How to independently estimate the amount of work and the cost of development?

Question

Task Management

I recently received an offer from a friend to develop his "super project". The challenge is that I'm developing an API for their mobile app. Basically, this is a standard application with authorization, categories, content, and a search for nearby establishments.
Prior to that, he only worked in companies and teams where he did not have to independently evaluate the amount of work in terms of time and cost. Tell me where to start the process of estimating time and how to evaluate the work of a java middle developer?

Answer 1

well, everything is simple - you
estimate your hour of work, or a monthly fee at full employment
and approximately estimate the time for development and testing
with a margin of one and a half or two times
development and design is an unpredictable and expensive process,
so normal designers have this motto:

expensive long fine

you either set deadlines with a margin, or work in a hurry and stress
stress and haste are incompatible with high-quality design
for good design it is important to work in a good mood

Answer 2

there are usually 2 ways:

1. Fixed price valuation
   
2. Or Time & Material.
   

The first path starts with the analysis of TK, of course. Sometimes, the customer has some kind of description in the form of a text file, mocap, design, or something else. If not, then you will have to somehow formalize his Wishlist, translate it into an understandable form and write the technical task yourself, not for him, but for yourself, first of all. In it, you describe all the functionality that is required, user behavior, approximate loads, etc.
The more detailed the TK is, the better.
For example:
Bad TK : In the application, you can log in via email / social networks.
Good TK: In the application, you can log in via email and social networks facebook, vk, instagram. The user can be authorized through any social network, and when authorizing through an unattached social network, we must update the user's data.
Next, you break the TOR into separate parts that you can evaluate.
For example :
Authorization by email : 2-3h + 3-5 for each social network.
For those things that you know and have done many times, you lay a smaller spread, then what you didn’t do - more. ON integration with third-party services such as social networks, payment systems, closed APIs, vaguely described things in the TOR, etc. you make a big range, because there is a high risk of problems with access, difficulties in obtaining, outdated documentation, and so on. If something is not clear, it is better to ask the customer in advance and describe possible problems, perhaps you will find another solution.
In the end, you will get 2 estimates - optimistic and pessimistic . You can also add some coefficient by which you need to multiply if your estimates are usually too optimistic ("Yes, there is work for 2 days", and you sit for a week) or vice versa, too pessimistic ("I need a month, but I do it in a week") .
Then you understand the work will take, for example, 250-300 hh, multiply it by your hourly rate. So you got the approximate cost of the work. Don't forget to add the cost of all material costs specific to (hosting, domain name, special hardware, paid API access).
Also, do not forget to take into account how much you can work, otherwise he may think that this is 250hh - this is 1 month of work, but in reality you have a main job / family / cat and you can only give 100hh / month.
2 way harder :
The customer likes to change the concept, but it’s not and it won’t be important for him to quickly test his ideas and he has a bad idea of ​​what will come out in the end - then you just tell him the cost of an hour and how much time you can devote to him. The main difficulty is that you can pile up shit code due to constant changes, and each further edit will take more and more time, and refactoring in this case is very difficult to push through.

Answer 3

Everything is simple with the price - we take the salary, divide by 190 hours, multiply by 1.4 - we get the price of an hour.
Estimating time is not easy. It is necessary
to 1) describe the project - what the customer will receive as an output
2) break the project into tasks, sub-tasks, and so on, until we reach tasks that we can estimate in time.
If the project is not defined, then the EU-but, it will not be possible to estimate by time, and it remains to work on time (which I personally do).

Answer 4

So you are doing the job. As a mid developer, can't you estimate how long this project will take you? Allow for contingencies according to your experience (only you know how often you have unforeseen problems before, only you know how fast you work)
And multiply the resulting time by the salary you expect to receive.

Answer 5

You can rate yourself as you like, even for a million.
But what's the point if Uncle Vasya makes it cheaper? =)
And also, to work for friends - to lose friendship.