Your September OKR to-Do List 📅  🔗 **Source:** [Your September OKR to-Do List 📅](https://mailchi.mp/whatmatters.com/guided-okr-cycle-sep-26) > [!note]+ Document Notes > tags: [okr, goal-setting, performance-management, team-rituals, measurement] > author: Measure What Matters (Audacious newsletter — no individual byline) > source: Measure What Matters / whatmatters.com > reference: https://mailchi.mp/whatmatters.com/guided-okr-cycle-sep-26 > date created: 2026-09-08 15:21 > > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > > # 🏷️ Title: Your September OKR to-do list 📅 > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [Your September OKR to-do list](https://mailchi.mp/whatmatters.com/guided-okr-cycle-sep-26) — the source; the "Audacious" newsletter from Measure What Matters. Captured with the per-recipient Mailchimp tracking parameter stripped. - [How to grade OKRs](https://www.whatmatters.com/faqs/how-to-grade-okrs/) — the scoring method the newsletter summarizes. - [Top-down OKRs and cascading](https://www.whatmatters.com/okrs-explained/top-down-okrs-cascading) — the alignment piece it links for the bottom-up argument. - [Dear Andy: converting KPIs to OKRs](https://www.whatmatters.com/faqs/dear-andy-converting-kpis-to-okrs) — the linked column on aligning a finance team's OKR upward. - [Audacious newsletter signup](https://mailchi.mp/whatmatters/get-audacious) — the subscription funnel the issue closes on. ## 📝 Summary - **What happened.** Measure What Matters — John Doerr's OKR organization — sent the September issue of its "Audacious" newsletter: a three-step ritual for closing out Q3 OKRs (score, self-assess, reflect together), followed by guidance on sourcing Q4 goals more from the front line than from the executive team. - **Why it matters.** It is one of the few pieces of OKR guidance that treats *grading* as a distinct activity from *reflection* and insists on doing both. The operational core is specific enough to act on: numeric bands with color thresholds, a stated trigger for reopening the objective, and a hard cap on company-level OKRs as the lever for bottom-up ownership. - **Key stakeholders.** Team leads running quarterly retros; anyone inheriting a goal cycle mid-quarter; DevRel and OSPO functions whose outcomes resist clean numeric scoring; open-source maintainers who have adopted OKRs from corporate practice without adapting them to volunteer labor. - **Risks and tradeoffs.** The scoring math is weaker than its presentation implies — averaging key-result scores into an objective score assumes equal weighting and commensurable units, which is rarely true and can hide a failed load-bearing result behind several easy wins. Self-assessment is described as bidirectional but is supplied with no forcing function for downward adjustment, so in practice it is an upward channel. And the strongest claims rest on institutional authority (Intel, Google) rather than on any outcome evidence. - **Bottom line.** Take the ritual, particularly the reflection-is-not-grading distinction and the zombie-OKR permission. Do not take the arithmetic at face value: weight your key results before you average them, or don't average them at all. ## 📖 Main Content - **The frame.** The issue positions September as the "back to business" moment — the point where organizations start eyeing year-end results and consider a fourth-quarter sprint. It follows a prior issue on CFRs (Conversations, Feedback, Recognition), and the through-line is that OKRs are a conversational practice with a numeric surface, not a scoreboard with a meeting attached. - **Step one, grading.** Two traditions are named. Andy Grove's Intel practice was binary: was the goal achieved, yes or no, and if no, why. Measure What Matters instead adopts Google's continuous scale, 0 for failure to 1.0 for complete achievement, bucketed into three bands: 0.7–1.0 green (delivered), 0.4–0.6 yellow (progress, short of completion), 0.0–0.3 red (no real progress). Each key result is scored, then averaged to produce the objective score. The stated action trigger is a score below 0.7, which should prompt a team conversation about whether the objective is still worth pursuing and what would have to change. - **The arithmetic is the weak point.** Averaging key results treats them as equally weighted and mutually substitutable. In practice one key result is usually the one the objective actually depends on. Three cosmetic KRs at 1.0 and one load-bearing KR at 0.0 average to 0.75 — green, and above the conversation trigger, for an objective that failed. The newsletter does not mention weighting, and the 0.7 threshold inherits the flaw: a bad average is what gets compared to it. - **Step two, self-assessment.** The initial score is described as binary in character — you hit the metric or you didn't — and self-assessment exists to adjust it for context. The examples given are circumstances outside the organization's control that obscured genuine effort, and, in the other direction, a strong score that was over-inflated. - **Asymmetry, unaddressed.** Both directions are named but only one has a natural advocate. The person adjusting is the person being assessed, and nothing in the process requires or rewards adjusting downward. Without a counterparty — a peer scorer, a calibration round, a required justification for upward moves — this is a grade-inflation channel with a fairness rationale attached. That is a fixable design gap, not an argument against self-assessment. - **Step three, group reflection.** Self-assessments are shared with the team, with prompts: did you adjust, why, what would you do differently, did your read match your colleagues' or surface something they missed. The section closes on the piece's best line — the adjusted scores matter less than the dialogue they provoke. - **Reflection is not grading.** The strongest structural argument in the issue: the temptation at quarter's end is to write the new OKRs and move, and the instruction is a flat "Don't." Reflection is framed as the built-in space for a team to learn, improve, and mark what worked. This distinction is underweighted in most OKR practice, where the retro collapses into the score review and then into planning. - **Sourcing Q4 goals.** Cascading top-down goals produce alignment, but over-alignment is claimed to stifle energy, creativity and individual motivation. The prescription is that more than half of objectives and key results should originate with people on the front line, on the reasoning that proximity to the problem and to the opportunity produces better goals and genuine ownership. The mechanism offered is a constraint rather than an exhortation: cap company-level OKRs at two or three, communicate direction clearly and concisely, and leave the *how* to departments, teams and individuals. - **This is the most transferable idea in the issue.** It converts a cultural aspiration — "we want bottom-up ownership" — into a structural limit that a leadership team can actually enforce on itself. Scarcity at the top creates room below. That mechanism generalizes well past OKRs. - **Carryover and zombies.