"Insulators" always conduct. But mostly the world divides into high-conductance (metals, salt-water) and low-conductance (dry wood, glass). The difference is many-millions to one. "Semi-conductors" stood out because they fall in between (though a thin well-doped semiconductor will carry a lot of current).
One problem with insulators is that our world is full of conductive crap. Sweaty finger grime is a popular contaminant on insulators. A think layer of salt-grime is not a great conductor, but it is at least as good as your body, which is all you care about.
Another problem is that electricity will jump through air if you push hard enough. You need big insulators to stop big voltages.
In dry air, your body can accumulate maybe 20,000 volts charge. (This will jump nearly 1/4", much further than the thickness of a typical key-rubber.)
The capacitance of your body to the universe is around 100pFd.
Skin resistance at low voltages is normally 10K-100K, but in ESD testing it is conventional to use 1K5. This 100pFd 1K5 model has been around for over a century (a spark in a coal-mine full of gas is explosive). There are other models, but this one seems to be appropriate for KV charges and general estimation of spark behavior.
Assume the doorknob has a low resistance (much lower than 1K5) path to ground or a large object (much more than 100pFd to the universe).
20,000 volts across 1K5 is 13 Amps.
20,000 volts and 13 amps is around a quarter million Watts. In a small area of the finger!
The time-constant of 100pFd and 1K5 is 0.15 microseconds. So that quarter-MegaWatt is down to an eigth-MegaWatt in less than 1/6,000,000 seconds. If you stretched the shock over a whole second, it would average something like 0.050 Watts for one second, way too small to notice.
My gal once worked on nylon carpet around metal cabinets. I made her a shock-ring: a 1-Meg 1/2-Watt resistor on a crackerjack ring. She touched the far end of the resistor to the cabinet first, and the charge was dissipated at the rate of about 20,000V/1Meg= 20 milliAmps, not Amps. The time-constant was under a milliSecond so even the quickest flick was enough to drain most of the charge. If your problem is a steel door (high capacitance and/or a direct path to general ground) you could screw a 1Meg resistor into the face of the door and brush that before you aim the key at the lock. Or you could build a 1Meg resistor into a key-like dongle: hold that and brush the door with the end before you go for the key.
It may also help to anti-stat your car seats, nylon parka, any carpet along your path. In some of our offices I have to anti-stat the carpets each winter so ESD does not lock-up keyboards when people shuffle and sit. There is magic spray for this. There is also spray for silk stockings and slips to keep your panties from getting in a bunch, and this works on carpets. But the cheapest trick is liquid fabric softener ("Downey" or ShopRite equivalent) mixed about 1:20 with water and misted with an old Windex bottle. What these really do is hold some atmospheric water in the fabric or carpet so charges drain much faster than they do in bone-dry air. Don't use too much or it gets slippery and slimey. And when air gets REALLY dry they can't do much. But in your Mid-Atlantic area there is always a little water in the air, you just have to make some of it stick to the fabric.
As an extreme trick, find a car grounding strap. This is conductive rubber clamped to the car so it drags on the ground. It was very common in the 1950s when the early nylon upholstery would zap you every time you slid out of the seat. It bypasses the high resistance of the tires. But I have not seen one in years. In fact the last time I saw one, I was putting it on a PC cart to keep the cart drained to the carpet, and that was 1986.
