Operating Cadence: Weekly Business Review and Quarterly OKRs

The WBR you inherited has 47 metrics on the dashboard, runs 90 minutes, and the only decision the room made last quarter was about the snack budget. The CEO has stopped asking sharp questions because the answers are buried in a slide pack nobody actually read. You've got one quarter to fix it before someone (probably the CFO) asks why we still hold this meeting at all.

Most operating cadences are theater. The slides are pretty, the room is full, the CEO nods, and nothing moves. The CoS owns the meeting but not the outcome, and quietly, everyone knows the cadence is a tax, not a tool. This guide is the rewrite. It's how to turn the WBR and the quarterly OKR cycle into a decision-making instrument that earns its calendar slot every week.

The test is simple. If the WBR were cancelled tomorrow and nobody noticed for two weeks, it shouldn't exist. Same for the QBR. Same for any review on the leadership calendar. A cadence that doesn't change behavior between week one and week two of the quarter is a status update with extra steps.

Strip the WBR to 3 leading + 1 lagging per BU

The first WBR I ran had 47 metrics. We cut it to 19, got 30 minutes back, and the discussion got noticeably sharper because the room could actually hold the numbers in their head. The math is brutal but it works: three leading indicators per function, one lagging metric per business unit. That's the ceiling, not the target.

Leading indicators are things a function controls this week. Sales: qualified pipeline created, demos held, proposals sent. CS: onboarding completion rate, NPS inflow, expansion calls booked. Product: tickets shipped against committed scope, P0 bug count, time-to-merge. The test for "is this a leading indicator?" is whether the function owner can change the number through their own actions in the next seven days. If they can't, it's a vanity metric or a lagging metric mislabeled.

Lagging indicators are the BU number the exec team is on the hook for: revenue, gross margin, NRR, CAC payback. One per BU. The lagging metric is the one the board cares about. The leading metrics are the ones you act on. If a metric is neither, kill it, or move it to a functional review where it belongs.

Here's the discipline: anything that doesn't fit the 3+1 rule gets a different home. Win rates by source belong in the sales team's own meeting. Sprint velocity belongs in product's. Customer health scores belong in CS standups. The WBR is for cross-functional pattern recognition, not for each function's internal hygiene. The moment the WBR becomes "every department reports their stuff," you've recreated the bloated meeting you cut.

One concrete example. A 60-person SaaS I worked with had a WBR with 31 metrics across five BUs. We cut to 17 (three leading per function, one lagging per BU). Within six weeks, the CEO started asking pre-meeting questions on Slack, because she could now actually see the connective tissue between marketing's MQL trend and sales's demo-held trend. With 31 metrics, she couldn't. The cut wasn't about saving time. It was about restoring signal.

The pre-read: number, why, what, ask

The single highest-leverage change you can make to a WBR is killing presentations. Nobody presents in the room. The room reads, in silence, for the first 10 minutes. Then questions only.

This works because the pre-read forces the metric owner to do the thinking before the meeting, not in the meeting. The format is four lines per metric, written by the owner, distributed 24 hours before the WBR:

• Number: the actual figure this week, with last week and trend in parentheses.
• Why it changed: the owner's hypothesis for the movement, in one sentence. Not a recap, a hypothesis.
• What we're doing: the action already underway, not what someone might do, what is being done.
• Ask: the specific decision or input the owner needs from the room. If there's no ask, write "no ask". That's a valid answer for a metric that's tracking on plan.

A real example, from a sales VP's pre-read line on a recent Tuesday:

Demos held: 41 (last week 47, 8-week avg 44). Why: SDR team had two reps out, mid-funnel automation tagged 12 leads as "low intent" that the AE team would have taken last quarter. What: Reverted the scoring change Tuesday morning, asked SDR manager to redistribute coverage for sick reps. Ask: Should we tighten the SLA on AE follow-up for borderline leads? Currently 4 hours, considering 1 hour for accounts > $50K ACV.

That's 80 words. The room reads it in 30 seconds. The discussion in the meeting starts at "should we tighten the SLA", not at "let me walk you through this slide." You've moved every WBR conversation forward by about 20 minutes of meeting time, every week.