** An unfinished Q3 objective should be re-evaluated for the next quarter rather than automatically rolled forward, and some will turn out to be no longer relevant. The phrasing — you never *have* to carry over a zombie OKR — is doing real work: it grants explicit permission to kill a goal, which is the thing teams most often lack and most often need. - **Uncertainties and omissions.** No outcome evidence is offered for either grading tradition. No guidance on weighting key results. No treatment of what to do when an objective's key results are genuinely non-numeric, which is the common case in developer relations, community, documentation and design work. No mention of how any of this survives contact with volunteer or part-time contributors, where the sanction for a red score is simply that someone stops showing up. ### 🗒️ Notes - **The two grading traditions are presented as a house preference, not a finding.** Grove's binary method is attributed to Intel; the 0–1.0 scale is attributed to Google and adopted because "the What Matters team adopts the Google approach." Neither is defended on outcomes. That is honest about its own basis but it means the reader is choosing between two authorities rather than between two pieces of evidence. - **The three bands are asymmetric in width.** Green spans 0.3 of the range, yellow 0.2, red 0.3. The green band being the widest, and beginning at 0.7 rather than at completion, is the well-known stretch-goal accommodation — but a newsletter aimed at teams new to the practice never says why 0.7 counts as delivering. - **The conversation trigger is the useful artifact.** "If you fell below .7, have a conversation" is a stated threshold with a stated response. Most measurement guidance omits exactly this, which is what turns a dashboard into wallpaper. It is worth stealing even by people who reject the rest of the scoring scheme. - **Reflection framed as celebration, not just correction.** The instruction is to celebrate what worked alongside identifying what to improve. Small, but it counteracts the standard failure mode where retros become exclusively remedial and people stop volunteering candor. - **"Zombie OKR" is the most portable phrase in the issue.** It names a specific organizational object — a goal that is carried forward because it was carried forward — and naming it makes it killable. - **The bottom-up argument contains a hidden definition.** "When OKRs work really well, more than half… should come from people on the front line" packages an empirical claim inside a definition of success. If your ratio is low, the sentence concludes your OKRs aren't working really well, without ever testing whether the ratio caused anything. Separating the two would make it a checkable claim. - **Editorial quality is noticeably below the standard the content advocates.** Multiple uncorrected errors in ~780 words: a stray space before a period in the scoring line; "Calculating the initial score… is initially a black and white process"; a missing sentence break at "other factors For example"; a stray comma in "circumstances beyond your organization's control that, obscured a strong effort"; "The same might be for a strong score that was over-inflated" (missing "true"); "the adjusted scores are less important than the dialog it provokes" (number disagreement); and a trailing "And celebrate what did work and improve." Not a reason to discount the substance, but a newsletter about disciplined measurement shipping unproofed copy is a small credibility cost it did not need to pay. - **Every outbound link carries `utm_campaign=September`.** This is content marketing for whatmatters.com, competently done and not disguised. It does not distort the guidance; it does mean the piece's job is to move readers to the site, and the guidance is calibrated to be useful enough to earn that. ## 📚 References - [Your September OKR to-do list 📅](https://mailchi.mp/whatmatters.com/guided-okr-cycle-sep-26) — Audacious newsletter, Measure What Matters, September 2026. - [How to grade OKRs](https://www.whatmatters.com/faqs/how-to-grade-okrs/) — the primary methodology page behind step one. - [Top-down OKRs and cascading](https://www.whatmatters.com/okrs-explained/top-down-okrs-cascading) — source for the over-alignment claim. - [Dear Andy: can KPIs ever be OKRs?](https://www.whatmatters.com/faqs/dear-andy-converting-kpis-to-okrs) — linked column on aligning departmental goals upward. - John Doerr, *Measure What Matters* (2018) — the book both grading traditions are drawn from; the Intel/Grove and Google practices are documented there. - Andrew S. Grove, *High Output Management* (1983) — primary source for the Intel binary-grading tradition the newsletter attributes to Grove. ## 🗣️ Paraphrase - **Opening.** A callback to the previous issue on feedback as part of CFRs — Conversations, Feedback and Recognition — with the hope that readers turned some 1:1s into two-way feedback. - **This week.** September reads as "back to business" for many organizations, the moment when year-end results come into view and a fourth-quarter sprint starts to look necessary. - **Step one: grade Q3.** Evaluate OKRs in both 1:1s and team meetings. A retro has three parts: scoring, subjective self-assessment, and reflection. - **Two grading styles.** Many teams grade simply — achieved, yes or no; if no, why not — which was Andy Grove's preference at Intel. The What Matters team uses Google's method instead: score each key result from 0 (failure) to 1.0 (fully achieved). Most scores fall into one of three ranges — we delivered enough, we progressed but fell short, or we didn't get far enough to call it success. - **The bands.** 0.7 to 1.0 is green, meaning delivered. 0.4 to 0.6 is yellow, meaning progress short of completion. 0.0 to 0.3 is red, meaning no real progress. Average the key result scores to score the objective. Below 0.7, talk with the team about whether the objective is still worth pursuing and what would need to change. - **Step two: self-assessment.** The initial calculation is black and white — you made the metric or you didn't — but the number may not tell the whole story. Self-assessment is the space to adjust for other factors: circumstances outside the organization's control that obscured a strong effort, or conversely a score that came out over-inflated. - **Step three: group reflection.** Share self-assessments with the team. Did you adjust, and why? What would you have done differently? Did your assessment match your colleagues', or surface something they hadn't considered? The adjusted scores matter less than the conversation they start. - **Reflection versus grading.** With Q3 ending, the temptation is to skip ahead into Q4 — write the new OKRs and go. Don't. Pausing to reflect is what creates the space for a team to learn and improve, and to celebrate what worked. - **Energy for Q4.** Cascading goals downward aligns a team, but too much alignment can suppress energy, creativity and individual motivation. When OKRs are working well, more than half of the objectives and key results come from front-line people — those closest to the problems and to the opportunities, who understand and own them best. - **The mechanism.