Pulsar · Metrics playbook · No. 10 · The monthly routine. Ties together No. 01–09 with the test pixel’s real numbers.
The 30-Day Ad Health Check
One monthly routine that catches every leak in this series — before it drains next month’s budget.
The whole series, one routine: No. 01 — Efficiency & Fatigue, No. 02 — Quantity vs. Quality, No. 03 — Garbage In, Lookalike Out, No. 04 — The Invisible Sales, No. 05 — Who Really Made the Sale?, No. 06 — Your Ad Is Dying, No. 07 — The Leak Map, No. 08 — Almost Bought and No. 09 — Are You Paying for Bots? — each one becomes a single monthly check here.
Contents
01 Routine > firefighting · 02 The nine checks · 03 The 30-minute version · 04 Where each check lives · 05 The compounding payoff · 06 The scorecard · 07 The whole series
Why a routine beats firefighting
Ad money leaks slowly and invisibly — a creative fatigues a few percent a week, a broken UTM quietly untags a few more buyers, the bot share creeps. By the time CPA obviously spikes, you’ve overspent for weeks. A fixed monthly pass catches the leaks early, in order of money impact, so you fix the biggest one first — instead of the one that happened to make the loudest noise this week.
You don’t need to watch everything every day. You need to check the right nine things, once a month, in the right order.
Every playbook in this series found one specific leak on the same real test pixel. This capstone turns those nine findings into one repeatable checklist — each check is one lecture, one number to read, one warning sign to react to. Section 02 is the full pass; section 03 is the 30-minute triage; section 06 is the printable scorecard you actually fill in every month.
The check, in order of money impact
Nine checks, each one a lecture in this series — what to look at, the warning sign, and this account’s real reading. Work top to bottom: the order is by how much money each leak touches. Tick each row as you review it.
The monthly pass — tick as you go
Each row shows this account’s real 30-day reading and an honest status chip. On this account, most read needs attention — that’s the point of the pass.
The checklist is interactive in the HTML version: each row ticks off and a progress meter fills to a “monthly pass complete” state at 9 of 9. On paper, the boxes on the left are for ticking by hand — the readings printed on each row are the real 30-day measurements from the test pixel: 25.3% bots (4,138 of 16,340), end-to-end 0.27% (3,675 visits → 10 buyers), view→cart −93.9%, T-shirt CPM 84→159 at frequency 1.05→1.66, 64% of buyers untagged, EMQ Checkout 8.0 vs Cart/Search 6.1 at 100% coverage, a 10-purchase seed, and 83 + 27 warm abandoners.
The 30-minute version — if that’s all you have
Busy month? These three checks catch the most wasted money the fastest. Everything else in the nine is refinement — important, but these three are the triage.
Is the data clean of bots?
4,138 of 16,340 events were machines. If bots aren’t excluded, every other number you read in the next 29 minutes is fiction. Check the share, confirm the filter is on, move on.
Where’s the single biggest funnel leak?
view→cart is the crack on this account. One step always dominates — find it, fix it, and ignore the temptation to tinker with the healthier steps first.
Is frequency creeping while CTR falls?
T-shirt frequency climbed while CPM ran 84→159 — the saturation signature. Catch it here and a refresh costs one creative; catch it in the CPA a month later and it costs the budget.
The 30-minute rule: clean data → biggest leak → fatigue. If any of the three trips its warning sign, spend the rest of the half hour on that one — not on all nine. The full pass can wait for a calmer week; the triage can’t.
Where each check lives — reading your own dashboard
The whole pass happens in six views you already have. One line each — no spreadsheet, no export, no SQL.
Check 9 doesn’t live anywhere — it’s the calendar entry. Same six views, same order, next month.
The compounding payoff
The nine checks aren’t nine separate chores — each fix makes the next one cheaper. Clean data gives you a better seed; a better seed gives the algorithm better targeting; better targeting lowers CPA; a lower CPA frees budget for what actually works. Checked monthly, the loop compounds in your favour instead of against you.
Fix the funnel before you feed it. Feed it clean, and it compounds — the right way.
