How I Decide Which of My 118 Accounts Needs Attention Today

Every morning I wake up to 118 Google Ads accounts. Some are fine. Some are slightly off pace. A few are in trouble. And I have no idea which is which until I look.

For a long time, my morning routine was opening the pacing spreadsheet, scanning every row, and building a mental list of what needed attention. It worked when I had 30 accounts. At 50 it got slow. At 80 it broke. I was spending 45 minutes just deciding what to work on before I’d done any actual work.

The Portfolio Health Prioritization skill exists because I needed a system that could look at 118 accounts and tell me — in under a minute — which 3-5 accounts need me today and which 100+ are fine without me.

The Core Problem: Equal Attention Is Wasted Attention

When you manage a portfolio, there’s a strong pull toward treating every account equally. Check each one. Review each dashboard. Scan each search terms report. It feels responsible.

It isn’t. It’s a misallocation of the scarcest resource you have — your focused attention.

On any given day, 70-85% of your accounts are on pace, performing within acceptable ranges, and don’t need you. They’re running. Let them run. The 15-30% that are off — some of those are slightly off and will self-correct. Some are trending in a direction that needs watching. And a very small number — usually 2-5 accounts — have a real problem that will get worse if you don’t act today.

The prioritization skill exists to separate those groups instantly, so your first hour of the day is spent on the accounts that actually need you, not on a scanning routine that treats a 3% pacing variance the same as a 15% pacing variance.

The Five Tiers

Tier 1: Critical — Investigate Immediately

Criteria: Pacing variance greater than ±15%, account completely off pace (2x expected or 50% of expected), major performance collapse (CPA doubled, ROAS halved).

These are the accounts where something is actively broken. A campaign paused by mistake. A conversion action that stopped firing. A budget that got changed to the wrong amount. Whatever it is, it’s bad enough that waiting until tomorrow makes it meaningfully worse.

On a typical day, I have 0-3 Tier 1 accounts. Some days, none. That’s a good day. When there are 3, that’s my entire morning.

Tier 2: High Priority — Investigate Today

Criteria: Pacing variance outside portfolio tolerance (the specific number depends on the client — some portfolios have tighter SLAs than others), zero spenders (campaigns with $0 spend month-to-date), or variance that’s gotten 5+ points worse in the last 7 days.

These accounts aren’t in crisis, but they’re outside the bounds of acceptable performance. A ±9% pacing variance isn’t going to blow up today, but if you ignore it for a week, it compounds into a month-end miss that requires a difficult client conversation.

I typically have 3-10 Tier 2 accounts on any given day. I investigate the top 3-5 after handling any Tier 1 issues.

Tier 3: Medium — Investigate This Week

Criteria: Pacing approaching the threshold, conversion tracking anomalies, multiple ad disapprovals, campaigns ending in 7-14 days.

These are the “keep an eye on it” accounts. Nothing is broken. Performance is acceptable. But there’s an indicator that suggests things could shift if left unchecked.

I review these in weekly batches, not daily. Trying to investigate 5-15 Tier 3 accounts every day is how you fall back into the “equal attention” trap.

Tier 4: Low — Monitor

Within tolerance but near the edge. Historical patterns that explain the variance (seasonal, post-budget-change ramp-up). Stable performance with minor fluctuations. These get a glance, not a deep dive.

Tier 5: No Action Needed

On pace. Performing well. No alerts. This is where 70-85% of the portfolio lives on a healthy day. The skill’s job is to confirm these accounts are fine so I can stop thinking about them and focus on the ones that aren’t.

The Four Focus Rules

Tiers tell you which accounts matter. Focus rules tell you which accounts matter most within a tier. When I have 8 Tier 2 accounts and time for 4 investigations, these rules decide the order.

Rule 1: Underspending Before Overspending

This is counterintuitive. Most people worry more about overspending — “we’re burning through the budget too fast.” But underspending issues typically have quick fixes: increase a budget, remove a bid cap, expand targeting. The investigation is short, the fix is fast, and the account gets back on pace quickly.

Overspending usually requires strategic changes — pausing campaigns, restructuring bid strategies, improving quality scores. These take longer to diagnose and longer to fix. Starting with underspenders means I can resolve 2-3 accounts in the time it takes to figure out one overspender.

The exception: overspending greater than ±15% is critical regardless, because it may exceed the client’s approved budget.

