How to correctly calculate the cost based on statistical data

Question

Mathematics

There is a survey with the question: how much will it cost to write a program according to the attached TOR.
There is a list of answers:
25,000, 32,000, 40,000, 45,000, 47,000, 50,000, 50,000, 60,000, 60,000, 80,000, 80,000, 120,000, 150,000, 185,000, 200.200 000, 370,000, 555,000.
Please tell the old engineer, who has forgotten the matan, how EXACTLY (by eye I can estimate this series of numbers myself and give an answer) to calculate the cost of work that will suit the majority of respondents.

Answer 1

1) We calculate the estimate of the expectation as an arithm. the average of all values
​​2) Calculate the unbiased estimate of the variance , remembering that since this is a survey, then it is biased and you need to divide in the formula not by (N) but by (N-1)
3) Now, if you want to know what cost you need to specify so that it suits K% of customers on average (for example, 80%), then you need to calculate (according to the data obtained in paragraphs 1 and 2) the left (if you are a performer) border of the quantile interval (100-K%) - this is and will be the desired value. If you set 50%, it will be exactly the expectation, and every second customer will agree, and every second will refuse.
If you want, publish your "desired percentage of customer coverage" and I will write you an answer for the above data.
PS True, I have a feeling that here, according to the data, the distribution is more lognormal than normal.

Answer 2

Thank you very much, Andrey!
I probably didn't fully explain my purpose. I would like to find not the cost that satisfies a certain percentage of customers, but rather the cost that has the highest weight among all. Speaking not mathematical language - the most adequate cost, based on the available data.