** To get more bottom-up OKRs, cap company-level OKRs at two or three. Leadership states clearly and concisely where the organization is going this cycle; departments, teams and individuals supply the how. - **Balancing old and new.** A Q3 OKR that didn't finish is fine — re-evaluate whether it belongs in the next quarter. Some will turn out to be irrelevant or unnecessary. You are never obliged to carry a zombie OKR forward. - **Close.** Next month covers carried-over OKRs in more depth. Happy grading, and here's hoping your team has found a workable rhythm and measurable Q3 results. ## 📰 Newsletter **Score the quarter, then actually stop and talk about it.** The September Audacious issue from Measure What Matters is a compact Q3-closeout ritual, and two things in it are worth lifting even if you never adopt OKRs. The first is a threshold with a consequence attached. Score each key result 0 to 1.0, average to the objective, and **below 0.7 you owe your team a conversation** — is this objective still worth pursuing, and what would have to change? Most measurement advice hands you a dashboard and no trigger, which is how dashboards become wallpaper. A stated number paired with a stated response is the whole difference. The second is the bottom-up mechanism, which is unusually honest about being structural rather than cultural. You don't get front-line ownership by asking for it — you get it by **capping company-level OKRs at two or three** so there's room underneath for anyone else's goals to exist. Scarcity at the top is what creates space below. That generalizes well past goal-setting. And "zombie OKR" is a genuinely useful bit of vocabulary: the goal that survives only because it survived last quarter. Naming the thing is most of what makes it killable. Where I'd push back: **the averaging is doing more work than it can carry.** Averaging key results treats them as equally weighted and interchangeable. Three easy KRs at 1.0 plus the one that actually mattered at 0.0 averages to 0.75 — green, above the conversation trigger, for an objective that failed. Weight your key results before you average, or skip the average and grade the objective directly. The 0.7 threshold is only as good as the number you compare to it. Also worth noticing: self-assessment is described as adjusting in both directions, but only upward adjustment has a natural advocate — the person doing the adjusting. If you use it, pair it with a calibration round or a peer scorer, or it becomes a one-way ratchet with a fairness story attached. For anyone applying this to open-source or community work: the whole scheme assumes key results reduce to numbers and that a red score has consequences. In volunteer contexts neither holds, and the interesting adaptation problem is what replaces them. ## ✅ Action Items #### DevRel - Developer advocacy: define what a 0.7 actually means for advocacy work before the quarter starts, since talks delivered and docs shipped are countable while trust and reputation are not. - Education: run the three-step retro on a workshop or curriculum program specifically, where completion rates flatter the score and learning outcomes are the thing you actually wanted. - Developer experience: weight DX key results explicitly — a docs gap closed is not equivalent in value to a quickstart that halves time-to-first-success, and averaging them says it is. - Developer marketing: adopt the below-0.7-triggers-a-conversation rule for campaign metrics, and write the conversation prompt in advance so it happens when the number is bad. - Community management: separate the reflection meeting from the scoring meeting on the calendar; if they share an agenda slot, reflection loses every time. - Internal/external/cross-project: use the two-or-three company OKR cap as the negotiating position when a partner team asks you to inherit their goals wholesale. #### OSPO - Policy & Governance: decide whether OKRs apply to volunteer contributors at all before cascading them; a red score with no employment relationship behind it is a request, not a goal. - License Compliance: compliance work is a poor OKR candidate — it is binary, continuous and non-negotiable. Grade it with Grove's yes/no rather than the 0–1.0 scale. - Security Management: same reasoning — vulnerability response SLAs should be thresholds, not scored objectives, since a 0.7 on patching is a failure dressed as progress. - Software Inventory: SBOM freshness suits the continuous scale well; set the band boundaries from your own risk tolerance rather than inheriting 0.7/0.4/0.3 unexamined. - Developer Support: publish the retro prompts so teams reflect consistently rather than each lead improvising. - Community Engagement: ask funded projects what *they* would set as objectives before proposing any, which is the front-line sourcing argument applied to grantmaking. - Strategic Alignment: cap org-level OKRs at three and enforce it against yourself first; the cap only produces bottom-up room if leadership actually holds to it. #### Community - Maintainers: run a quarterly zombie-OKR sweep — list every goal carried forward twice and require a live justification for each or drop it. - Contributors: reflection sessions are where new contributors learn what the project values; make at least one per cycle open rather than maintainer-only. - Sponsors: report both the score and the self-assessment adjustment to sponsors, with the reasoning; an unexplained 0.6 reads worse than an explained one. - Ecosystem: compare band definitions across adjacent projects before treating any cross-project score as comparable — the same 0.7 usually means different things. #### WordPress - Core: distinguish release-criteria thresholds (binary, non-negotiable) from improvement objectives (scored) rather than running both through one grading scheme. - Core Performance: performance budgets are a good match for the continuous scale; set the bands from field data rather than from the newsletter's defaults. - Core AI: any AI-related objective needs a stated falsifier at the start of the cycle, or self-assessment at the end will rationalize whatever happened. - **Core Program Team:** own the reflection ritual across teams — it is exactly the coordination work that no single team will do for itself under deadline. - Mobile: weight key results by user impact before averaging; app-store metrics and internal delivery metrics are not commensurable. - Playground: a good venue for the "what would we have done differently" prompt, since experiments there are cheap enough to actually redo. - Community: bring the zombie-OKR permission to meetup and WordCamp planning, where recurring commitments persist mostly through inertia. - Training: teach the reflection-versus-grading distinction directly; it is the part practitioners most consistently skip. - Marketing: apply the below-0.7 conversation trigger to campaign objectives rather than reporting the number and moving on. - Support: response-time targets belong in the binary tradition, not the scored one. - Design: