Cost caps vs bid caps at scale

Quick Answer

Cost caps work when you have a hard CPA target you can't exceed and are willing to give up some volume to hit it. Bid caps work when you understand auction dynamics and want maximum control over per-impression cost. Lowest cost (no cap) works when you trust Meta to find your best ROAS and are willing to accept variance. At scale, most accounts use a mix: cost cap on profit-critical campaigns, lowest cost on testing.

The Framework

1. Lowest cost (no cap): use for cold prospecting and testing

When you don't yet know what a winning ad looks like, lowest cost gives Meta the freedom to find unexpected pockets. Use it on 60-70% of cold spend and on 100% of your testing campaign.

2. Cost cap: use when you have a hard CPA ceiling

If a customer above £40 CPA loses you money, set a cost cap at £40. Meta will scale up to that ceiling but stop before it. Be aware: cost cap kills volume — you'll spend less than your budget if Meta can't find profitable conversions.

3. Set cost cap 15-20% above your true target, not at it

If your true target is £35, set the cap at £42. This gives Meta room to scale during the day and only catches genuinely expensive conversions. A cap exactly at target is too tight and chokes spending.

4. Bid cap: use rarely, only when you understand auction mechanics

Bid cap is the most aggressive control — it sets a maximum auction bid. If you set it too low, you barely deliver. It's useful for high-CPM categories where you want control but it requires constant monitoring. Most accounts shouldn't use it.

5. Mix bid strategies across campaigns, not within

Use one strategy per campaign. Don't mix cost cap and lowest cost in the same campaign — they create conflicting signals for the algorithm.

6. Check cap delivery daily for the first week

When you switch to cost cap, monitor for under-delivery for 5-7 days. If you're missing budget by >25%, your cap is too tight. Loosen by 10-15% and re-check.

Real Numbers from the Field

A subscription brand we worked with had been using cost cap aggressively across all campaigns, set at exactly their target £28 CPA. They were under-delivering by 35% on most campaigns. We loosened cost cap to £33 (18% above target) on profit-critical campaigns and switched testing campaigns to lowest cost. Total spend grew 47% within 3 weeks while blended CPA stayed within target at £29.50. The over-tight caps had been the limiter, not the audience or creative.

Frequently Asked Questions

Should I use cost cap on Advantage+ Shopping?

Yes, when you have a strict CPA target. ASC + cost cap (set 15-20% above target) is one of the most reliable scaling combinations in 2026.

Why does cost cap reduce my spending?

Because Meta won't deliver impressions it can't fulfill within the cap. If your true cost-per-result drifts above the cap, Meta starves the campaign rather than overspend.

Is bid cap obsolete now?

Almost. There are niche scenarios (high-CPM auctions, retargeting tight audiences) but for most accounts at scale, cost cap or lowest cost are the right choices.

Can I A/B test cost cap vs lowest cost?

Yes — duplicate a winning campaign and run one with cost cap and one without. Compare ROAS and volume after 10-14 days.

What's the biggest cost cap mistake?

Setting it at your hoped-for CPA instead of slightly above your true profitable CPA. Meta needs headroom to scale, and a too-tight cap denies it that headroom.

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.