The best way to answer your question would be to create a hypothesis.......then test it.
You might, for example, allocate 1 hour for every $1000 of project price. That $25,000 job would, under this hypothesis, take you about three days to handle the management functions.
Break those hours into phases. You should be able to ascertain how many hours were spent prior to inking the deal and how many were spent after the contract is signed.
These two primary accounts subdivide into specific actionable items. Measuring the job site, for example, is not the same as engineering the cabinets and cut-listing is not the same as engineering either.
Building a production strategy is different than cut-listing or engineering. Passing out tasks and monitoring status is different than strategizing.
To the extent that you parse these processes out formally you will be able to implement systems that can improve the outcome while lowering the cost.
It's all about mitigation. There are a million things you can't control but there are a few that you can and these should be controlled. The customer is at one level the random part of the equation but looked at from another perspective is the easiest part. All customers want exactly the same things and they are highly motivated to make it happen. These people will sit up in their pajamas till midnight poring over Pinterest. Your employees, for the most part, can not be inspired.
I would first solve your problem through the lens of your customer. You will get more help from them than anybody else.