Around the time the Cambridge Analytica scandal was dominating headlines, I was at a wedding, seated next to a digital marketer. Unsurprisingly, the illicit gathering of millions of people’s detailed Facebook data in order to create psychographic profiles came up in our conversation, and we both agreed that compiling personal information without consent was clearly wrong and absolutely inexcusable.

But, we sheepishly admitted, wouldn’t it be cool to see that data!?

This, of course, is the dilemma facing marketers and researchers today, even those without a shred of malicious intent. Massive amounts of personal data can be incredibly valuable, and not just in a financial sense. But ensuring that this data is collected with informed consent can be difficult. Many researchers have a defined set of participants who can be asked for their consent in person, with a human being available for questions. But in an “open” study, with a free app downloaded from the app store, the consent process is generally relegated to a pop-up dialog with a lot of text which–as a rule–is never read.

Furthermore, the definition of “informed consent” can be murky. In the case of the Cambridge Analytica scandal, Facebook’s initial non-apology apology was a PR failure, but to some extent accurate: the original users of the app were informed their data would be used for a study and they did consent to give it away. What wasn’t clear was how that data would be used (sold for use for political rather than academic purposes), nor that their entire social network would be scoured as well.

Our partners and we have struggled with this issue (or ones like it) in almost every project we’ve done. How much explanation is necessary to consider “consent” to be “informed”? If the user doesn’t or won’t read your carefully worded legalese, is that on you or them? How do you balance the user experience between one which informs completely and one which causes the user to give up and uninstall?

These are some of the subtle but important questions you’ll need to ask yourself as you design apps for research purposes.