The one-page scorecard — print it every month
This is the takeaway: one table, four columns, filled in once a month. The “this month” column below holds the current real readings — next month, yours replace them. Hit Print / Save as PDF and this section is a clean one-page cheat-sheet.
| Check | Healthy target | This month | Action |
|---|---|---|---|
| 1 · Bot share | Excluded by default; watch spikes | 25.3% · 4,138 of 16,340 | Keep the bot filter on; treat any sudden jump as a signal, not noise |
| 2 · End-to-end conversion | ≈1% visit→buyer | 0.27% · 3,675 → 10 | Stop buying more visits until the funnel below is fixed |
| 3 · Biggest funnel leak | No step far worse than its neighbours | view→cart −93.9% | Product page + offer work first — the top-ranked fix this month |
| 4 · Efficiency & fatigue | Stable CTR at steady frequency | CPM 84→159 · freq 1.05→1.66 | Refresh the T-shirt creative — saturation, not a broken product |
| 5 · Untagged buyers | Under ~20% of buyers | 64% untagged | Fix UTM tags and repair broken URL macros |
| 6 · Match quality (EMQ) | Money events at 8+ | Checkout 8.0 · Cart/Search 6.1 · coverage 100% | Push advanced matching on Cart and Search up toward Checkout’s 8.0 |
| 7 · Lookalike seed | 50+ clean buyers | 10 clean purchases | Seed on buyers only — never the 1,529 viewers; let the list grow |
| 8 · Warm recovery | Refill audiences on + syncing | 83 cart + 27 checkout waiting | Leave the abandoner audiences armed — cheapest sales in the account |
| 9 · Re-map & repeat | Full pass monthly | The leak moves as you fix | Book 30 minutes next month; redo the pass in this order |
Filled in: ____ / ____ / ______ · account: ______________________ · biggest leak this month: ______________________
The series, in one place
Ten playbooks, one test pixel, one routine. Each check in section 02 unpacks into a full lecture — this is the index.
Efficiency & Fatigue
Read CPM, CTR and frequency together — the three-line chart that separates an expensive audience from a tired ad.
Quantity vs. Quality
Count buyers, not visits — the end-to-end rate (here, 3,675 visits → 10 buyers = 0.27%) is the only conversion number that pays rent.
Garbage In, Lookalike Out
A lookalike is a lesson you teach the algorithm — seed it on your 10 best buyers, never on 1,529 window-shoppers.
The Invisible Sales
Match quality decides how many of your sales the platform can even see — push money events toward EMQ 8+.
Who Really Made the Sale?
Attribution is only as honest as your tags — here 64% of buyers arrived untagged, so every channel report was guessing.
Your Ad Is Dying
Every creative fatigues on a schedule — catch the frequency-up / CTR-down crossover and refresh before the CPA spike.
The Leak Map
Rank the funnel’s step-to-step drops and fix the biggest crack first — here, view→cart at −93.9%.
Almost Bought
Abandoners are your warmest audience — 83 cart + 27 checkout people already said “almost yes”. Win them back first.
Are You Paying for Bots?
25.3% of events weren’t people. Exclude the machines before reading — or optimizing on — anything at all.
The 30-Day Ad Health Check
This deck — the capstone. All nine leaks, one monthly pass, in order of money impact.
Meta Says 6, You Sold 1
Meta’s purchase count is not your real sales. Appendix — Meta-attributed vs pixel-recorded, and which one to trust for money decisions.
The Learning Phase
A new ad’s first-week numbers are Meta guessing. Appendix — why a small budget gets stuck in Learning Limited, and how to get out.
Too Many Ad Sets
One real campaign done right — “Bible Cover” concentrates 225/day into a single ad set and clears learning (~92 Add-to-Carts a week). Appendix — why budget is a per-campaign fact, how splitting a small budget too many ways starves learning, and how to consolidate.
Custom vs Lookalike
Re-reach the warm, or clone them to find the cold. Appendix — custom audiences vs lookalikes, why the seed is everything, and the exclusion you’re missing.
Pulsar vs a Tag Manager
A tag manager fires the pixel; Pulsar makes the spend work. Appendix — why a pixel + GTM isn’t the whole job: identity, clean CAPI, match quality, audiences.
One honest line to close: Pulsar surfaces all nine checks in one place, on real people, with the data already cleaned — so the monthly pass takes minutes, not a spreadsheet.