DevShop!
I had this idea 10 years ago for a game called DevShop and finally decided to implement a minimal version of it.
It's a kind of 'lemonade stand' game, but instead of buying lemons ๐ and making lemonade ๐น, you hire developers ๐ป, testers ๐ฌ and business analysts ๐ฃ, and complete tasks on a kanban board.
Play it online here:
https://devshop.secretgeek.net
Walkthrough
- click "๐ start"
- click "๐ find project ($100)"
- Select on the newly created project in the
Inbox
column of the kanban board. - Select the "๐ค Founder". (In the animated version this would cause the Founder to walk to the board, grab the card, take it back to their desk, work on it, create a bunch of story cards, and take them back to the board, putting them in the
backlog
column.) - Once the story cards are in the
backlog
column, click a card to select it. - Select the "๐ค Founder" again. This causes the Founder to begin developing the card. (In the animated version, they would walk over, get the card, take it back to their desk, swear a lot, etc.)
- The story card eventually ends up in the
Test
column. Now the card needs to be tested. - The founder is capable of developing, testing, and acting as a BA. They don't do any of these things particularly well...
- Select the card in the
test
column, then select the Founder, so it can be tested. - If it passes testing it is done... and you make money ๐ฒ. (It's possible a bug ๐ is found during testing... or if the bug is missed by the tester, the customer can find the bug ๐ once they receive the card.)
- when the project is completed you get a completion bonus ๐ฒ.
- And you can begin another project...
Repeat this until you have enough money to buy a dev or a tester or a ba.
Features include:
- upskilling people ๐ ๐
- training people to increase their efficiency
- giving them 'initiative' so they can sometimes grab cards for themselves.
- โ coffee and donuts ๐ฉ
- ๐ถ dogs and cats ๐ธ
- inflation. (i.e. things will cost more and be higher stakes as the game progresses.)
- (People will get faster as you increase their skills, but jobs will be bigger... thus creating a steady state, a hedonic treadmill, a red queen's race ๐.)
Tips
Some keystrokes....
1. Press 1 to select the first available item in the `Inbox` column
2. Press 2 to select the first available item in the `Backlog` column
4. Press 4 to select the first available item in the `Test` column
..thanks to Richard Mason.
Make sure to buy "โญinitiativeโญ" from the store (when it is available) -- once the workers have initiative you are free to manage the work at a higher level.
Lessons
If you consider it as a tutorial on Kanban or Lean, there are some lessons you can learn by playing this game. These include the following:
- If work took no time to complete, you would only need 1 worker.
- The earlier in the process a bug is found and fixed, the cheaper it is to rectify.
- Bugs should be removed before they get to the customer.
- Work should flow left to right.
- You need higher proportions of workers for steps that take longer (e.g. development in this game)
- The smaller the stories the better. They are less likely to have bugs, and faster to work with.
- Multi-skilled workers are helpful for alleviating bottlenecks.
- In a system where bugs and irregular work happen, bottlenecks can appear anywhere.
- Throughput is a more useful measure than idle time.
- Good teams are self-organizing, and this requires initiative.
- A newbie thinks "I wish I could hire a manager to assign the tasks automatically" -- a wise person thinks "I wish the workers had enough initiative to select the tasks for themselves."
- The system needs to have some slack in it, or there will be no way to respond to bottlenecks.
- A founder, in the early days, needs to be able to do anything. Later employees can be increasingly specialised.
- To state the above more generally: in a small company, or when work is volatile or sporadic, generalists are very valuable (even if their skill levels are not very high). In a larger company, or when work is more predictable, certain specialists are very valuable (and busy).
- Resource levels need to be appropriately balanced.
- Given a choice between doing something well or doing it quickly, it's better to do it well. It saves time soon enough.
- People can only work on one thing at a time. If they worked on more than one thing at a time, it would decrease throughput. Multi-tasking is a waste of time. (But multi-skilling is not!)
- A functioning and growing business can afford to pay a lot for good coffee โ.
Non-lessons
There are many features of the real world, and of workplaces, that are not modelled in this game.
"All models are wrong, some models are useful."
In the real world:
- not all bugs are so easily dealt with: a single bug can do unlimited damage. That's a vital feature of real bugs, but isn't modelled in the game.
- workers are not machines. They have good days and bad days, they have real lives. All of this is abstracted away.
- customers don't pay immediately, sometimes they don't pay at all, even when you've done good work. Life isn't fair like that.
- you have to pay workers after you hire them. Recurring paychecks aren't modelled in the game because they're not the problem I'm trying to demonstrate. But this isn't a sufficient answer to use in the real world.
- skill levels are not a simple ladder, "dev skill == 7". They would be better modelled as a bifurcating directed acyclic graph, where every node represents a different atom of skill or knowledge. But there's no emoji for that currently.
- before you hire a person, you don't accurately know their skill level. Often you don't know afterward either.
- training takes time and it doesn't have a consistent outcome.
- sometimes workers leave.
- projects never really finish. If they went well there will be ongoing support and changes, indefinitely. If they went badly, the lawsuits can also drag on.
- a situation where "any developer can develop any card" is not achievable. But you can take steps to move in that direction.
- procurement of goods in the enterprise is not as simple as clicking a "buy" button.
- workers spend non-zero amounts of their time in meetings, not shown in this game.
- some related practices I find useful are Daily Standups, Retrospectives, Project Pre-Mortems and Code-reviews. There was no way to integrate these into the game without disturbing the flow.
- context-switching takes time and effort.
- some projects are not worth doing at all.
- some bugs are not worth fixing at all.
- some bugs are not bugs.
Regarding the risk of skilled employees leaving, I love this:
"What if we train them and they leave?
What if we don't train them and they stay?"
Contributions
Pull requests not actually welcome at this moment. I may take this commercial, and haven't thought through the implications of accepting PRs on it prior to that.
I love PRs normally. Please send PRs to other projects.