Brainforge Knowledge

Data Formatting Principles

4 min read

Zero-Defect Analysis & Formatting Standards

How We Make Data Clear, Trustworthy, and Executive-Ready

Formatting is not cosmetic. It is how we encode clarity, trust, and judgment into our work.

A correct analysis that is poorly formatted or poorly sense-checked is not done.

Zero-defect standard:
Never put a number in front of leadership that you cannot confidently explain without opening your laptop

Units & Scale (Non-Negotiable)

A number without a unit is not a number.

Rules

  • Metrics should explicitly label unit whether that be in the column or row header or in the cell itself
  • Never mix units within the same column
  • If units change, visually separate the table or section

Gut-check

If this table were screenshotted, would the meaning of every number still be obvious?

Precision & Rounding

Precision should reflect decision quality, not data availability. We also want the key takeaways from data to be clear (vs. muddying with extra digits).

Rules

  • Percentages: 0–1 decimal max
  • Dollars: whole numbers rounded

Red flags

  • Excess decimals
  • Different rounding for comparable numbers
  • False precision in forward-looking estimates

Formula Transparency vs Hard-Coded Inputs

If a model can’t be audited, it can’t be trusted.

Rules

  • Clearly distinguish calculated vs input cells (shading or labels)
  • No mixing formulas and hard-codes in the same column
  • Assumptions live in one clearly labeled place
  • No buried constants in formulas

Gut-check

Could someone else recreate this logic without asking me questions?

Time & Date Logic

Time comparisons must be structurally correct, not visually convenient.

Rules

  • Use formulas for time windows (quarters, last 13 weeks, YTD)
  • Explicitly label time logic (calendar vs fiscal)
  • Never manually select date ranges for recurring analysis
  • Logic should hold when the file refreshes next period

Red flag

  • “I just highlighted the last 13 rows”

Usability & Readability

If an executive can’t scan it in 30 seconds, it fails.

Rules

  • Freeze headers and key identifier columns
  • Collapse rows/columns that aren’t needed often
  • Organize left-to-right or top to bottom: inputs → calculations → outputs (or rows of data to show a funnel)
  • Avoid excessive horizontal scrolling by organizing the data well and not freezing too many columns at the beginning

Gut-check

Can someone understand the structure without reading every cell?

Zero-Defect Sense-Checking (Before Sharing)

Every number must pass three checks.

  1. Plausibility: Is this even possible? If it feels off, stop.

    1. Order-of-magnitude (10×) check

    2. Percentages within bounds

    3. Directionality makes sense

  2. Consistency: Does this make sense relative to other numbers? If I flip to another slide, will this contradict me?

    1. Compare to prior periods, forecasts, or other tabs/decks

    2. Same metric = same answer everywhere (or clearly labeled why not)

    3. Any change >5–10% needs an explanation

  3. Coherence: Do the story and the math agree? “If I say this out loud, do I believe it?”

    1. Headlines match the data

    2. Drivers ladder to outcomes

    3. No orphan metrics

Cross-File & Cross-Deck Discipline

  • Declare a source of truth

  • Explicitly label known differences (gross vs net, bookings vs revenue)

  • No silent number drift between versions

Variance Discipline (No Silent Surprises)

Unexpected results require extra rigor.

Rules

  • Any material variance from expectation must be:
    1. Explicitly called out
    2. Either explained or labeled as under investigation
  • Never “bury” a surprise number in a table
  • Never assume leadership will catch it themselves

Material variance includes:

  • Large deviation vs forecast, budget, or prior trend
  • Directional surprise (up when down was expected, or vice versa)
  • Step-change without a known driver

Acceptable states

  • “This variance is explained by X and Y”
  • “This variance is real but under investigation — here’s what we know so far”

Unacceptable state

  • Presenting a surprising number with no commentary

Gut-check

“If this surprises me, it will surprise leadership — have I addressed that?”

Final Readiness Check (Required)

Before sharing:

  • Can I explain what this is?
  • Can I explain why it changed?
  • Can I explain what it implies?

If not, it’s not ready.

Why This Matters

Zero-defect analysis protects:

  • Your credibility
  • Leadership trust
  • The team’s time

This is not perfectionism.This is professional reliability.

Graph View