichols said.

The company pondered incorporating the Facebook application programming interface with its education section. In other words, parts of its site not associated with accounts.

"It absolutely would catapult our ability to use data. It would make us use it more effectively, but we don't think all of our customers would be comfortable with it," Nichols said.

He added that it also may have been like drinking water from a firehose.

"Millennials may have used it," Nichols said, "but the reality is I don't know that we would have had the bandwidth available to use the data if there was great adoption."