People care less about large groups of people than individuals, per capita and often in total. People also care more when they are one of very few people who could act, not part of a large group. In many large scale problems, both of these effects combine. For instance climate change is being caused by a vast number of people and will affect a vast number of people. Many poor people could do with help from any of many rich people. Each rich person sees themselves as one of a huge number who could help that mass ‘the poor’.
One strategy a charity could use when both of these problems are present at once is to pair its potential donors and donees one-to-one. They could for instance promise the family of 109 Seventeenth St. that a particular destitute girl is their own personal poor person, and they will not be bothered again (by that organisation) about any other poor people, and that this person will not receive help from anyone else (via that organisation). This would remove both of the aforementioned problems.
If they did this, I think potential donors would feel more concerned about their poor person than they previously felt about the whole bunch of them. I also think they would feel emotionally blackmailed and angry. I expect the latter effects would dominate their reactions. If you agree with my expectations, an interesting question is why it would be considered unfriendly behaviour on the part of the charity. If you don’t, an interesting question is why charities don’t do something like this.
Meteuphoric 
Some charities do do something like this: ‘big brother’ programs, say, or programs where you ‘sponsor’ some poor child somewhere.
Mostly though they don’t, and I think you are right that if charities did this not-so-voluntarily, it would be considered unfriendly by donors. I think the reason is that donors would experience this as forcing-them-to-care in a situation where they never, in fact, cared in the first place. Much charitable giving is less an expression of actual ‘caring’ than it is of a desire to signal-how-much-I-care to others in and below one’s social class (say, at some expensive charity dinner-event). A shift to one-on-one charity would shine a spotlight on this distinction (either compelling tangible action or exposing one’s hypocrisy/false feelings) and thus make most charity enthusiasts feel torn, imposed upon and uncomfortable, quite understandably so.
See here: http://en.wikipedia.org/wiki/Kindernothilfe
Also: http://en.wikipedia.org/wiki/Child_sponsorship
this has been done for years
http://www.google.co.uk/search?q=sponsor+an+african+child
I think charity is done for self-beneficial reasons, such as signaling that they are a nice and helpful person, or that they have the resources to give, or done to increase the help they can expect from those given to.
In cases where there are many who could help, you face less of a cost of looking unhelpful by not stepping forward yourself, and you also have less opportunity to demonstrate you unique ability to help.
I admit to my model being confused about being less interested in helping groups than individuals, much beyond the way in which that reduces the extent to which they have made a strong impact on some set of people (or in which it is easier to defend your claim that you made a difference). So if you sponsor an orphan or something, that orphan is going to know that *you* are the one who helped them, and having made a personal difference you might expect to receive more help in return. If your contribution provides about a third of one meal to a large group, once, and they don’t know it was you and have no way of distinguishing you, you gain less in future alliances. Of course an orphan in Africa is very unlikely to actually be helpful, but we seldom update our EEA heuristics.
To force a particular person on someone, when they weren’t looking for someone to help, and say they will receive help from this person or not at all, imposes a cost on someone by either forcing them to give or hoisting on them the responsibility of that person’s loss and responsibility. Most people aren’t interested in giving (and weren’t expecting to gain much from it), but now they look really bad if they let another person suffer when they could clearly help.
To those of you who mention sponsoring a child, note that you don’t get paired with the child until you make the decision to sponsor them. If you don’t sponsor that child someone else can, and there are a huge range of children to sponsor. I agree the strategy there is in the same direction of drawing attention to individuals rather than the masses, but it is quite different.
Here’s another charity that does this: http://www.donorschoose.org/
It lets donors pick which classroom/project they want to donate to directly.
charity is all about overhead costs…..when they spend more than 60% of the funding on administration and/or running the charity they have to hide how they spend the money.
I would love to see this charity actually started. In fact, it’s one of very few charities other than SIAI that I can easily imagine donating to and I hope others feel similarly. It would be much more effective, however, if it could credibly assert that no other charity would come along and make a similar offer/threat in the future.