Most marketing does not fail because nobody knew what to do. It fails for want of a marketing operating rhythm. It fails because the work happened in bursts — three weeks of intensity, six weeks of nothing, then panic when the numbers slip.
What actually works is boring: a small amount, on a schedule. Here's the cadence we run, and how long each part honestly takes.
Daily — 15 minutes
Short, and mostly about catching problems while they're cheap.
- Check spend delivered. A campaign that stopped, got rejected or exhausted its budget by noon is the only genuine emergency in daily marketing.
- Answer every lead. Not a review task — an operational one. In competitive categories the first business to reply wins a disproportionate share.
- Scan for policy warnings. An ad disapproval caught today is an afternoon. Caught in a week, it's an account restriction.
What not to do daily: change budgets, pause ads after one bad day, or read ROAS. Daily numbers are noise, and reacting to them resets learning.
Weekly — about an hour
This is where decisions happen.
- Blended cost per customer — total marketing spend divided by new customers. The number that can't be gamed.
- Qualified lead rate. Rising volume with a falling qualified share means you're buying cheaper, worse people.
- Kill the losers. Anything past your kill threshold with nothing to show. Decide it by rule, not by mood — the framework is here.
- Check frequency. Above 2.5 on cold audiences means fatigue, not a bad concept.
- Read the search terms report if you run Google. An hour here regularly finds 20–30% of spend going somewhere useless.
- Ship one thing. A new creative angle, a page fix, a follow-up message. One shipped thing a week compounds faster than a quarterly overhaul.
A weekly review that doesn't end in a decision is a status meeting. The decision is the point.
Monthly — half a day
- Full numbers review: cost per customer against your ceiling, trend over three months, not one.
- Creative concentration. If one ad carries more than about 70% of spend, you're one fatigue curve from a bad month.
- Plan next month's tests from what you learned about the customer, not from a calendar.
- Publish content. Two to four genuinely useful pieces beats twelve thin ones.
- Update Google Business Profile — photos, a post, replies to reviews. Fifteen minutes, and it compounds.
Quarterly — a day
- Recalculate your CAC ceiling. Margins, pricing and repeat rate move.
- Technical SEO pass — indexing, speed, structure.
- Audit tracking end to end. Submit a test form, watch the event arrive. It breaks more often than anyone expects.
- Review the offer. The highest-leverage thing in the account, and the least often revisited.
If you post on social too
A realistic, sustainable cadence for a small team:
| Activity | Cadence | Time |
|---|---|---|
| LinkedIn / social post | 2–3 a week | 30 min each |
| Blog article | 2–4 a month | 2 hrs each |
| New ad creative | 8–12 a month | Batched, one day |
| Reply to comments and DMs | Daily | 10 min |
Three posts a week that continue for a year beats daily posting that stops in five weeks. Pick the cadence you can hold in your worst month, not your best one.
The honest failure mode of any marketing operating rhythm
Nobody abandons this schedule because it's hard. They abandon it because a busy fortnight eats it, and then restarting feels like admitting something.
So make the daily fifteen minutes non-negotiable and let the rest slip when it has to. Spend delivered, leads answered, policy warnings checked — those three, every day, prevent most of the expensive problems on their own.
A dashboard is not a report. This is what we send.
The version of this rhythm we run on client accounts is written up on our process page.
Questions people ask
What should I check daily?
Almost nothing, and it should take fifteen minutes: that spend is pacing, that nothing has been disapproved, and that new leads have been answered. Daily is for catching breakage, not for making decisions.
What belongs in the weekly review?
What changed and why, what the numbers did against the baseline you agreed, anything unexpected, and what is running next. An hour, same day each week, written down so the month can be reconstructed later.
Why is checking performance daily harmful?
Because daily variance looks like signal and invites changes that reset the learning phase. An account edited in response to noise never finishes learning, which makes the noise worse.
What is the most common failure?
Working in bursts. Three weeks of intensity, six weeks of nothing, then panic when the numbers slide. A modest cadence held every week beats an ambitious one abandoned in month two.
Want someone running this rhythm for you? Tell us what you're working with.