Bidding and Budget Pacing Without the Fire Drill

Two months ago I watched a paid manager finish March $14,000 under budget. The CMO didn't say anything in the Monday review, which is somehow worse than getting yelled at. Then April happened. The same manager spent $22,000 over plan in the last six days because someone in finance asked, "are we actually using the channel?" That's not a budget problem. That's a system problem. And it gets fixed on day 5, not day 25.

I've run paid for accounts ranging from $8K/month to $400K/month, and the pattern is the same. The managers who never end the month in panic mode aren't more disciplined. They aren't smarter. They just have a dashboard that tells them what to do on a Monday morning and a bidding choice that matches the conversion volume they actually have. That's it. The rest of this article is just spelling those two things out.

Bidding Strategies — When Each One Actually Works

Here's the part most blogs get wrong: they tell you Manual CPC is dead. It isn't. It's just been mis-sold to people running mature accounts who should have moved on years ago. Each strategy has a window where it's the right answer, and outside that window it'll quietly cost you money.

Manual CPC. New accounts. Brand campaigns where you need a hard ceiling. Anything under 15 conversions a month where the algorithm has nothing to learn from. I still run Manual CPC on three brand-search campaigns right now because I refuse to pay $14 for a click on my own brand name when $4 holds position 1. Smart bidding doesn't care about that ceiling. You do.

Maximize Conversions (no target). Ramp phase only. You're feeding the algorithm and watching CPA spike before it stabilizes. This is the strategy I use for the first 10 to 14 days after launching a new campaign that's expected to do volume. Watch CPA daily here, but don't touch anything. The instinct to "fix" a $180 CPA on day 4 when your target is $90 is what kills the learning. It usually settles by day 11 if your creative and targeting are within reason.

Target CPA. Use this when you've crossed 30 conversions per 30 days and your CPA history is stable, meaning the standard deviation across the last four weeks is under 25% of the mean. One conversion action, not a Frankenstein "leads + demos + downloads" stack. And here's the part nobody tells you: set your target 10 to 15% above your actual rolling average, not below it. Setting tCPA below your average doesn't make CPA come down. It throttles delivery and makes your spend collapse, which is the exact opposite of what most managers think will happen.

Target ROAS. E-comm or any campaign where conversions carry real revenue values from a clean source. If you're a B2B lead-gen account and you've been told to "use tROAS with a fake $50 lead value," don't. The algorithm bids for that fake outcome. I switched a $40K Meta account to tROAS once with assumed lead values and CPA doubled for three weeks before I undid it. The bidder was optimizing for a number I made up, and it was very, very good at it.

The 30-Conversion-Per-Month Threshold (And Why You Keep Breaking It)

Google publishes the 30 conversions per 30 days threshold for smart bidding. It's not arbitrary. It's the volume the algorithm needs to model auction patterns reliably. The same logic applies to Meta CBO and LinkedIn auto-bid even though those platforms don't publish the number. I'd argue LinkedIn needs more like 40 because the auction is thinner.

Below 30 conversions a month, smart bidding does the wrong thing. It over-corrects on small data, flattens delivery, and produces CPAs that swing 2-3x week to week. Managers see the swing, panic, and switch strategies again, which resets the learning to zero. I've audited accounts where the manager switched bid strategies four times in eight weeks. They had no signal at all. They were just paying Google to learn the same thing four times in a row.

Two rules I won't break:

Don't switch bid strategies if you're under 30 conv/month. Stay on Manual CPC or Max Conversions until you're comfortably above the threshold for two consecutive months.
After any bid strategy change, give it 14 days minimum before judging performance. The first 7 are trash. Day 8-14 is when you see the real CPA.

Pausing campaigns during the learning phase also resets you. So does any change above 20% to budget or target. If you have to make a big budget change, do it in 15% steps spaced 4 days apart.

The Pacing Math (One Formula, That's It)

Here's the only formula you need:

Daily target = (Monthly budget − spend-to-date) / days remaining × platform delivery quirk

That's it. Print it. Tape it next to your monitor.

Platform quirks matter more than people admit:

Google. Front-loads delivery, especially on search. Multiply by 0.95 in week 1, 1.0 after that.
Meta. Steady delivery within reason. Multiply by 1.0. The exception is iOS conversion windows where you'll see lumpy reporting on weekends. That's a reporting issue, not a delivery issue.
LinkedIn. Under-delivers Friday through Sunday. Multiply by 1.10 Monday-Thursday, 0.85 Friday-Sunday. Saturday on LinkedIn is functionally a half-day.

Worked example. $50K monthly budget split across Google ($25K), Meta ($15K), LinkedIn ($10K). Today is day 10. Spend to date: Google $9K, Meta $5K, LinkedIn $3.5K. Total $17.5K, behind plan by about $750 (plan would be $50K × 10/30 = $16,667... actually slightly ahead, but the channels are uneven).

