On 08/03/2016 01:27 AM, mray wrote:
> By definition the carry over is lower than the limit where fees make
> sense - I expect this to be low.
> For this low amount of money to trigger an unfortunate un-matching the
> total would have to be full to the brim already. This will hardly
> happen, and IF it happens it is only an indicator of a bad situation
> that will soon get an auto-un-match anyway. There is not much to gain.

I was going to make this point myself. I agree, it just happens when
we're already approaching the limit, which is already an issue, this
just triggers it sooner.

> This corner case of a corner case is *NOT* worth breaking the *ONE*
> limit the user trusts us with!

I must say, I get *less* sympathetic with your views when they are
expressed hyperbolically. Charging 2 months at once instead of in two
separate charges does not break the limit or the trust. I actually
*agree* with you that the experience is cleaner if we make the limit
even harder and have less to explain. When you equate "we have to
explain combining two months into one charge" with "we broke the trust",
it actually makes me less respectful of your view because I disagree
with what you are saying.

So, to be clear: when you give exaggerated or even bad justifications
for something that might be the correct approach, it makes me more
skeptical. So, please, try not to use this sort of hyperbole.

> I would want to be able to make a promise to honor the limit *without
> any restraints*. We can do that. Not doing it makes us look desperate or
> needy. People setting a $10 limit should never find a $11 transaction
> fee in their payment processors accounting. No matter what wordplays we
> come up on our site to differentiate between "monthly pledge" or
> "monthly total".

So, in the end, you are right that keeping the limit totally clear means
less confusion, less to explain (maybe), and I'm okay with going this
way. I think you're simply entirely wrong that it makes us look
desperate or needy.

> If we let the user set a limit we need a darn good reason to ignore it
> *ever*. This is not a good reason.

Well, again, I think it may just be simplest and best user experience to
not combine two month's charges into one if that results in a higher
than one-month-limit charge. But I don't doing so is *not* correct to
describe as ignoring their monthly budget limit, it's just that it's
more confusing to have to explain the slightly weaker form of limit, and
I agree that making it the very hardest and easiest-to-explain approach
is indeed best.

So, here's my proposal:

If there is a carry-over: *never* make the carry-over result in *either*
over-1-month-limit charge *and* never make the carry-over result in
suspension of pledges. Instead, just charge a *portion* of the
carry-over that the monthly limit will allow.

Example: $2 carry-over and $1 fee and $7.50 of
current-month-pledge-values with a $10 limit: Charge precisely $10
total. Now the remaining carry-over to be included with the following
month is $0.50.

Basically: we will charge the *entire* limit (and no more), in order to
slowly widdle-down the carry-over. That's the right way to do the corner
case. Thus, carry-over will never be a factor for suspensions and we
never violate hardest limit version: absolute monthly-charge limit.

I see zero downsides to this approach.

Attachment: signature.asc
Description: OpenPGP digital signature

Discuss mailing list

Reply via email to