Slavery Footprint is a site that asks you a bunch of questions about your lifestyle and tells you the number of slaves working for you. Except the way they calculate it the number will always be much too high, and lower numbers don't even mean less slavery.

Going through the survey, and giving numbers well below the American median I got "26 slaves work for you". This would imply there are 8+ billion slaves just to support the lifestyles of Americans, plus billions more supporting wealthy people in other countries. Why are the numbers so high?

Let's look at their methodology:

The Product Score represents the likely number of forced laborers that have been involved in creating the product at some stage in the process of production. Slavery must be known to exist to a significant degree in some stage of production.

In other words, they count someone as "your slave" if you use anything they produce. Imagine a certain kind of shirt is created entirely by slave labor, and it's made on an assembly line where each shirt is worked on by five slaves. If you buy one of these shirts each year, then they would count this as five slaves working for you. If those five slaves together produce 1,000 shirts a year, then we're telling 1,000 people that they each have five slaves working for them. Most people would instead expect that if 1,000 consumers are purchasing the output of five slaves then each consumer has 0.5% of a slave working for them.

If this metric simply overstated the number of slaves supporting you by some large constant factor, it could still be useful if it got ratios right. If I thought about ways of changing my consumption and ran through the calculator until I got my 26 down to 13, would that mean half as many slaves? If I did it solely by cutting my consumption of everything in half then yes. But eating half as much food, for example, isn't really practical. Instead you need to shift your consumption between goods. Say both eggplant and cabbages have a "slavery product score" of 1, indicating that both tend to involve one slave in their production. If I'm eating some of both this will count as "2 slaves working for you," but if I switch from 50% eggplant and 50% cabbages to 100% of either they'll count me as having half as many slaves working for me, even though the total amount of slave labor I'm benefiting from hasn't changed.

Slavery is a problem, but it's not the only problem. Twisting definitions to get impressive numbers hurts the discourse around charities, making it harder for us to find and support the best ones.

(See also: saving lives for 7¢ each.)