Something that really bothered me back in GalCiv 2 was that some planet improvements were not well designed. The worst was the Industrial Sector (in the TA expansion). It increased its maintenance cost (compared to the previous factory) more than it increased its industrial output. It more than doubled it's build cost over the previous factory type. The only reason why I would want to research the tech was to nab the last mining module that was tied to it.
So if I don't like how certain planet improvements turned out, enough that I want them changed, then why not change them? To be more precise, why not design them to begin with? For factories, that might mean changing 3 variable, maintenance costs, industrial output, and build cost. Other planet improvements might change different variables. There might even be cross over between different planet improvement and variables. Imagine a factory that also boosts the economy (produces goods among other things), or boosts population growth (produces drone servants among other things).
Likewise, technology might be able to influence how much you can push things in various directions. Maybe you start out with a base design that you can modify only a little at first. Researching techs might allow you to modify things even more. Maybe researching techs will unlock new base designs that be modified more that other designs and/or modified in ways that other designs can't.
Factories I would like to be able to design would be fast to build factories to develop my worlds quicker, cheap to maintain factories for when I'm done building things, and high output factories for those rare worlds that I want the absolute maximum output possible.