It's a simple & practical, albeit ugly, way to build a dummy load.
Keeping the resistor and the connector close together is good; if they are separated they should be joined by a 50-ohm line, either co-ax or the microstrip/stripline for which this kind of package is intended.
As 'BYU said, there is adequate electrical connection to the resistor flange which, in the "TC" model, is the ground end of the internal resistor.
The return loss/SWR is very good; 30dB RL = ca. 1.07:1 SWR.
The device is specified at max. SWR of 1.30:1 over its 3 GHz range.
If the amount splattered-around is any indication, there may be far too much heat transfer compound; the thickness of compound between the mating faces should be virtually zero, say 0.001".
Too much compound gives less heat transfer than no compound at all.
But, if you are happy with the appearance and with the 90-second time limit, you have a useful item of test equipment.
Oil will increase the 90-second period but it will also increase the cooling-down time (unless measures are taken to cool the oil).
SWR may also increase but probably not significantly for most work.