What is Cold Audience?
Also known as: Prospecting audience, Top-of-funnel audience
What is a cold audience?
A cold audience is a group of users who have no prior relationship with the brand. They have never visited the site, clicked an ad, watched a video, or signed up for anything. They sit at the top of the funnel.
Per HubSpot's inbound funnel research, only about 3 percent of any market is actively buying at any moment. The other 97 percent is cold to some degree. Cold-audience advertising is how brands reach that 97 percent before competitors do.
Cold traffic shares three traits across every platform:
- No first-party signal exists for the user.
- Click-through rates run lower than warm traffic.
- CPMs run cheaper because competition for attention is broader.
The job of a cold audience campaign is rarely a same-session purchase. It is to create the warm pool that the next 30 days of retargeting will live on.
Cold vs warm vs hot audience
The funnel splits cleanly into three temperature bands. Each band needs different creative, different bids, and different success metrics.
| Stage | Definition | Typical CTR | Typical conversion rate | Primary metric |
|---|---|---|---|---|
| Cold | No prior brand interaction | 0.5-1.5% | 0.5-2% | CPM, reach, video view-through |
| Warm | One or more touches (site visit, video view, follow) | 1.5-3% | 3-7% | CPC, add-to-cart rate |
| Hot | High-intent action (cart abandon, demo request, repeat visit) | 3-8% | 5-20% | CPA, ROAS |
Warm audiences and hot audiences live inside custom audience lists built from pixel events, CRM uploads, and engagement triggers. Cold audiences are everyone else the platform can show an ad to.
How to target cold audiences
[UNIQUE INSIGHT] The biggest cold-audience mistake in 2026 is over-targeting. Stacking three or four interest layers on a cold ad set sounds precise. It usually shrinks reach, raises CPMs, and starves the algorithm. Broad plus strong creative beats narrow plus average creative on most accounts we run.
Meta: broad with Advantage+
Per Meta's prospecting documentation, Advantage+ Audience now treats manual targeting as a suggestion rather than a hard filter. The recommended cold setup on Meta in 2026: country plus age 18-65, no interests, Advantage+ on, creative variants stacked.
Google: in-market and custom segments
Per Google's audience targeting reference, in-market audiences capture users actively researching a product category. Custom segments let advertisers seed an audience with keywords, URLs, or apps. Both work as cold-prospecting layers on Search, Display, Demand Gen, and Performance Max.
TikTok: interest plus content behavior
TikTok's cold-audience graph runs on what users watch, complete, and share. The platform's Smart Performance Campaign mirrors Meta's Advantage+. Broad reach plus creative volume wins. Narrow interest stacks rarely outperform on prospecting.
Lookalikes from a thin seed
Lookalike audiences built from a 1,000-user purchase seed sit between cold and warm. They are technically cold (the modeled users have no brand touchpoint) but the seed makes them more efficient than pure broad on most accounts.
What ad creative works on cold audiences
[PERSONAL EXPERIENCE] Cold creative has one job: stop the scroll. The viewer has no brand context, no patience, and no reason to stay. Three patterns work consistently across our customer accounts:
- Hook in the first 1.5 seconds. Pattern interrupt, bold claim, or a face-on-camera question.
- Problem before product. Show the pain the buyer recognizes before the solution shows up.
- UGC over polish. User-style content outperforms brand films at cold by 30 to 60 percent on CTR in our tests.
Static image ads still work cold, especially on Meta feed placements. Video wins on TikTok and Reels because the format demands motion. Carousel ads win for product-led offers with multiple SKUs.
Cold audience CPMs vs warm CPMs
[ORIGINAL DATA] Across Coinis customer accounts in 2024-2025, cold Meta CPMs averaged $14 to $22 in the US. Warm retargeting CPMs averaged $28 to $45 on the same accounts. Cold runs cheaper per impression because the auction has more eligible inventory. Warm runs richer per click because the user already knows the brand.
The math that matters is not CPM. It is cost per acquisition after the full funnel runs. Cold spend that costs $4 per click and converts at 1 percent produces a $400 CPA. Warm spend at $2 per click converting at 8 percent produces a $25 CPA. The cold spend is what makes the warm spend possible.
Real-world example with numbers
A direct-to-consumer coffee brand splits a $300/day Meta budget across two ad sets for 30 days. Same creative, same offer, same landing page.
Ad Set A: Cold prospecting. Broad US, 25-55, Advantage+ on. Audience size: 180M. Result: 12,400 link clicks at $0.65 average CPC. CPM: $16. Purchases: 142. CPA: $56. ROAS: 1.4.
Ad Set B: Warm retargeting. Last 30-day site visitors plus video viewers. Audience size: 84K. Result: 2,100 link clicks at $1.10 average CPC. CPM: $38. Purchases: 168. CPA: $13.80. ROAS: 5.7.
The warm set looks like the winner on its own. It is not. Without 30 days of cold prospecting feeding the retargeting pool, the warm set has nobody to advertise to. Cold spend created the 84K-user warm pool. Warm spend converted it.
Cold audience targeting in 2026
Three forces define cold-audience advertising in 2026.
First, AI delivery beats manual stacks. Meta's Advantage+ Audience, Google's Performance Max, and TikTok's Smart Performance Campaign push budget toward broad cold reach. Manual interest stacking still wins in regulated and B2B verticals.
Second, creative volume is the new targeting precision. When platforms cannot slice users finely, they slice ads instead. The right creative reaches the right user because the algorithm pairs them. Conversion rate on cold lifts most when ad volume goes from 3 variants to 30.
Third, cold-to-warm velocity matters more than cold CPA in isolation. A campaign that builds a 100K warm pool in 30 days at a 1.2x ROAS often outperforms a campaign that hits 2.5x ROAS but builds only 10K warm users. The warm pool is the asset.
The Coinis platform addresses creative volume directly. Paste a product link, generate dozens of ad creative variants for cold audiences, ship them to Meta, TikTok, and Google. Targeting becomes a question of "did we go broad enough and feed retargeting fast enough" rather than "did we pick the right interest stack."
Related terms
Frequently asked questions
What is the difference between a cold and a warm audience?
Cold audiences have zero prior touchpoints with the brand. Warm audiences have at least one: a site visit, a video view, a follow, a form fill. Cold audiences cost less per thousand impressions but convert at 0.5 to 2 percent. Warm audiences cost more but convert at 5 to 15 percent on retargeting campaigns.
How big should a cold audience be?
On Meta, between 1 and 50 million users gives the algorithm enough room to optimize. On Google, broad match plus audience signals on Performance Max effectively makes the audience the entire in-market pool. Per Meta's targeting guidance, narrow cold audiences under 500K starve delivery and inflate CPMs by 30 to 50 percent.
Do cold audiences ever convert directly?
Yes, but the rate is low. Direct-response cold campaigns convert at roughly 0.5 to 2 percent on first impression, depending on price point and category. The job of cold audience ads is mostly to seed the retargeting pool. Per HubSpot's funnel research, only 3 percent of any market is actively buying at any given time.
What budget split should go to cold versus warm?
A common starting split is 70 percent cold, 20 percent warm retargeting, 10 percent loyalty. Without cold prospecting, the warm pool dries up in 30 to 60 days. Without retargeting, cold spend leaks because the average buyer needs 6 to 8 touches before purchase.
Are cold audiences still effective in 2026?
Yes, more than ever. Privacy changes since iOS 14 shrank retargeting pools, so the cold-to-warm pipeline has to feed itself constantly. Broad cold audiences plus AI-driven delivery (Advantage+, Performance Max) now outperform narrow interest stacks in roughly 70 percent of accounts in our data.