Startup Board Pack

Startup board pack

A board pack is a decision document, not a data dump. Its job is to get a small number of important decisions made and to keep the board from being surprised.

The governing rule:

No surprises. A board that learns about a problem in the meeting learns two things: the problem, and that you sit on problems. The second is more expensive. Bad news goes early, quantified, with a plan.

Load saas-metrics for definitions and tie-out for the proof obligation. Every number in a board pack must be tied out to closed figures before it leaves the building.

The arc

Board attention decays fast. Order by importance, not by chronology or by org chart.

Sections 1-3 must survive on their own. Assume some readers get no further.

The scoreboard

One slide, 6-10 lines. Every line: actual, plan, prior period, and a verdict.

                        Actual    Plan     Prior     vs Plan    Verdict
ARR                     $4.82M    $5.10M   $4.61M     (5.5%)    Behind - new logos light
Net new ARR             $210K     $290K    $255K      (27.6%)   Behind - see growth section
NRR (TTM)               104%      108%     106%       (4 pts)   Watch - contraction in SMB
Gross margin            76%       75%      75%        +1 pt     On track
Net burn                $412K     $455K    $388K      +9.5%     Better than plan
Runway                  14.2 mo   13.0 mo  15.1 mo    +1.2 mo   On track
Cash                    $5.85M    $5.92M   $6.26M     (1.2%)    On track
Headcount               47        50       45         (3)       Behind on hiring - deliberate

The verdict column is the value. A board reading a table without verdicts has to derive your judgment, and different members will derive different ones.

Use exactly four verdicts: On track, Watch, Behind, At risk. Define them once — for instance, Watch = within 10% of plan but trending wrong; Behind = missed plan, plan for recovery exists; At risk = missed plan, no credible recovery inside the period. Consistency across meetings is what makes them meaningful.

Cash and runway

For a startup this is the slide that matters most, and it is the one most often softened.

State plainly:

Cash on hand (2026-07-31)                      $5,850,000
Net burn, trailing 3-month average               $412,000/mo
Runway at current burn                              14.2 months
Runway at plan burn (hiring resumes Q4)             11.8 months
Zero-cash date at current burn                  2026-10-02
Next raise: target open 2026-12, close 2026-Q2 (leaves ~6 months buffer)

Rules:

• Show runway on both current and planned burn. Planned burn is usually higher; the gap is the real decision.
• Give a zero-cash date, not just a month count. A date is concrete; "14 months" is abstract.
• Never present runway on a single month's burn. Use a trailing 3-month average and say so.
• If runway is under 12 months, it goes in section 1 as an ask, not in section 3 as information.

Stage-appropriate framing

The same number means different things at different stages. Always frame against your stage, and say which stage you are framing against.

Do not report a metric your data cannot support. LTV:CAC computed on 11 customers with 5 months of history is a made-up number wearing a suit. Say "not yet meaningful at n=11" instead. Boards respect that; they do not respect a number that falls apart under one question.

Presenting a miss

There is a shape to this and it works:

1. The number, first and unhedged. "We closed $210K net new ARR against a $290K plan. We missed by 28%."
2. Why, decomposed. "Two of the three enterprise deals we planned slipped: Acme ($55K, now signed 08-14) and Northwind ($40K, verbal commit, contracting). The third, Delta ($30K), we lost to a competitor on integration depth."
3. Recurring or one-time. "Two are timing; one is a real loss. Adjusted for the slips, we were ~$45K light, which is 15% not 28%."
4. What changes. "August already has $95K of the slipped ARR closed. We are holding the Q4 number. The Delta loss goes to the product roadmap discussion in section 7."
5. The trigger. "If September net new is under $240K, we cut the Q4 hiring plan by three roles. That decision point is 10-05."

Never lead with the explanation. Lead with the number. Explaining before disclosing reads as managing the board, and boards notice.

Charts that earn their slide

Per fin-artifact for rendering conventions:

• ARR bridge (waterfall) — the single best growth visual. New / expansion / contraction / churn.
• Cohort retention heatmap — with cohort sizes shown, small cells suppressed.
• Cash and runway line — actual to date, projected forward on both burn scenarios, zero-cash date marked.
• Actual vs plan with a variance band — not two lines the reader has to subtract mentally.
• Rule of 40 scatter over time — growth on one axis, margin on the other, your path traced.

Charts that do not earn a slide: pie charts of revenue mix, dual-axis anything, 3D effects, month-by-month tables of every GL account, and any chart whose caption is a restatement of its title.

Pre-send checklist

• Every number ties out to closed figures (tie-out blocks exist and PASS)
• Metric definitions are unchanged from last period, or the change is flagged and history restated
• The ask slide contains actual asks, not a summary
• Runway appears in the first three slides
• Every miss has a driver, a classification, and a trigger point
• No customer-identifying detail below the aggregation threshold (finance-guardrails Rail 5)
• No employee compensation detail
• Appendix contains the definitions page
• Sending is a human action. Per Rail 1, this skill assembles and stages the pack. A human sends it

Degraded mode

Entirely a reasoning and structure skill. With no tooling it produces the pack outline, the scoreboard structure, the framing, and the drafted commentary — the numbers come from wherever you can get them, with unverified figures explicitly marked as such.

Related skills

• saas-metrics — definitions and benchmarks for the scoreboard
• flux-analysis — section 5 commentary
• three-statement — the appendix statements and their integrity checks
• fin-artifact — building the pack as a live or snapshot artifact
• finance-guardrails — Rails 1 and 5 govern sending and disclosure