Ad account health signals to watch at scale

Quick Answer

Account health at scale isn't measured by ROAS — it's measured by leading indicators that predict ROAS movements 5-10 days ahead. The five signals to watch: average CPM trajectory, average frequency trend, creative win rate from your testing pipeline, audience saturation by ad set, and total active winning ads. When any of these degrades, ROAS will follow within a week. Watch them daily.

The Framework

1. Daily check: CPM 7-day moving average

Rising CPMs are the most reliable leading indicator of impending ROAS drops. A 10%+ increase over a week means audience saturation, creative fatigue or rising auction competition. Diagnose immediately, don't wait for ROAS damage.

2. Daily check: number of active winning ads

Define a 'winning ad' as ROAS above account average AND spend above $200/day. Count them daily. If the number drops below 4-5, you're approaching fragility — one fatigue and the whole account stalls.

3. Weekly check: creative test win rate

From your testing pipeline, what % of creatives passed graduation criteria last week? If it's below 10% for two consecutive weeks, your creative production is producing the wrong stuff. Investigate before scaling further.

4. Weekly check: frequency by ad set

Pull frequency by ad set every Monday. Anything above 4.0 on cold needs creative refresh this week. Anything above 5.5 on retargeting needs a budget cut this week.

5. Monthly check: audience overlap report

Run Meta's audience overlap tool monthly. If two ad sets have >25% overlap and both are scaling, they're cannibalising each other. Restructure.

6. Monthly check: incremental lift via hold-out

Once a month, run a small hold-out test (turn off one campaign in one geo) to validate incrementality. If the campaign isn't producing incremental revenue, restructure or kill.

Real Numbers from the Field

An accessories brand we monitored had ROAS quietly drift from 3.1 to 2.4 over 3 weeks. The CPM 7-day moving average had been rising 4-6% per week the whole time — visible weeks before ROAS dropped enough to alarm anyone. We caught it on the daily CPM check, refreshed creative aggressively and recovered ROAS to 2.9 within 10 days. Without the leading-indicator monitoring, the drop would have continued for another 2-3 weeks before the team intervened.

Frequently Asked Questions

How automated should the monitoring be?

Use a dashboard that pulls these metrics daily — Looker Studio, Triple Whale, Northbeam, or a custom Sheet via Meta's API. Manual pulls miss things.

What's the worst single signal?

Rising CPM. It's almost always a leading indicator of trouble and is the metric most often ignored until it's too late.

Can I just watch ROAS and skip the rest?

No. ROAS is a lagging indicator. By the time ROAS confirms a problem, you've already lost a week of budget. Watch leading indicators.

Who in the team should monitor these?

Ideally the lead media buyer or a dedicated analytics person. Monitoring without authority to act on the signals is pointless.

How do I set thresholds for alerts?

Start broad (alert on 15%+ weekly changes) and tighten as you understand your account's normal variance. Each account has its own normal range.

Stop Guessing What Your Ads Look Like at Scale

When you're spending £5k+ a day, you can't manually screenshot every ad on every device, in every placement, in every locale. Things break silently. Creators ship the wrong aspect ratio. Localised text overflows. Safe zones get clipped by Reels UI.

Pix-Vu lets you preview every ad creative across Facebook, Instagram and Reels placements before launch — desktop, mobile, all aspect ratios, all crop variants. Spot the broken creative before you spend £20k discovering it the hard way.

Try it free at https://pix-vu.com — built for media buyers who can't afford a single wasted impression.