I hate to be a Debby downer, but I'm constantly unsure if I'm happy with Google being able to read my emails, let alone a random company out of nowhere.
It's unfortunate, because there is a huge opportunity for cool things to come out of handing your email over to some company (including this app I'm sure) but the risk right now is just too great.
Hmm. Unroll.me is owned by Slice, I wonder if Slice has a different policy regarding selling aggregate user data (since it also scans your inbox, but to find shipping and delivery emails).
The subsections "Vendors and Suppliers" and "Anonymous Use and Sharing" on section V "How we share your information" are quite vague ...
If you're going to charge me a subscription instead of mining and selling (even abstract insights), I'm game. But first you have to promise you'll never mine and sell my inbox.
When we ask for you to say "we will never mine and sell your inbox" and you instead say "mining and selling your inbox data isn't our business strategy," I hope you understand why that would make people nervous about using your product.
I think what is making people nervous would be something like this phrase from your privacy policy:
>We may collect other information that cannot be readily used to identify you, such as (for example) your Email Meta Data and IP address of your computer.
You make clear earlier that you will not sell or rent personally identifiable information:
>If you do provide personally identifiable information to us, either directly or through a reseller or other business partner, we will: not sell or rent it to a third party without your permission
But it is unclear what "email meta data" means and whether or not you will rent or sell anonymized data "that cannot be readily used to identify you". Other questions might include how long you store this information.
With all due respect "mining and selling your Inbox data isn't our business strategy" is not the same as "we will never ever ever ever sell or mine your Inbox data". My inbox is too precious for any potential loopholes.
Yeah, I understand collecting data is a part of training your model but it could be worth explicitly saying that zero data (either personally identifiable or anonymized meta data or any data at all) will be sold or rented. And if a change to that policy happens, include in the language of the user agreement that it isn't retroactive and only applies to new accounts.
The vague language is what bothers people. Probably having very strong language in the opposite direction (that you will NEVER sell any data, and would shut down rather than sell it) would probably win you favor.
Edit: nm. This only happened this week, and I missed the news. Just Googled it and found out.
https://www.nytimes.com/2017/04/24/technology/personal-data-...