design outcomes resist numeric key results more than most; consider paired qualitative assessment rather than forcing a score. - Documentation: docs coverage scores well numerically and badly qualitatively — pair every coverage KR with a task-success measure or the average will lie. - Photos: submission and moderation throughput fit the continuous scale cleanly; this is one of the easier adoptions. - TV: caption quality and content indexing both suit scored key results with explicit weighting. - Polyglots: locale coverage percentages are a natural fit, but weight by locale usage or a long tail of small locales will dominate the average. - Accessibility: audit findings closed is a countable KR that can hide severity — weight by WCAG level and user impact before averaging, which is the same critique as this issue's central flaw. - Meta: instrument whether teams that hold reflection sessions separately from scoring show different carryover rates; the project has the data to check this. - CLI: command coverage is straightforwardly scoreable; a good candidate for a first OKR cycle if a team wants a low-risk trial. - Hosting: cross-host checklist adoption is a bottom-up OKR candidate — the hosts closest to the problem should set it. - Tide: scoring transparency is itself the subject here; publish how bands are defined, not just the resulting scores. - Openverse: ingestion and moderation volumes score well; provenance quality does not, and shouldn't be forced to. - Themes: review throughput as a scored KR risks trading speed for review quality — pair it with a defect-escape measure. - Plugins: same pairing; throughput alone is exactly the metric this newsletter's own averaging flaw would let a team hide behind. #### The WPOCC - Programs: fellowship and stipend programs should set objectives with recipients rather than for them — the front-line sourcing argument, applied where it matters most. - Projects: governance clinics are a natural place to teach the two-or-three cap, since over-proliferation of stated priorities is the common failure in volunteer organizations. - Funding experiments: milestone payouts map almost directly onto key results; adopt explicit weighting from the start rather than discovering the averaging problem mid-grant. #### FAIR - Technical independence: independence objectives are long-horizon and quarterly scoring will misrepresent them; consider annual objectives with quarterly key results rather than forcing the cycle. - Policies and covenants: a covenant is a threshold, not a scored goal — grade it yes/no in the Grove tradition. - Productized governance artifacts: artifact delivery is countable and suits the continuous scale; adoption is the outcome that matters and needs its own measure. ## 🧪 Hypotheses - **H1.** Teams that hold reflection as a separately scheduled session score no better in the following quarter but carry forward fewer objectives, because reflection's effect is on pruning rather than on performance. - **H2.** Objectives whose key results are explicitly weighted show larger divergence between the weighted score and the simple average as the number of key results rises — quantifying how much information averaging destroys. - **H3.** Self-assessment adjustments are net upward across a population of teams, and the asymmetry grows where the assessment is visible to the person's manager. - **H4.** Capping company-level OKRs at two or three increases the proportion of team-authored objectives, and the effect is mediated by the cap rather than by any stated encouragement of bottom-up goal setting. - **H5.** Teams using binary Grove-style grading report higher confidence in their scores and lower confidence in cross-team comparison than teams using the 0–1.0 scale. - **H6.** In volunteer or open-source contexts, red-scored objectives predict contributor attrition rather than corrective action, inverting the intended function of the grade. ### 📊 Methodology - **For H1:** Cohort comparison across teams within one organization over four quarters. Independent variable is whether reflection is separately calendared; measure carryover rate, objective count, and next-quarter scores. Carryover rate is the discriminating outcome. - **For H2:** Purely computational — no fieldwork needed. Take any historical OKR dataset, elicit weights retrospectively from the owning team, and compute the divergence between weighted and unweighted objective scores as a function of key-result count and score variance. The claim stands or falls on arithmetic. - **For H3:** Compare pre- and post-adjustment scores across a population, testing whether the mean adjustment differs from zero. Split by whether the assessment was private or manager-visible to test the moderator. - **For H4:** Interrupted time series on organizations that introduce the cap, measuring authored-by-level proportion before and after. Needs a control for organizations that merely exhorted bottom-up goal setting without capping. - **For H5:** Survey instrument administered to both populations measuring confidence in own score and confidence in cross-team comparability, with the two questions kept separate — the hypothesis predicts they move in opposite directions. - **For H6:** Longitudinal contributor-activity data in open-source projects that adopted OKRs, comparing activity trajectories after red versus yellow scores. Ethically sensitive and best done with project consent and aggregate reporting only. ### ❓ Questions - Where did 0.7 come from? It is stated as the delivery threshold and as the conversation trigger without derivation, and it is doing a great deal of work in the scheme. - What is the right procedure when key results are not commensurable — weight them, grade the objective holistically, or refuse to average and report the vector? - What would make self-assessment symmetric in practice rather than only in description? - Does the two-or-three cap work because of the number, or because a cap of any kind forces leadership to choose? A cap of five might produce most of the benefit. - How should any of this be adapted where labor is voluntary, the sanction for a red score is departure, and "the front line" is self-selected rather than assigned? - Is the reflection-versus-grading distinction actually about the calendar? The failure mode described — rushing to Q4 — is a scheduling problem, and it may have a scheduling fix rather than a cultural one. ### 🔮 Future Research - Empirical work on OKR weighting: how much objective-level signal simple averaging destroys, across real datasets rather than constructed examples. - Adaptation of goal-setting frameworks to volunteer and mixed paid/volunteer organizations, which is a live and largely unaddressed problem across open source. - Whether stated response thresholds — "below X, do Y" — improve metric follow-through independent of the metric's own quality. The general form of the most useful idea here. - Study of self-assessment calibration mechanisms borrowed from performance review research and applied to goal scoring. - Longitudinal work on zombie goals: how long objectives actually persist unexamined, and what interventions kill them. ### 💡 Implications - **Practice.