I can be sure of one thing, if I start out with a beat and percussion, the idea ends up a lot more driving than anything else as it seems to be more percussive orientainted, whereas if I start with a few chords or a busy melodic bass line, the idea develops into something a lot more musical so I tend to know this and just go with the flow.
I always used to find this. Although now the percussive elements emerge from the musical elements.
As part of my learning process, I always start with what I last had the most difficulty with. This way I learn without distractions.
I rarely have a need to put in a 4x4 kick clap pattern. Its in my blood now
I think it's more to do with ones musical background. I started in music playing keyboard way before I knew anything about composing/producing. It's an almost unconscious act now for me to 'get busy' when I have a sequencer and a keyboard at hand. The 'danger' of that, is that I quickly end up playing a natural way ( which is all I know ) and everything ends up sounding like a 1980's electro funk record. The way I write bass lines, the way I write chords and harmony/melody are all naturally geared towards a structure that is most definitely not within the realms of what I would class as Techno or House, I mean my left hand will disappear off the keyboard to access the pitch bend wheel and mod wheel and I only realise I have done it afterward.
It sounds stupid but it's true and also, for a keyboard player, when you have a blank canvas like an empty sequencer and just a beat or a metronome, it's extremely difficult NOT to suddenly fill up the entire empty space with things like bass fills, pitch bends, sudden flirtatious keyboard trills and other ridiculous phrases that are completely irrelevant, thing is, if I choose to prgram stuff in, like in a step sequencer, I will still get those important accents and 'ear candy' bits but I will do it across a number of different sounds, does that make sense??? If not, here's an example: Say I write a 4 bar bass line using my keyboard as the input method. I will use the same sound, the same synth and it will be a fairly busy affair with perhaps a change at bar 2 and a bass fill at bar 4. It's a basic example but it illustrates perfectly my dilema. Now, if I wrote the same melodic sequence using say samples in a step sequencer, I would almost certainly spread the part across a number of sounds. For instance, I might use a single deep bass sample to reinforce any events on any offbeats. I might call up a different sound to work with that sequence for more syncopated measures and any 'bass fills' or points of interest could well be made up with a number of say Toms or other pitched sounds. So you see, on say staff paper, the sequence would be notated the same but in pure audio, it would end up far more rich in terms of timbre and subtelty than it all coming from a single bass patch off a synth. I'm trying to make it as clear as possible how different I sound from playing stuff in to programming stuff in. By programming, I think holistically and say to myself, hmmmm, another sound could play that note and so on etc etc, whereas with my synth bass and my fingers on keyboard, the rhythmic/melodic idea comes from one source sound and in my opinion, ends up sounding like a bass line ( or whatever ) from a Herbie Hancock wannabe or something.
If I start out and unplug the controller keyboard and work in step sequencers and pattern sequencers, I can find the grooves and subtleties I cannot find when relying soley on a keyboard to input my ideas/ I'm not sure if that makes sense but it's definitely a thing with me. It's almost like, I can't PLAY Techno/Techy House, but I can definitely program it.