Looks like you’re right. This https://insideenergy.org/2015/11/06/lost-in-transmission-how... claims it’s only 7%. I’ve updated the original comment. You still have the problem of non green generation sources, and you could still do better with a high efficiency gas furnace on-site, although I think most people go with low efficiency (~80 or 85%) units.

It’s complicated.

If you’re talking about transmission between a nearby generation facility and end user, losses are low. If you’re talking about losses in April/May, it’s low. The statistics are difficult because the average losses don’t matter, it’s the marginal losses that matter. In mid-August when electricity demand peaks, you need to generate twice as much electrify at a higher price to get another marginal unit of electricity.

For New York, which is planning to ship power in from Quebec, losses at that distance are very high, and require significant investments in new transmission infrastructure to improve yield. At one point, shipping additional grid supply from far away to NYC resulted in losses as high as 65% in certain weather conditions. That’s why so much was invested in peak load reduction… stuff like voluntary shutdowns in exchange for discounts, efficiency programs, etc. Every dollar invested there avoided like $7 in infrastructure investment.