The enforcement rule matters more than the format. If the doc isn't in by T-24, the metric isn't reviewed that week. The owner explains why next week, in writing, before the room. This is the part most CoS skip because they don't want to be the bad guy. Skip it once and the deadline is gone forever. The pre-read discipline is the cadence. The meeting is just the residue.

A pre-read template I've used across three companies:

Four cells per metric. Distributed Monday EOD for a Tuesday WBR. Read in silence. Decisions in the discussion.

Quarterly OKR design — three max, outcome over output

Most OKR cycles fail at the design stage, before a single number gets scored. The three failure patterns: too many objectives, output framed as outcome, and target thresholds that don't matter.

Three is the cap per BU. Not five with two "stretch," not three plus a bucket of "key initiatives." Three. If a BU leader says they have to do five, they don't have OKRs. They have a project list. The point of an OKR is forced prioritization. Anything beyond three loses the prioritization signal.

Outcome over output is the second test. "Ship onboarding redesign" is output. "Reduce time-to-first-value from 21 days to 7" is outcome. The first you can hit by shipping something nobody uses. The second you can only hit if the work actually moved the customer's clock. Outcome OKRs are harder to write because they force you to commit to whether the work mattered, not just whether it happened. That's the entire point.

The 0.7 scoring rubric, in plain language:

• 0.0–0.3: Missed badly. Either we mis-scoped or something broke. Needs a written post-mortem regardless.
• 0.4–0.6: Made progress, didn't hit the bar. Default zone for genuinely ambitious OKRs in a healthy company.
• 0.7: The target. We hit what we said we'd hit. Aiming for 0.7 means setting a goal where 0.7 is achievable but not inevitable.
• 0.8–0.9: Strong delivery. If most OKRs land here, the targets were too soft.
• 1.0: Sandbagged. We knew we'd hit it. Needs a post-mortem on planning, not execution.

Anything below 0.4 or above 0.9 needs a written post-mortem. Both signal a planning problem, not an execution problem, and the CoS owns surfacing that distinction. Most companies only review the misses. Reviewing the over-deliveries is where the planning rigor actually compounds.

A before-and-after rewrite from a real OKR cycle:

Before: "Improve onboarding experience" with KRs "Ship onboarding redesign," "Hit 80% NPS on onboarding," "Reduce support tickets by 30%."

After: "Reduce time-to-first-value from 21 days to 7 days for new mid-market customers." Single objective, single metric, one number that moves or doesn't. The redesign is implied work; the NPS is a downstream check; the support ticket reduction is a side effect. None of those needed to be the goal. The time-to-first-value number was.

The rewrite shrunk the OKR doc from three pages to half a page and made the scoring trivially clear at end of quarter. That's the smell test for a good OKR. If the scoring conversation takes more than 10 minutes, the OKR was poorly written.

MBR vs QBR — keep them separate

The monthly business review is a trend check. The quarterly business review is a strategy check. Don't merge them. The MBR exists to catch leading-indicator drift before it becomes a quarter miss. The QBR exists to ask whether we're still pointed at the right outcome at all.

MBR audience: operators. Function leads, BU heads, the CoS, the CEO. The agenda is "are leading indicators predicting the lagging metric correctly, and if not, what's broken in our model of how the business works?" One example: a marketing team's MQL number is on plan, but pipeline created is 30% below plan. The MBR is where you notice the gap and decide whether the conversion ratio assumption is wrong (recalibrate the plan) or the lead quality is wrong (fix the targeting). The MBR doesn't change strategy. It changes the model.

QBR audience: leadership team plus board-adjacent (sometimes the actual board). The agenda is "is the strategy still right?" Three questions get asked: are the OKRs we set still the right OKRs, did the quarter teach us anything that changes the next quarter's plan, and are there second-order effects we missed. The QBR is where strategy actually gets reset. The MBR doesn't have the audience or the standing to make those calls.

