Agile Mastery Vault · Tools · Multitasking Name Game

🎯 Multitasking Name Game

Browser version of the classic simulation by Henrik Kniberg (Crisp, 2011) — shows how costly it is to run many tasks in parallel.

What's coming

Time: ~5 minutes. Best played on a physical keyboard.

Made possible by
Agile Mastery Vault
A knowledge base for Scrum Masters & Agile Coaches — scenarios, articles and AI agent in one offline app.
Learn more →

Step 1: Estimation

Before we begin, a quick gut check:

Remember these numbers — we'll come back to them after the game.

Round 1: Multitasking

📋 Corporate Policy

"Never keep a customer waiting."

The earlier you start a project, the earlier you finish it. (Right?)

How it works

Type on your keyboard. Wrong letters are ignored.

Policy: Never keep a customer waiting
Delivered: 0 / 5
00.0s

Type the letter highlighted in orange on the active card.

Round 1: Results

Your estimate
5 names
Total time
5 names (actual)
Median delivery
per name

🤔 Hmm…

Let's check which factors caused this. The killer question: did any of these really cause the delay?

  • Length of name? — no
  • Complex spelling? — no
  • Bad tools? — no
  • Lousy Developer? — probably not 😉
  • MULTITASKING.

Let's see what happens when we change the corporate policy.

Round 2: Limit WIP

📋 New Corporate Policy

"Limit Work In Process = 1."

You work on one customer at a time. You don't start the next one until the previous is delivered.

How it works

Customer #5 has to wait the longest. Does it bother them?

Policy: WIP Limit = 1
Delivered: 0 / 5
00.0s

Type the entire name of the active customer. When you're done — it'll automatically jump to the next.

🎉 Comparison

Estimate
5 names
Round 1 — total
median per name:
Round 2 — total
median per name:

📊 Delivery chart (Gantt)

Each bar = length of one customer's project. Shorter is better for the customer.

Round 2 — WIP Limit:

💡 What we learned

  • "Start early = finish early" — busted. In Round 2 almost every project started later and finished earlier.
  • Little's Law — doing X things in parallel makes each one take ~X times longer. Even with zero context-switching cost.
  • Quality: a bug in Round 1 wouldn't surface until near the end; in Round 2 — right after the first delivery.
  • Forecasting: halfway through Round 1 you have no idea when anything will be done. Halfway through Round 2 you've already delivered 2–3 cards and can predict the rest.
  • WIP limit doesn't have to be 1. Any limit is better than none. In practice: 1 main project + 1 background (when the main one is blocked).
Enjoyed the game? Support the project ☕
☕ Buy me a coffee
Made possible by
Agile Mastery Vault
A knowledge base for Scrum Masters & Agile Coaches — scenarios, articles and AI agent in one offline app.
Learn more →