No development project can possibly benefit everyone equally. E.g. a project aimed at creating jobs and other economic opportunities will disproportionately benefit the able bodied. Most people would agree that is not a reason not to undertake such a project.
Some donors attempt to get around this by mainstreaming support for so-called disadvantaged groups, and requiring projects they support to incorporate some kind of support for disadvantaged people into their work. I think this is a fairly silly approach. For a start you are never going to capture every single disadvantaged person: women and HIV/AIDS sufferers are the most commonly supported, mentally ill people rather less often. Secondly bolting on such adjuncts comes at the expense of project focus, and leads to project managers overseeing work in areas in which they are far from experts. Much better, I reckon, for donors instead to support a wide portfolio of projects that ensure disadvantaged groups are given a chance. (Different donors could even specialise in supporting different disadvantaged groups.)
But if we are to accept a certain inevitable degree of inequity in project design, how much inequity is acceptable? And by this I do not just mean how much is acceptable not just to us, but to the would be community of beneficiaries? Here is my suggestion for a convenient yardstick.
Probably the single biggest dimension of inequality in the world is the nationality of your parents and/or where you were born. And yet, a few philosophers apart, this is not an inequality that gets many people especially riled. Envious: yes, angry: not so much. My guess is this is because there is no human agency involved. The outcomes are very unequal, but, by and large, poorer people have not suffered a specific recent injustice perpetrated by identifiable rich people to cause this inequality. Even with colonialism, when taken as a whole, it is hard to argue that people born today in ex-colonies are poorer as a result, and plenty of people argue colonialism brought various developmental goods in exchange for lower freedoms.
Instead most people can accept without a deep upwelling of anger that even if you are born in the bottom 5% of the population in the UK, you will probably be able to watch TV every day of your life that you want to do so. Indeed, even if you do not have a job, the UK government will pay you enough money that you can afford to watch TV when you want to. Whereas if you were born in rural Malawi, say, you might only occasionally ever get to watch TV.
Thus my yardstick is this: does your project appear to introduce more inequity than this basic starting inequality? If so you should be worried? If not, you may well be ok. In many ways it is not a very good yardstick: rural Malawians and poor Brits are not living side by side, and so the inequity is remote, hidden even. Moreover the key point is that poor people do not really have anyone to blame for this starting inequality, whereas project staff and managers are clearly identifiable for local ire if unfairness is perceived. But it does provide a philosophical anchor point that could be useful during project conception if ever you are worried about differences between winners and losers that could arise from a new project.
Maybe someone else has a much better yardstick?