** Adopt the threshold-plus-response pattern immediately; it costs nothing and it is the difference between a measured system and a monitored one. Weight key results before averaging, or report them unaveraged. - **Policy.** Organizations importing OKRs into governance contexts should decide explicitly which commitments are thresholds (binary, non-negotiable) and which are objectives (scored). Grading a compliance obligation on a 0–1.0 scale quietly authorizes partial compliance. - **Security.** Anything with an SLA belongs in the binary tradition. A 0.7 on patch response is not seventy percent of safe. - **Education.** The reflection-versus-grading distinction is the most teachable idea here and the most commonly skipped in practice — good curriculum material, and it transfers to retros generally, not just to OKRs. - **Funding.** Milestone-based grantmaking is structurally identical to key results, which means it inherits the same averaging flaw. Funders who average milestone completion into a project health score are making the exact error described above. ### 🧠 Learning Styles Assessment - **Reading level estimate:** Flesch Reading Ease **72.0**, Flesch-Kincaid Grade **6.1**, computed over the 692-word body (60 sentences, 11.5 words/sentence, 1.46 syllables/word). Deliberately plain — appropriate for a newsletter meant to be actioned the week it arrives rather than studied. - Structurally strong for scanning: emoji-prefixed headings, numbered steps, short paragraphs, and a bulleted scoring key. #### 👀 Visual learners - Present: the three color bands are the one genuinely visual device, and they map numbers to a familiar traffic-light schema that carries meaning without explanation. - Improve: a single diagram of the three-step cycle, or a worked example showing four key results scored and averaged, would do more than any prose here — especially since a worked example would expose the weighting problem. #### 🔊 Aural learners - Present: nothing — email text only. - Improve: the content is already conversational enough to read aloud unchanged; a three-minute audio version would suit the "do this during your commute before the retro" use case. #### 🗣️ Verbal learners - Present: strong. CFR is expanded on first use, "zombie OKR" is a memorable coinage, and the Grove-versus-Google contrast is a clean verbal frame. - Improve: define "key result" and "objective" at least once per issue for readers arriving mid-series, and gloss what a 0.7 means before using it as a threshold. #### 🧩 Kinesthetic learners - Present: the strongest dimension. Three numbered steps, a concrete scoring procedure, and explicit meeting prompts make this directly executable. - Improve: ship a one-page retro template or scoring worksheet; the newsletter describes an artifact it never provides. #### 🧠 Logical learners - Present: a clear procedure with numeric thresholds and a stated decision rule at 0.7. - Improve: the averaging step needs justification it never receives, and the band boundaries need a derivation. A logical reader will notice both gaps immediately and may discount the rest. #### 🧑‍🤝‍🧑 Social learners - Present: excellent. The entire second and third steps are structured group activity with named discussion prompts and explicit attention to whether colleagues' readings converge. - Improve: guidance on facilitating when self-assessments conflict sharply, which is the moment the ritual actually gets tested. #### 🧘‍♀️ Solitary learners - Present: step two is genuinely individual, with reflective prompts that work alone. - Improve: nothing for someone practicing OKRs without a team — a common case for independent consultants, and one this framing never acknowledges. #### 🌿 Naturalistic learners - Present: none directly, though "zombie OKR" is an organic metaphor for accumulation and the pruning frame is implicit throughout. - Improve: make the pruning metaphor explicit — a goal portfolio needs seasonal cutting back, and that framing would land the carryover section harder than the current phrasing does. ### 📊 Accessibility Assessment #### Screen reader semantics - Present: a heading structure with descriptive section names, and short paragraphs in reading order. - Improve: headings are prefixed with decorative emoji (🚦, 🕑, 🪞, 💥, ⚖️) that screen readers announce by name — "traffic light Step 1: Grading Q3 OKRs." Two of the step headings reuse the same 🕑 clock, so the announcement adds noise without distinguishing them. Mark decorative emoji up so they are skipped, or move them out of headings. #### Keyboard navigation - Present: an email rendered as a web archive; links are native anchors in document order with no interactive widgets. - Improve: nothing observed. Confirm the Mailchimp archive template exposes a visible focus indicator, which its defaults do not always guarantee (SC 2.4.7, 2.4.11). #### Contrast and readability - Present: not verifiable from captured markup. - Improve: Mailchimp templates commonly set footer and preference text in light gray on white, which typically fails 4.5:1. The unsubscribe and preferences links are the specific risk, and they are the ones a reader most needs to be able to find (SC 1.4.3). #### Text resizing and reflow - Present: table-based email layout, which is the norm for the medium. - Improve: verify reflow at 320px and 200% zoom; fixed-width table layouts are the standard failure point here and email templates frequently do not reflow (SC 1.4.10, 1.4.4). #### Media and captions - Present: a header image with no meaningful text content, so no caption obligation is triggered. - Improve: confirm the header image carries either descriptive alt text or an explicit empty alt if decorative. Mailchimp defaults frequently emit the filename, which is worse than nothing (SC 1.1.1). #### Forms and interactions - Present: no forms in the body; the signup and preference links are plain anchors. - Improve: the linked signup and preference-center pages need their own labeling and error handling review — out of scope for this capture but in scope for the sender (SC 3.3.1, 3.3.2). #### Standards reference - Gaps map to WCAG 2.2 AA: **1.1.1** (header image alt), **1.4.3** (footer and utility-link contrast), **1.4.4 / 1.4.10** (email table reflow at zoom), **2.4.7 / 2.4.11** (focus visibility in the archive template), plus a non-normative but real usability issue with decorative emoji inside headings. Nothing observed is a content-level Level A failure; the risks live in the email template. ### 🌍 DEIB Assessment #### Cultural breadth - Present: minimal. "Back to business" in September, quarters aligned to a calendar year, and a fourth-quarter sprint are all assumptions that hold in some markets and not others. - Improve: fiscal years differ widely, and September is a holiday-heavy or monsoon-season month in parts of the world. Naming the assumption once would cost a sentence. #### Inclusive language - Present: good. Gender-neutral throughout, plain vocabulary, jargon expanded on first use (CFR). - Improve: "Hi team!" assumes the reader leads or belongs