In order to properly assess the performance capability out of each runway, a proper analysis of the obstacle field needs to be done.
This involves looking at type A charts, topos of an appropriate scale and any other information that you can get your hand on (google earth is an excellent tool for general orientation and "reasonableness")
I would suggest that most performance engineers would automatically look at straight ahead first in order to see if the terrain allows an unrestricted (or close enough to unrestricted) payload.
If the terrain is such that a turn is required then they draw up the splays account for all the obstacles and use those obstacles to produce the runway charts.
In many cases they only publish the turn procedure - simply assuming that crews understand, or the ops manual clearly states, that in the absence of a published procedure the procedure is straight ahead.
Personally I prefer to publish a procedure for every runway - even if the procedure is straight ahead clean up at 1000 ft.
This eliminates any doubt and also gets people into the habit of looking for an engine failure procedure for every runway.
One observation of performance engineers who are not pilots is that in order to milk the last kg of payload out of a procedure they often make the procedure so complicated that the average pilot can't actually fly it - or can't fly it accurately, which kind of defeats the purpose.
I am thinking of the procedures like "runway head to 5 dme then left track 125 to 9 dme then turn left track to the VOR and hold".
I use (or used to when I did this stuff) the "dark and stormy night" test - could I do this myself in the aeroplane on a dark and stormy night having just flown 4 busy days, am totally shagged, and have an ineffective offsider.
The other thing I dont like is procedures that use heights and/or headings - "at 1500 ft turn right heading 300" - how can you be sure that you are tracking where the performance engineer intended you to track?
So Alex - after that rant - it might be worth checking with your performance engineering people - who might be an internal department or Jepps, Navtech or somesuch other external organisation, that if no procedure is published then straight ahead is the appropriate procedure.