Rule 2: Stricter SLAs First

Not all portfolios have the same tolerance. Some clients report weekly and expect tight pacing. Others report monthly and are more flexible. A ±6% variance is a non-issue for a flexible client and a problem for a strict one.

The skill encodes these differences. When two accounts have the same variance percentage, the one with the stricter SLA gets investigated first.

Rule 3: Recent Changes Over Historical Patterns

An account that suddenly shifted from on-pace to ±10% in the last week is more urgent than an account that historically underspends by 10% every Q1. The sudden shift means something changed — a new competitor, a bid strategy update, a landing page issue. The historical pattern means the business is seasonal and the account will self-correct.

Both might show the same pacing number in a dashboard. The context makes one urgent and the other routine.

Rule 4: High Variance Over Near-Threshold

A ±12% variance needs attention before a ±6% variance, even if the ±6% account is from a stricter portfolio. The absolute magnitude of the problem matters — a ±12% variance is likely to get worse and is harder to correct at month-end.

What This Looks Like in Practice

Here’s a typical morning. My AI agent runs the portfolio health check and returns:

• Tier 1: 1 account — a property that dropped to $0 spend overnight
• Tier 2: 7 accounts — mix of underspending and one overspender
• Tier 3: 12 accounts — approaching thresholds, some ad disapprovals

Applying the focus rules:

1. First: The Tier 1 zero-spender. Something broke. Investigate, find the paused campaign, fix it. Done in 10 minutes.
2. Second: The 3 underspending Tier 2 accounts, starting with the highest variance. Quick fixes — budget increases, bid cap removals. Done in 20 minutes.
3. Third: The overspending Tier 2 account. Deeper investigation — check change history, review bid strategy, look at impression share. This one takes 30 minutes.
4. Remaining Tier 2 accounts: Listed in my briefing for tomorrow if they don’t self-correct.
5. Tier 3: Noted, reviewed in Friday’s weekly batch.

Total time deciding what to work on: about 1 minute (reading the prioritized output). Total time working: about an hour. Five accounts investigated, 4 fixed. The other 113 accounts didn’t need me today, and I didn’t waste time confirming that manually.

The Daily Briefing Integration

The prioritization skill is designed to feed into a morning briefing workflow. The pattern:

1. Overnight health check pulls pacing data for every account
2. Prioritization skill classifies all accounts into the 5 tiers
3. Investigation agents run in parallel on the top 3-5 accounts, diagnosing root causes
4. Briefing lands in my inbox with: critical issues (with root causes already identified), high-priority accounts (with investigation summaries), and everything else (counts only)

By the time I sit down with coffee, the system has already identified the problems, investigated the worst ones, and drafted recommendations. My job is to review the recommendations and decide which actions to approve. Not to find the problems. Not to figure out what’s wrong. Just to make the final call.

That’s what the prioritization skill makes possible. Not by being smart about any individual account — but by being systematic about which accounts deserve attention in the first place.

Get the Portfolio Health Prioritization Skill

mkdir -p .claude/skills/portfolio-health-prioritization
curl -o .claude/skills/portfolio-health-prioritization/SKILL.md \
  https://raw.githubusercontent.com/fourteenwm/ppc-ai-skills/main/portfolio-health-prioritization/SKILL.md

The full repo has 32 PPC AI skills I use in production — mutation safety, SQR classification, RSA refresh, and more. All at github.com/fourteenwm/ppc-ai-skills.

The Bigger Point: The Bottleneck Is Decision-Making, Not Execution

Most PPC automation focuses on doing things faster — generating ads faster, pulling reports faster, uploading negatives faster. And that matters. But the bigger bottleneck for portfolio managers isn’t execution speed. It’s decision speed.

When you have 100+ accounts, the question “what should I work on?” takes longer to answer than the actual work. The scanning, the mental sorting, the context-switching between “this is fine” and “this needs me” — that’s where the morning goes. Not on fixing problems, but on finding them.

The prioritization skill doesn’t fix accounts. It doesn’t generate reports. It doesn’t change bids. It does one thing: it tells you where to point your attention. And for portfolio managers, that’s worth more than any optimization tool, because attention pointed at the right account is the only thing that actually moves performance.

The accounts that are fine will stay fine without you. The accounts that are broken will get worse without you. The skill’s entire job is to make sure you can tell the difference in under a minute, so the rest of your day is spent on work that matters.