Cylinder heads flow a max amount at lift of, say .550. Any reason to have a cam with a higher lift?

Yes. You use the most lift the valve gear will tolerate. I can put any head on a flow bench and with enough pressure drop can cause the lift curve to drop.

I try and get all the flow loss out, but I no longer go berserk trying to find all of it. Most of it comes from using a valve that is too big for the port to feed.

Impossible to say without actually trying it. If the engine needs more airflow then the additional lift will probably help. If the engine doesn't need more airflow then it is just a waste of time and money.

You can look at this several different ways, if you install a cam that has .700 max lift for 15 degrees of crankshaft travel and has 45 degrees of crankshaft travel at or above .550 lift will the motor make peak power longer or not compared to a camshaft that has 15 degrees valve opening at .550 lift
You can have cams made that will hold the valves opened for a bunch of duration or not, your choice
A round nose long duration cam lobe compared to a standard looking lobe will have completely different idle and driving manners

I'm reflecting the cam choices we made back in the 60-70"s. FT Isky cam with .550 and 292 duration. Bumped up to what you get for a bracket cam with a lift in the range of .520-.590 and a duration of 300-310. Looking at specs on street builds with rollers in the .650+ range. Overkill or I need to know something else?

What do you have for valve gear? I net .600 lift and that's about all you can do unless you run .100 long valves and then it's a [censored] to get the geometry correct. I'm assuming stock heads.

I'm also only 255@.050 with 280 advertised. There is no reason to limit lift more than you have to. And there is no need to keep adding duration. Your valve gear needs to right to do it.

I looked at some old flow #s for a pair of W2 heads that I have. At .550 they are flowing around 300 and pretty much steady to .650. Since the flow remains steady then the higher lift is keeping that flow going and making more HP, right? There is no drop off to hurt power, right?

The correct size valve, a proper valve job and some short turn and guide work will take all of that break over out until the test pressure gets to about 40ish inches of water. At that pressure, no matter what you do you'll get some break over.

Use a 2.100 valve (yes they fit in there) and you'll make the port not be able to feed that valve and the break over will move to a lower lift and make it not come back. Until you make the port square. That's a whole nother issue.

Like I said...I try and get as much lift as the valve gear will take in the reliability you expect. If you've done much testing with a flow bench and looked at areas of pressure differential, I'm not so sure that testing higher lifts at higher pressures is actually the best way to do it.

I have access to a Flow Data bench that will pull 60 inches on most heads and you can learn a bunch. But I'm not sure at .500 or more lift the port is actually seeing that much pressure drop.

I do some testing at 10 inches at say .800 lift and then test at lower lifts at 40ish and then change the test pressures and see what happens. It's not cut and dry always.

More lift is better as long as the valve train is correct.