Being very experienced product designer/developer after few talks with the client I could fairly accurately split big project in few comprehensible stages with the price for each. I do not remember ever being off by any significant margin. I've done way too many of projects either directly or as a director during the course of my life so it is mostly an autopilot at this stage for me. Also my clients tend to be real business people with real business needs. They do not really like to go into methodology level. they just look at my resume, and check references. They are much more interested in how I can solve their problem on conceptual levels before really engaging.

I concur. Milestones are comically similar from project to project.

Software in internal development, software available in testing, beta deployment on the customer site with real customers/data, go live in production, first user, first thousand users, etc...

For hardware projects (I did a lot of joint software/hardware) hardware design, ready to manufacture, first prototype, second prototype, first production series, first delivery to customers, etc...

Last but not least, the primary use cases are major milestones, the software allows to do A then B then C. Gotta determine the main use cases of the project as early as possible.

Large projects (10+ people over years) are split into components, each component has its own milestones and should stand on its own as a deliverable. Add major milestones for integrations, as soon as any 2 dependent pieces are in a working state, they need to be integrated together and tested.

I mean, I've worked on plenty 7-8 figure programs for Fortune 50 companies of all kinds and that has not been my experience at all. I've even worked for FAANG clients who didn't know what they wanted until we tell them. The kind of work I did (I am exiting the consulting world as of this week) was very consumer-facing and creative driven. I can estimate tech work as well as anyone, but there's always too many moving parts to a big project to know how everything will fall into place.

I did work on similarly sized contracts but I was at the time a CTO and had like team of 30 on that project under my supervision.

Everything fell into places. But that was after few month long and expensive exploration phase that along qualified for decent contract.

After I went on my own I handle smaller contracts that are easier to eyeball. Well small is relative as in one case licensing fees alone over a period of time went well into 7 figure territory. But yeah I no longer had to direct 30 people and did not have 20 persons on a client side bugging me every day and being under constant stress ;)