Days remaining: 20. So:

Google: ($25K − $9K) / 20 × 0.95 = $760/day. Wait, that doesn't pass the smell test for a $25K monthly Google budget. Let me redo it: ($25K − $9K) / 20 = $800. Multiply by 0.95 because we're past week 1 (actually 1.0 now since we're in week 2). So $800/day for the next 20 days. Done.
Meta: ($15K − $5K) / 20 × 1.0 = $500/day.
LinkedIn: ($10K − $3.5K) / 20 = $325 base. Of the next 20 days, about 14 are Mon-Thu and 6 are Fri-Sun. Weighted target: ~$355 Mon-Thu, ~$275 Fri-Sun.

The point isn't the exact arithmetic. The point is that "spend $1,667 a day" is wrong, and most managers run on that average without thinking about platform delivery shape.

The Day-5 vs Day-25 Dashboard

If you only check pacing on day 25, you're not pacing. You're explaining. By day 25 the only levers left are bad ones: jacking up daily caps, panic-launching new ads, dumping spend into one ad set. The whole game is catching the trend on day 5, when you still have 25 days to course-correct gently.

The dashboard has six columns. Every paid manager I've worked with has built some version of this. The columns are non-negotiable:

Platform	MTD Spend	MTD Target	% to Plan	Projected EOM	Action
Google Search	$9,000	$8,333	108%	$27,000	Hold
Google PMax	$4,200	$5,000	84%	$12,600	Raise daily cap 10%
Meta	$5,000	$5,000	100%	$15,000	Hold
LinkedIn	$3,500	$3,333	105%	$10,500	Hold

Anything over ±10% of plan goes red. Anything in red gets an action assigned the same Monday it shows up. The "Projected EOM" column is just (MTD spend / days elapsed × days in month). It's the most useful number on the sheet and the one most managers don't compute.

The IC version pulls from each platform's API into a Google Sheet, refreshes Monday and Thursday mornings, and takes 30 seconds to read. The lazy version is 15 minutes of manual entry twice a week. Both work. The version that doesn't work is "I'll check pacing in the platform UI when I get a chance." You won't.

End-of-Month Surge Fixes (When You're $15K Behind on Day 25)

Sometimes you inherit the dashboard problem instead of preventing it. Day 25, $15K behind, the CFO email arrives. Here's the order of operations that actually works.

Don't just raise daily caps. That's the lazy fix and it dumps your money into low-quality auctions. The platform will spend it. The conversions won't be there.

What to do instead:

Raise tCPA by 15-20%. This is the cleanest unlock. You're telling the algorithm it's allowed to bid into more expensive auctions. Conversions usually arrive within 48 hours. Don't go above 20% in one move or you reset the learning.
Expand audiences. LinkedIn: open job function targeting from "VP, Director, Manager" to include "Senior Manager, Head of, Lead." Meta: shift lookalike from 5% to 10%, or layer interest stacks. Google: add a broad-match counterpart to your phrase-match campaigns with a 30% lower bid.
Turn on Performance Max if you've been holding it back. PMax will spend whatever you give it. It's the easiest channel to absorb $5-10K of unspent budget on short notice. Just make sure your asset groups are set up before you flip the switch.
Increase frequency cap on retargeting. From 3/week to 5/week. Diminishing returns kick in fast but you'll get a few extra conversions.

What NOT to do, ranked from worst to merely bad:

Pausing campaigns "to save budget for the strong ones." This always backfires. You lose learning, the strong campaigns can't absorb the freed budget, and you end the month worse off.
Switching bid strategies in the last week of the month. The 14-day learning period eats your remaining runway.
Dumping $5K into a single ad set overnight. CPA will spike 3-4x because you've broken the audience saturation curve. The algorithm will normalize, but not by month-end.
Launching new campaigns on day 26. They won't exit learning before the month closes. You'll look busy and spend nothing.

The 90-Minute Monday Cadence

Pacing is a weekly job, not a daily one. Daily fiddling kills the algorithm. Set a 90-minute block every Monday morning and don't touch the accounts between Tuesday and Friday unless something breaks.

0-15 min, pacing dashboard check. Open the sheet. Read the % to plan column. Anything red gets noted.
15-45 min, fix red flags. Apply the actions from the action column. Daily cap adjustments, tCPA tweaks, audience expansions. Make changes inside the 20% movement guardrail.
45-75 min, bid adjustments based on weekend data. LinkedIn especially. Friday/Saturday data shows you what's happening when the bid pressure is lower. Look at which placements drove the cheapest leads over the weekend and shift Mon-Thu bids accordingly.
75-90 min, queue creative tests for the week. Two or three new variants per channel, scheduled to launch Tuesday. Brief the designer if needed.

That's the whole week. You'll be tempted to "just check on it" at 4pm on Wednesday. Don't. The platforms reward consistency, and your CPAs will be more stable if you let the bidder do its job.

Pacing Isn't About Discipline

I used to think the managers who paced well were just more rigorous. After running this in a dozen accounts, I don't think that anymore. It's the dashboard. Build the dashboard once. Refresh it twice a week. Match the bid strategy to your conversion volume. Stop switching strategies because a single week looked bad.

The fire drills disappear. Not because you got tougher. Because the system started doing the work for you.

Paid Ads Manager Playbooks