• Goal

  • BizOps (Vincent, Sriram) is doing 2026 revenue planning: setting an MRR target ($330K) and working backwards to determine how many leads they need at each stage to hit it.
    • https://media.secondbrain.lelouvincx.com/2026/03/9b7d19a1a916c3868a43f25c1b3939c0.png
    • https://media.secondbrain.lelouvincx.com/2026/03/daf21e663f6916c8cec1a4c418167e6e.png
  • Conversion Rate By Sales Motion: for won deals, what % went through sales-assisted (call first => trial) vs self-serve (trial first => call).
    • Sriram and Vincent intuitively feel call-first has a higher win rate but don’t have data to prove it.
    • https://media.secondbrain.lelouvincx.com/2026/03/233fc4decf6a75e71f71ef06780ce53e.png
  • Why it matters: assess which entry point produces better outcomes => to invest more on.
  • Decision-making questions:
    • Given we need ~59 raw leads/month to hit 330K MRR, how should we allocate between call-first or trial-first deals?
    • Should we consider removing self-serve trials if call-first converts significantly better?

• Sales Motion Classification

  • Leads are classified based on timestamp ordering of first demo call vs first trial submission:
    • call-first: first_call_at < submitted_at
    • trial-first: submitted_at <= first_call_at
    • trial-only: trial submitted, but first_call_at is NULL
    • call-only: call booked, but trial_submission_skey is NULL
  • For same-day events, uses exact UTC timestamps to determine sequence.

• Data Model

  • Main working dataset: DTS106

• {{video https://youtu.be/y7rWVhihlb8?si=3NL2Cnh4CQKxpk21}}
• {{video https://youtu.be/F2c1ORtkmuU?si=uEE403a6bAhP0ZBH}}

• https://justoffbyone.com/posts/measuring-engineering-productivity/
• The system should work for your engineers, not against them.
  • If your measurement system makes engineers less productive, you’re failed.
  • If it makes them feel surveilled and distrusted, you’re failed.
  • If it becomes an end in itself rather than a mean to build better products, you’re failed.

• How does Olist work?
• How does Olist make money?

• Tasks

• ((691029d8-69c2-4627-b93c-7328fcebda27))

• Written guide
  • Surviving the Early Game
  • Surviving the Mid-Game
  • Useful Construction Patterns
  • Dupe Builds
  • Surviving the Late-Game

• wa

• https://public.lelouvincx.com/parquet-how-it-works.png
• {{video https://youtu.be/1j8SdS7s_NY?si=mBplsyA4pzArccsj}}
• [[Predicate Pushdown]]
• Row Group
• Column Chunk
• Page
• [[Parquet Zone Maps]]
• And then there is Delta Lake, which is a storage layer sitting on top of parquet files. In short, we have Delta Lake to manage, provide database’s functionalities, and the files are stored as parquet.

• Gemini: https://gemini.google.com/app/c44f8c1670b8187a

Notes

• Section 4, Result Delivery & Reporting, L2 - “Can create an effective story with data, maintaining objectivity in their analysis”
  • It’s possible not doing inside Holistics but with my side projects for [[Data Journalism]], [[Data Visualization]]