06-reference

mostlymetrics masking your metrics

2026-08-02·reference·source: Mostly Metrics·by CJ Gustafson
cfo-perspectiveinvestor-relationsnarrative-alignmentunit-economicsmetrics-disclosuresanity-check-voice

"Masking Your Metrics" — CJ Gustafson

Why this is in the vault

A CFO-voiced argument that the metrics a company chooses to externally disclose quietly drift from the metrics actually driving the business — and that gap, left unmanaged, forces either an awkward walk-back or "unnatural" decisions made just to keep the number looking right. Direct craft input for Sanity Check's storytelling-CFO voice work and a live example of narrative-vs-instrumentation drift, a pattern RDCO already tracks in its own operating discipline (status-as-fact laundering, cache-stale critic signals).

⚠️ Sponsorship

Top-of-email paid placement for Abacum — an FP&A platform pitching "AI Spaces," AI-generated board-ready dashboards and planning workspaces built on the company's own model/variables rather than generic templates. Explicit sponsor tag: "Mostly metrics is proudly powered by Abacum." Abacum is a known repeat sponsor of this newsletter (previously seen 2026-05-24 sponsoring the same slot with a NetSuite SuiteApp launch). Not a new sponsor. Separately, Koyfin is credited as "our data partner" via an affiliate link for the recurring valuation-metrics section — a data-source credit/affiliate relationship, not a paid editorial sponsor block. The editorial essay itself is unsponsored and not about FP&A tooling. Checked against the known sponsor roster (Brex, Intuit, Samsara, Rivian, MLB, Mostly Talent) — none of those appear in this issue; Intuit and Samsara show up only as ticker names inside the public-comps dataset, unrelated to sponsorship.

The core argument

CJ opens with his daughter tracing paper masks that drift shape with each new tracing — a metaphor for companies whose externally-communicated metrics drift from the ones actually driving the business internally, usually toward a "sexier" narrative (e.g., touting large-enterprise-customer counts while sales is actually comp'd on quarterly cohorted activation numbers). Aurelien Nolf (CFO of Navan, ex-SVP IR at Lyft) is quoted twice: metrics a company doesn't actively "drive" eventually derail, forcing either an awkward walk-back or unnatural business decisions made purely to protect the optics of the number. At Lyft, the fix was consistent investor education on what actually drives growth and profitability, so a bad quarter was never a surprise. The CFO/FP&A takeaway: align management on which metrics get disclosed externally, keep those tightly correlated to the internal metrics the business is actually run on, and proactively narrate the "why" every quarter — not just on the bad-news day — because investors (or any external audience) who lack the right handles will invent their own, worse narrative. Closes with three self-check questions for finance leaders. The issue's recurring second half is a Koyfin-sourced "Weekly Valuation and Efficiency Metrics" roundup across 9 sectors (~132 public companies) covering revenue multiples, CAC payback, Revenue/Employee (rule of thumb >$450k/employee), Rule of 40, and OPEX targets (>25% profitability at scale).

Mapping against Ray Data Co

Direct hit on RDCO's own "narrative drifts from instrumentation" failure pattern. The "Workflow agent output integrity" memory (feedback_workflow_agent_output_integrity) names exactly this drift internally — status: proposal laundered as live fact, FALSE "verified against primary text" stamps, cache-stale critic signals treated as current. Nolf's framing gives the CFO-world name for the same failure: a metric (or a status claim) that isn't actively "driven" by the reporting process eventually derails and forces either a walk-back or a decision made to protect the number rather than reflect reality. The fix pattern is identical too — Nolf's "consistent investor education so a bad quarter is never a surprise" is structurally the same discipline as "one gate per chain MUST hit the primary source." Worth an explicit cross-reference the next time that memory gets revisited: external metric-disclosure discipline and internal agent-output-integrity discipline are the same problem wearing different clothes.

Sanity Check storytelling-CFO voice study. This is the third piece filed against the "narrative must track instrumentation" theme (alongside the storytelling-CFO arc from CFO Secrets). The "what must this audience believe, and does the number I'm showing them actually support that belief" framing is reusable SC craft: an SC draft that shows a number without owning the narrative around it is unfinished by the same test CJ applies to a board deck.

Anti-mapping caveat: this is CJ's home turf (IR/disclosure discipline for VC-backed companies reporting to a board or investor base) — RDCO is a solo-founder operation with no external cap table to manage, so the investor-education mechanism doesn't map literally. The transferable piece is narrower: internal-vs-external narrative drift as a general failure mode, not the specific IR cadence practice.

Related