Post 3 — The real 10/20/70 rule

  1. Print to PDF: Click the button below → Save as PDF (1080×1080px per slide).
  2. Upload to LinkedIn: New post → document icon → upload PDF → swipeable carousel.
  3. Post tip: Paste the post text from linkedin_posts.html → put the task2vec.com link in the first comment.
Slide 1 / 7
Post 3  ·  378 Engineers · Apache Camel, Spark & Hadoop
The real 70/20/10 rule.
70
No change
/
20
Coasting
/
10
Growing
We analysed 45,000 Jira tickets and git commits across three Apache projects and tracked whether engineers actually grew into harder work over their careers.
1 / 7
Slide 2 / 7
The study — Apache Camel, Spark & Hadoop · 2007–2022
We matched Jira tickets to git commits and tracked complexity over career
378
engineers with 10+
linked tickets
45k
tickets matched
to git commits
15yr
of engineering
history
Complexity score per ticket: how many modules were touched, whether core files were changed, breadth of impact across the codebase. Then for each engineer: is their complexity score rising, falling, or flat over their career?
2 / 7
Slide 3 / 7
Here is the real rule.
11%
Growing
Took on progressively harder work — more modules, more core codebase, and got faster at it over time
20%
Coasting
Drifted toward simpler work — fewer modules, less core, getting slower even on easier tickets
70%
No effect
Work experience had no measurable impact on the complexity they could handle
3 / 7
Slide 4 / 7
The 11% — what growing actually looks like
The work got harder. Then they got faster.
1
Repetitions in one area
Engineer works on similar ticket types repeatedly — same cluster, same module. Resolution time starts to decline. Pattern recognition building.
2
Velocity improves, scope expands
Tickets take less time AND touch more files, more modules. Getting faster while handling broader scope — the transfer signal.
3
Complexity ceiling rises
Cross-cutting problems, core modules, more people consulted per ticket. The work itself raised the ceiling.
The growth curve is measurable before the advancement happens — it can be designed, not just hoped for.
4 / 7
Slide 5 / 7
The 20% — coasting in slow motion
Productive. Closing tickets. Coasting.
Complexity score
— — tickets closed: unchanged
Yr 1
Yr 2
Yr 3
Yr 4
Yr 5
Yr 6
Yr 7
Ticket count: normal — closing work at a steady pace, no red flag in any dashboard
Complexity declining — fewer modules per ticket, less core codebase involvement each year
Resolution time rising — getting slower even as the work gets simpler
5 / 7
Slide 6 / 7
The implication — growth can be designed, not just hoped for
The answer is already
in your own data
Who is in which trajectory? Your Jira and git history already shows which engineers are growing, stagnant, or drifting. No new tooling required.
What work would stretch them? The complexity map shows which modules each person has touched — and which adjacent ones they haven't yet.
No training platform needed. The 11% didn't learn from a course. They got the right tickets. That's designable.
Open source data from Apache — enterprise teams may look different. But the signals come from data every company already has.
6 / 7
What does your team's 10/20/70 look like?
Your ticket history already contains the answer.
We can compute the growth distribution for your team and show you exactly who needs what work next.
Based on 45,000 Apache commits · Camel, Spark & Hadoop · 2007–2022
7 / 7