That's not always the problem, unforseen problems, "I forgots", untested designs, etc., can all contribute to unexpected scope in design. And, of course, there's the "winning" bid that might have little correspondence to reality; for aerospace, the estimators are the engineers, but the GM and marketing have the final say-so on bids, so estimators and engineers were aligned, but monkey wrenches abound.
I consider all of that poor alignment. It certainly is if you talk to the Accountants, who expect every single change to be perfectly costed and documented at all times. I generally don't believe in that approach, but in
most situations I've experienced the alignment truly isn't good enough. Another hour or two improving the scope definition or estimation will save dozens of hours (or thousands of dollars) shortly thereafter.
I forgot = lost margin. There will always be a difference between quoted scope and final bill of materials, a good estimating process will filter out the costly/risky items consistently. If you've built this kind of stuff before it should be trivial to go down the list of major components and make sure you're not missing anything important. At my company we estimate major components by actual cost and add 5% for hardware. That 5% could be higher or lower depending on your past performance.
Untested design = not understanding technical risk and not building in appropriate cost to mitigate the risk. The cost of mitigating a technical risk is almost impossible to estimate accurately, so you either have to cost it conservatively or try to cost accurately knowing that some will overrun and it will all come out in the wash. That's a business strategy decision there. We as engineers should be striving to identify what is a risky untested design and advise what can and cannot be done about it. And still there will be painful escapes because "physics doesn't give a f&(*" about how hard we work or what the issue is costing us.
That said, some businesses develop only untested designs and that makes this element wildly different from businesses where the expectation is a solution made up wholly of well-tested elements. I've worked across that spectrum and they all have a strategy and method for estimating it.
"Winning" bid = I infer that is when you lose the order to a competitor who probably estimated cost very inaccurately. The vendor who 'won' will be bogged down losing money and the customer may still come back to you if that competitor's mistake becomes a problem for the customer. Short term loss = long term opportunity.
Sales / GM / Marketing making late changes = a good technical estimation will make it possible for them to be accountable for those concessions later. (Now whether your company ever chooses to hold them accountable is another matter!)
I'm not ragging on the individuals - in many cases the processes are where the really big estimation errors happen. A lot happens quickly in the days/weeks around a purchase order - which puts a lot of stress on processes, and very few corporate cultures will fully follow their processes when the PO is floating above their heads and end-of-quarter or end-of-year is looming. And many customers know that is the perfect time/place to put pressure on a vendor, getting as much as possible at the last minute, taking advantage of the vendor's lack of self-regulation.
In the real world, with well-run processes, major jobs have a post-mortem to assess what went well, what did not, and whether a process change would be worthwhile to mitigate the risk of it happening again. That's how you really learn how to estimate well for your business needs. Not all companies have the discipline to do this.
