The only reason they're doing full reusability is to enable Mars missions. Had they only wanted cheap flights for satellite constellations they could have launched with what they have now (i.e. reusable booster and single use starship), and it would be cheaper than anything on the market with huge upmass capabilities.

Something that I rarely see mentioned about SpaceX's goals and financial viability of a reusable Starship is how significant of a boost they get to their talent pool by working on the absolute bleeding edge on things that seem impossible. People go work at SpaceX because they work on crazy things like Starship instead of only working to make "cheap flights for satellite missions". SpaceX attracts the absolute cream of the crop (at below market rates even but that's besides the point) and those people work better/harder than they ever would working on more boring goals. It's the power of vision and motivation.

The constellations they can launch now are only a small fraction of the size they aspire to launch. Their plans for Starlink alone call for almost 10x as many satellites as they currently have. Then there are all the other constellations they might launch in addition to Starlink, including defense contracts that might call for rapid replenishment of expended/destroyed satellites.

As for Mars, Starship might be ready for that in a few years (a year ago I was saying in a year or two, but I've kicked that back). But where are the Mars customers? Who is developing Mars habitation plans and hardware that will be ready in even twenty years? Commercial demand for large satellite constellations obviously exists. Demand for Mars colonization is nonexistent, it exists only in the hopes and dreams of sci-fi junkies.

I really hope I'm wrong, because I'm a sci-fi junkie!

That doesn't mean they're going to achieve full reusability, although obviously it looks more and more likely.