to a team; individual contributors and solo practitioners are addressed by implication only. #### Neurodiversity support - Present: strong. Numbered steps, consistent structure, short paragraphs, clear visual chunking, and an explicit decision rule rather than an implied one. - Improve: the copy errors noted above disrupt exactly the readers who parse literally — a missing sentence break at "other factors For example" and the garbled "The same might be for" cost more comprehension than they would for a skimming reader. #### Representation - Present: two named exemplars, Andy Grove and Google, both from a narrow slice of American technology management. - Improve: goal-setting practice outside large US tech — public sector, nonprofits, cooperatives, volunteer-run projects — is entirely absent, and those readers get a framework with no adaptation guidance. #### Psychological safety - Present: the piece's best dimension. It explicitly makes room for scores that don't tell the whole story, invites disagreement in group reflection, instructs teams to celebrate what worked, and grants permission to kill goals rather than carry them in guilt. The framing throughout is learning-oriented rather than accountability-oriented. - Improve: it does not address power. Sharing a self-assessment with a team that includes your manager is a different act than sharing it with peers, and the guidance treats them identically. One sentence acknowledging that would strengthen the rest. ### ✅ Logical Validity #### Strong reasoning - **Procedural, threshold plus response.** "If you fell below .7, have a conversation with your team: Is the Objective still worth pursuing?" is strong because it binds a measurement to an action in advance. A threshold with a predetermined response resists post-hoc rationalization in a way that a reported number does not. To clarify further: state what happens at a green score too, so the rule is symmetric rather than only a failure trigger. - **Causal, over-alignment to lost motivation.** The claim that cascading goals produce alignment but excess cascading suppresses energy, creativity and ownership is a plausible mechanism with a specified pathway — distance from the problem reduces both goal quality and felt ownership. It is stated as mechanism rather than assertion, which makes it testable (H4). - **Structural, cap as the lever.** Recommending a hard limit of two or three company-level OKRs rather than exhorting bottom-up culture is the piece's most rigorous move. It identifies a constraint that produces the desired behavior without depending on anyone's good intentions, and it binds leadership rather than the front line. - **Distinction, reflection versus grading.** Separating the two, and instructing readers explicitly not to skip to planning, correctly identifies that these activities compete for the same meeting slot and that one reliably loses. The "Don't." is doing real rhetorical work. #### Fallacies - **Appeal to authority** — the entire basis for both grading methods. "This was Andy Grove's preferred style at Intel"; "The What Matters team adopts the Google approach." Neither is supported by any outcome evidence, and the reader is offered a choice between two institutional reputations rather than two results. **Fix:** either cite evidence that one method produces better decisions, or state plainly that this is convention rather than finding — which would cost nothing and gain credibility. - **Unsupported precision** — the 0.7/0.4/0.3 boundaries and the 0.7 action trigger are given to one decimal place with no derivation. Presenting uncalibrated cutoffs at that resolution implies a measurement basis that is not in evidence. **Fix:** say where the thresholds came from, or present them as conventions readers should calibrate to their own risk tolerance. - **Suppressed premise (the averaging step)** — "After each Key Result is scored, take their average to score the Objective" silently assumes key results are equally weighted and commensurable. Three cosmetic KRs at 1.0 plus one load-bearing KR at 0.0 yields 0.75, which is green and above the conversation trigger, for a failed objective. The scheme's own action rule can be defeated by its own arithmetic. **Fix:** require weights, or grade the objective directly rather than deriving it. - **Circular definition / unfalsifiable framing** — "When OKRs work really well, more than half of the Objectives and Key Results should come from people on the front line." Success is defined partly by the practice being recommended, so a low ratio proves the OKRs aren't working well rather than testing whether the ratio matters. **Fix:** separate the definition from the causal claim and state the latter as checkable — teams with more front-line-authored OKRs achieve X. - **Asymmetry presented as balance** — self-assessment is described as adjusting in either direction, with over-inflation named. But the adjuster is the assessed party and no mechanism, counterparty or justification requirement is supplied for downward moves. Describing a one-directional incentive as bidirectional is a real, if unintentional, misrepresentation of how the process will behave. **Fix:** add calibration, peer scoring, or a required written rationale for upward adjustments. - **Not a fallacy, recorded for fairness:** the piece is transparently promotional — every link is UTM-tagged to a September campaign — but the guidance is not distorted to favor a product, and nothing is being sold beyond the newsletter and the site. The commercial interest warrants noting, not discounting. #### Tone of voice - Warm, collegial, coaching. Opens "Hi team!", uses first person plural throughout ("we find," "we delivered"), and issues instructions as encouragement rather than mandate — with one deliberate exception, the flat "Don't." at the reflection section, which lands precisely because the surrounding register is so gentle. Closes on "Keep up the momentum." The overall posture is a peer who has done this before rather than an authority issuing a standard, which suits a subscriber audience practicing the method between issues. ### 🛡️ DARVO Analysis - **Deny:** Not present. Nothing in the piece responds to an accusation or minimizes a harm. The closest thing to minimization is the unexamined averaging step, which obscures a methodological weakness — but that reads as an oversight in a short newsletter, not as a denial. - **Attack:** Not present. No person, organization, or competing framework is criticized. Grove's binary method is presented as a legitimate alternative rather than as a foil, which is notably even-handed given that the piece then adopts a different approach. - **Reverse victim/offender:** Not present. No party is cast as wronged. The guidance on missed goals explicitly removes blame — an unfinished objective is "okay," and killing it is permitted rather than penalized. - **Emotional cues and rhetoric:** Low-arousal and consistently reassuring. The persuasive mechanism is permission-granting ("That's okay," "You never have to carry over a zombie OKR") rather than urgency or fear. The single imperative — "Don't." — is directed at a behavior, not at the reader's character. Related fallacies are catalogued above under Logical Validity. - **Assessment: DARVO is not applicable to this source.