Usually when you're looking at this, what's going to matter a lot is final revenue and (as you noted) margin.Well, you are only looking at the US. Worldwide Detroit sold 1 million copies in 2 weeks where Heavy Rain needed 5 weeks. Digital sales mean those 1 million units also generated higher revenue, there was no digital copy for Heavy Rain. 2 weeks vs. 5 weeks seems like significant improvement no? That's reaching the milestone more than twice as fast + higher revenue, while its budget isn't twice as high as Heavy Rain's.
Heavy Rain came out on a €16.7 million budget, cost about €40 million total with marketing included, and ended up with "earned Sony" €100 million (I can't tell if that's profit or revenue).
We know that Detroit cost at least €30 from Le Monde in development costs. We're not sure what the marketing budget is. The NPD results we have are 20% up from Heavy Rain in revenue, and we know it seems like it might be up a bit more than that overall (though number of weeks doesn't really tell us as much as might be assumed because we don't know what the individual weeks looked like leg wise). Obviously digital margins are higher than retail margins, so Sony would be getting something like 95% digitally (they only have to handle payment processing fees) versus 80% at retail (no license fee cost).
It's certainly possible that this shakes out to a healthy profit in the end, but I don't think that's 100% clear.
However, what's generally going to determine if they greenlight another game is how they expect that to perform relative to costs. If they expect the budget to linearly increase to €45 million, that's a pretty healthy chunk of change. If they expect it to keep geometrically increasing, they would get a €60 million budget, which is a very healthy chunk of change. Do they expect sales to keep up with that?
Maybe they don't expect the budget to increase at all though, even with PS5 quality graphics. Maybe they're interested in greenlighting two or three smaller titles instead, like how most games in the genre are budgeted and priced. There's kind of a lot of factors there.
Even if the final numbers come in and the conclusion is "Well, that was a huge slam dunk, and we're going forward like normal.", I still feel the results we're seeing here fit my original thesis that cost is seemingly increasing faster than sales, which is why most publishers seem to want additional revenue streams past the initial sale.