When companies merge MBR and QBR, two things happen. First, the operators stop showing up to the QBR because it's now too long and too high-altitude. Second, the strategic questions get rushed because the meeting got eaten by tactical reporting. Keep them separate, even if it costs an extra two hours a quarter.

"We hit the OKR but missed the quarter" — the diagnostic

This is the single most common pattern in the post-quarter wrap-up, and it has three distinct causes. The CoS's job is to diagnose which one, because each requires a different fix.

Cause A, wrong OKR: We measured the easy thing, not the right thing. The classic example is a sales team that scored 0.9 on "increase pipeline by 40%" but missed revenue by 20%. The pipeline grew because the team chased volume; the conversion rate cratered because the leads were unqualified. The OKR rewarded the wrong behavior. Fix: rewrite the OKR. The next quarter's OKR isn't "pipeline by 40%", it's "qualified pipeline by 25%, with conversion rate above 22%."

Cause B, right OKR, wrong threshold: The OKR pointed at the right outcome but the target wasn't ambitious enough to actually move the BU number. Sales hit 0.7 on "expand into mid-market," meaning we landed eight mid-market deals. Eight wasn't enough to move revenue. The OKR was right, the threshold was wrong. Fix: keep the OKR, raise the bar. Next quarter, target is 18 mid-market deals.

Cause C, right OKR, right threshold, wrong attribution: We hit the OKR, the threshold was correct, but something else broke that the OKR didn't cover. Sales hit mid-market expansion. Revenue still missed because gross margin compressed on the deals (the new mid-market segment has 40% lower margin than the SMB book). The OKR was fine; we missed because a non-OKR variable broke. Fix: the next quarter needs an additional OKR or a guardrail covering the variable that broke.

The diagnostic is the conversation. Walk the leadership team through which cause it was (A, B, or C) and force a pick. The most damaging move is to vaguely acknowledge "we missed the quarter" and write generic improvement notes. That guarantees the same failure pattern next quarter. Naming the cause forces the next OKR cycle to do something specifically different.

Decision logs — the cadence's actual deliverable

Every WBR and QBR ends with a written decision log. What was decided, who owns it, what the next checkpoint is. Stored in one place (Notion DB, Linear, the company wiki), doesn't matter as long as it's one place. The log is the cadence's deliverable. No log = the meeting didn't happen.

Five columns, no more:

That's it. One row per decision. Owner is one human, not a team. Next check is a specific calendar event, not "soon." Status updates between meetings, in the doc, not in the meeting.

The CoS writes the log live during the meeting and circulates within four hours. Same-day. If the log waits until tomorrow, half the room will have a different memory of what was decided, and you'll spend the next WBR re-litigating the same call. The four-hour window is the discipline that matters most.

A side benefit nobody talks about: the decision log becomes the artifact that earns the cadence its calendar slot. When the CFO asks "why are we still doing this WBR?", point to the log. Forty-six decisions made in Q1. Twelve of them changed the trajectory of a BU. That's the answer. If you can't produce a log with that density, the question is fair, and the cadence probably does need to be cancelled.

What the CoS owns vs what the CEO owns

The cleanest split: CoS owns the mechanics, CEO owns the calls. Mechanics means pre-read on time, agenda discipline, decision log written and circulated, dashboards updated, the right people in the room. Calls means the actual decisions: should we raise the SLA, should we pause the feature flag, should we rewrite the OKR.

If the CoS is making the calls, the cadence is broken. Either the CEO is checked out (and the org will start to feel that), or the CoS has overreached and the leadership team will quietly stop bringing real questions to the meeting because the answers feel like proxy decisions, not real ones. Both situations need a separate conversation, not a workaround.

The honest version: a junior CoS often gets pulled into making calls because nobody else wants to. Resist it. Surface the call back to the CEO with the recommendation framed clearly, and let the CEO make the call in the room. That's the job: to make the CEO's decisions higher-quality and faster, not to substitute for them.

Learn More

• A Day in the Life of a Chief of Staff
• Your First 30/60/90 Days as a New Chief of Staff
• Chief of Staff Job Description
• Running 1:1s Your Reports Look Forward To
• Setting Goals for a Reluctant Team