** This is instructional content with a commercial funnel attached, containing no adversarial or defensive rhetoric. Recorded because the section is structurally required, and because a clean negative result is itself informative — it separates ordinary marketing from the defensive pattern DARVO describes. ## 🎯 Conclusion - **Bottom line.** A tight, usable quarterly-close ritual whose best ideas are structural rather than numerical. The threshold-plus-response rule, the insistence that reflection is not grading, the cap on company-level OKRs as the actual mechanism for bottom-up ownership, and the explicit permission to kill a zombie goal are all worth taking. The scoring arithmetic underneath them is weaker than its presentation implies. - **The flaw that matters.** Averaging key results into an objective score assumes equal weight and commensurability. Because the 0.7 conversation trigger is applied to that average, the scheme's most valuable rule — talk to your team when you fall short — can be silently defeated by its own math. Weight the key results, or grade the objective directly and use the key results as evidence rather than as inputs. - **The second-order issue.** Self-assessment is described as symmetric and structured as asymmetric. If you adopt it, add a counterparty: peer scoring, a calibration round, or a required written rationale for upward adjustment. Otherwise it is a ratchet. - **Next steps for readers.** Before the next cycle, write down your band boundaries and your response rule, and make sure the rule fires on something you cannot manipulate. Cap top-level objectives at three and hold leadership to it first. Calendar reflection separately from scoring, because when they share a slot reflection is what gets cut. And if a goal has been carried forward twice without progress, say out loud that it is dead. ## 🔄 Refactor Appendix — Your September OKR to-do list > Density: 4 Literature atoms (H3), 2 Permanent candidates (H4). > Resonance filter applied: dropped the CFR callback and the "back to business" > framing, which set context rather than advancing a claim. ### 📚 Literature Atoms (H3) #### [H1] Reflection matters more than the adjusted score > [!quote] Source: Audacious newsletter, "Step 3: Group Reflection" > Keep in mind that the adjusted scores are less important than the dialog it provokes. **Context & use**: - Closing line of step three, after the group-reflection prompts. - The clearest statement of the piece's actual thesis: the number is instrumentation for a conversation, not the output. Worth keeping because it is the standard most measurement rituals fail — including ones I've built. It is also the sentence that most undercuts the scheme's own precision about band boundaries, which is an honest tension rather than a contradiction. **Tags**: #xref #aphorisms **Highlight**: anchored inline in Reader on the source paragraph <!-- atomicity: 1 idea ✓ --> #### [H2] Capping top-level goals is what creates bottom-up room > [!quote] Source: Audacious newsletter, "Finding Energy and Motivation for Q4 OKRs" > The trick to encouraging more “bottom-up” OKRs is to limit the number of company-level OKRs to just two or three. **Context & use**: - The Q4 sourcing section, immediately after the over-alignment claim. - The most transferable idea in the issue because it is structural, not cultural. You do not get front-line ownership by asking for it; you get it by leaving room for it to exist. Scarcity at the top as the enabling constraint generalizes well past goal-setting — the same shape as limiting WIP, or capping the number of P0s. **Tags**: #xref #articleseed **Highlight**: anchored inline in Reader on the source paragraph <!-- atomicity: 1 idea ✓ --> #### [H3] Over-alignment suppresses the motivation alignment was meant to serve > [!quote] Source: Audacious newsletter, "Finding Energy and Motivation for Q4 OKRs" > Cascading “top-down” goals will align your team. But, over-alignment can stifle energy, creativity and individual motivation. **Context & use**: - Opens the Q4 sourcing argument and supplies its premise. - Useful as a stated mechanism rather than a slogan: alignment and motivation are not monotonically related, so there is an optimum rather than a maximum. Names a real failure mode in organizations that treat cascade completeness as the goal. **Tags**: #xref **Highlight**: anchored inline in Reader on the source paragraph <!-- atomicity: 1 idea ✓ --> #### [H4] Unfinished goals get re-evaluated, never auto-renewed > [!quote] Source: Audacious newsletter, "Balancing Old and New in Q4" > You never have to carry over a zombie OKR. **Context & use**: - Closing the carryover section. - Keeping this for the vocabulary as much as the rule. "Zombie OKR" names a specific object — a goal surviving only because it survived — and naming it is most of what makes it killable. The permission framing matters too: teams usually know the goal is dead and lack authorization to say so. **Tags**: #xref #aphorisms **Highlight**: anchored inline in Reader on the source paragraph <!-- atomicity: 1 idea ✓ --> ### 🧠 Permanent Candidates (H4) #### [P1] Averaging sub-scores can defeat the threshold built on top of them **Claim** (one sentence, Courtney's voice): When a composite score is a simple average of unweighted parts, any action threshold applied to that composite can be silently defeated by high scores on the parts that didn't matter. **Evidence pointer**: [[#H1]] — the newsletter averages key results into an objective score and then triggers a team conversation below 0.7. Three cosmetic key results at 1.0 and one load-bearing key result at 0.0 average to 0.75: green, no conversation, failed objective. The arithmetic is the whole argument and needs no fieldwork to check. **Scope & conditions**: Holds wherever a composite metric is an unweighted mean over components of unequal importance and a decision rule reads the composite. Does not hold where components are genuinely equally weighted, where the aggregation is a minimum rather than a mean, or where the threshold is applied per-component. The fix is to weight, to aggregate by minimum, or to read the vector instead of the scalar. **Links** (≥1 with annotation): - [[Alarm fatigue before adding alarms]] — because contrasts on the same defect from the other end: there a constant alarm carried no signal, here a composite hides signal that exists. - [[Site health audit methodology]] — because applies-in: my own audit scoring aggregates unequal checks, and I should verify it doesn't have this shape. - [[Metrics that resist gaming]] — because extends: unweighted averaging is not just lossy, it's the specific structure that makes a metric cheap to game from the easy end. <!-- atomicity: 1 idea ✓ --> #### [P2] Self-assessment without a downward mechanism is a ratchet **Claim** (one sentence, Courtney's voice): A review step that lets the assessed party adjust their own score in either direction will move upward on average, because only one direction has an advocate. **Evidence pointer**: [[#H1]] — step two invites adjusting for uncontrollable circumstances that obscured strong effort, and names over-inflated scores as the symmetric case, but supplies no counterparty, calibration round, or required justification for upward moves. **Scope & conditions**: Holds where self-assessment is unwitnessed and unarbitrated. Weakens where a peer scorer, manager calibration, or public rationale is required — and inverts where the assessment is visible to someone with power over the assessed, since visibility can suppress upward adjustment through a different distortion. The general point is that a described symmetry is not a structural one. **Links** (≥1 with annotation): - [[The maker should not grade its own homework]] — because extends: the same principle I already apply to agent pipelines, appearing here in a human process. The maker/checker split is the fix in both cases. - [[Calibration in performance review]] — because applies-in: the existing literature on this asymmetry is in HR research, not goal-setting writing. <!-- atomicity: 1 idea ✓ --> ### ✏️ Fleeting Queue - [F1] Check whether the site-health-audit scoring aggregates unequal checks into one average — if so it has the [[#P1]] shape and needs weighting or a per-check threshold. #fleeting #inbox - [F2] "Zombie OKR" has an obvious analogue in issue trackers: the ticket carried across milestones purely by inertia. Is there a name for it, and is a periodic forced-justification sweep worth adding anywhere? #fleeting #inbox - [F3] Goal-setting frameworks assume employment as the enforcement mechanism. What actually replaces a red score in volunteer-run open source? Nothing I've read addresses this and it affects every WordPress team that has adopted OKRs. #fleeting #inbox ### → Things 3 (max 2 actions) - [ ] Audit my own audit checklists for unweighted-average scoring — site-health-audit first, then the website-audit playbook — and add weights or per-check thresholds where a composite drives a decision #things - [ ] Before the next quarter starts, write down band boundaries and the response rule for any metric I'm tracking, and confirm the rule fires on something I can't adjust after the fact #things <!-- deep-read:sections end --> 📄 **Full text:** [[Full Document Contents/Articles/Your September OKR to-Do List 📅.md|Your September OKR to-Do List 📅 by Measure What Matters]] ## Highlights - After each Key Result is scored, take their average to score the Objective. ([View Highlight](https://read.readwise.io/read/01m217gny81r8n5zk8azxdceva)) - Note: The scheme's structural flaw, stated in one sentence. Averaging assumes key results are equally weighted and commensurable. Three cosmetic KRs at 1.0 plus the one load-bearing KR at 0.0 averages to 0.75 — green, above the conversation trigger, for a failed objective. The arithmetic defeats the scheme's own best rule. Fix: weight them, aggregate by minimum, or grade the objective directly. → P1 - Tags: fallacy/suppressed-premise, atom/zettel/z1 - If you fell below .7, have a conversation with your team: Is the Objective still worth pursuing? If so, what can we change to achieve it? ([View Highlight](https://read.readwise.io/read/01m217grxqtxftre0gjfgrwgn2)) - Note: Strong: a threshold bound to a predetermined response, which is what separates a measured system from a monitored one — steal this pattern regardless of the rest. Weak: 0.7 is given to a decimal place with no derivation, and it's applied to an unweighted average (see the averaging highlight), so the trigger can be defeated by the math feeding it. - Tags: fallacy/unsupported-precision, atom/literature/h2 - Share self-assessments with your team. Did you adjust the score from the initial calculation? If so, why? What would have done differently if you could have? Did your self-assessment reveal similar insights to others on your team? Or perspectives that others didn’t consider? Keep in mind that the adjusted scores are less important than the dialog it provokes. ([View Highlight](https://read.readwise.io/read/01m217ggp0nhtvsnr5whqpsav1)) - Note: H1 — the piece's actual thesis: the number is instrumentation for a conversation, not the output. This is the standard most measurement rituals fail, including ones I've built. Note the tension: it also undercuts the scheme's own decimal-place precision about band boundaries. - Tags: atom/zettel/z1, atom/literature/h1 - [Cascading “top-down” goals](https://www.whatmatters.com/okrs-explained/top-down-okrs-cascading?utm_source=audacious&utm_medium=email&utm_campaign=September-26) will align your team. But, over-alignment can stifle energy, creativity and individual motivation. When OKRs work really well, more than half of the Objectives and Key Results should come from people on the front line — those who are close to the problems, and close to the opportunities to understand them — and own their own goals — better. ([View Highlight](https://read.readwise.io/read/01m217gwxs67q52p6hbzvk50nj)) - Note: Good mechanism, unfalsifiable packaging. The causal claim — alignment and motivation aren't monotonically related, so there's an optimum rather than a maximum — is real and testable. But "when OKRs work really well… more than half should come from the front line" defines success by the practice being advocated, so a low ratio proves your OKRs aren't working rather than testing whether the ratio causes anything. Fix: separate the definition from the claim. - Tags: atom/literature/h3, fallacy/circular-definition - The trick to encouraging more “bottom-up” OKRs is to limit the number of company-level OKRs to just two or three. As a leadership team, you should communicate very clearly and concisely where you want to go this cycle. It is then up to your departments, teams, and individuals to share in their OKRs of how to get there. ([View Highlight](https://read.readwise.io/read/01m217gk7vqh90kaq9skenvh90)) - Note: H2 — the most transferable idea here, and the most rigorous move in the piece: structural, not cultural. You don't get front-line ownership by asking for it; you get it by leaving room for it to exist. Scarcity at the top as the enabling constraint — same shape as limiting WIP or capping P0s. Binds leadership rather than the front line. - Tags: atom/zettel/z1, atom/literature/h2 - Did you have a Q3 OKR that didn’t make it across the finish line? That's okay — re-evaluate it for inclusion in the next quarter. Or you may discover some are no longer relevant or necessary. You never have to carry over a zombie OKR. ([View Highlight](https://read.readwise.io/read/01m217gzv85mq6rq2sdzccj28n)) - Note: Keeping this for the vocabulary as much as the rule. "Zombie OKR" names a specific object — a goal surviving only because it survived — and naming it is most of what makes it killable. The permission framing matters too: teams usually already know the goal is dead and lack authorization to say so. Obvious analogue in issue trackers: the ticket carried across milestones by inertia. - Tags: atom/literature/h4, atom/fleeting/f1