FreeYour first ad is on us, built from your own ad data.Your first ad, on us.Book your free ad →

Every ad makes the next one better.

Built from your data. Trained on every result. Winning performance ads, every cycle, from 100 to 1,000+ a month.

MetaMeta Partner AppLovinPartner ShopifyPartner
How it works

Your data in.
Winning ads out.

Four steps, on repeat. Connect once and every round after that starts from what already worked in your account.

01

Connect your accounts

Joops plugs into your ad accounts, store and attribution, then listens everywhere your product is discussed.

Meta Ads
Google Ads
Ads
Triple Whale
02

Joops finds the next winner

It reads your ads, your competitors, seasonality and customer objections, then turns the evidence into concepts and hooks.

HookSpendCTRCPM
States Problem$7.9K1.45%$64.78
Curiosity$3.3K1.15%$62.42
Social Proof$393.56%$51.70
Bold Claim$31.44%$23.29
"96% of your magnesium gets pissed out."Your winning ad
"Magnesium oxide did nothing for my sleep."Reddit thread
"The drugstore bottle never mentions absorption."Competitor ads
03

Your ads get made, you approve

Joops builds each hook into an ad, then our human editors finish and claim-check it. Nothing ships until you approve.

Hook A
Hook B · Winner
Hook C
04

Every batch beats the last

Joops tracks the metrics that move your account, scales the winners, retires the rest, and builds the next batch from what won.

Creative metricBatch 1Batch 4
Hook rate3-second views22%22%+86%
Hold rate15-second views9%9%+100%
CTRlink clicks1.1%1.1%+118%
CPAnew customer$52$52−35%
ROAS7-day1.3×1.3×+62%

Illustrative values

Case studies

What this looks like in a live account.

The platform

Your view of Joops.

Managed-service brands see three screens. Everything else runs underneath.

01Review

Approve in one tap.

Every ad lands on your board with its batch and platform. Watch it, send it back with a note, or clear it to launch. Nothing goes live without your tap.

  • Approve, or request a revision in one line
  • Full history of every change on every ad
  • Download any finished ad, any time
02Top Ads

Know what is winning. Every morning.

Joops reads your account overnight and ranks every creative against your own average. The winners are obvious before the week starts.

  • AI summary: what leads, what to watch, where budget sits idle
  • Hook, CTR, CVR and CPA on every card
  • Tagged by what the ad does, not just how it looks
03Competitors

Every competitor ad. One library.

What the brands you compete with are running on Meta and Google, by platform, described frame by frame so Joops can learn from them and so can you.

  • Live counts per brand, updated as they launch
  • Filter by platform, saved or expired
  • Search hooks, copy and formats across all of them

The problem

It's a supply problem.

When performance stalls, most teams blame the hook. But the algorithm burns through creative, and the more you spend, the more it needs. The brands that scale aren't making better ads. They're testing more of them.

Creative volume scales with spend

Average monthly video uploads per advertiser, by daily spend tier

  • $1M+ / day5,307
  • $500K–$1M1,082
  • $250K–$500K365
  • $100K–$250K237
  • $25K–$100K204
  • $5K–$25K177
  • Under $5K107

Source: Meta & AppLovin 2026 advertiser data reports

Introduced

Joops isn't a tool.
It's a loop.

Your data goes in. Ads come out. Results come back. Every cycle, Joops knows more about who buys your product and why — and the next batch is built from that.

Most creative starts over every month. This compounds.

Month one we're guessing less than your last agency. Month six we're not guessing at all.
Healora

Zero to $7,053 a day in three weeks — while acquisition cost fell.

Daily spend growth
9.9×$716 → $7,053 per day
New-customer CPA
−46%$45.44 → $24.35
New customers
2,441$123,005 new-customer revenue
New-customer ROAS
1.99×2.19× total attributed

The problem

Healora had no media history on the channel. No account, no baseline, no winning creative to copy. Every dollar and every customer here was built from a standing start — day one spent $716.

What we did

Twenty new ads a week, launched in batches with fresh end cards. Budget stepped on Aug 13 and again on Aug 17, each time after new-customer efficiency confirmed headroom. Supply never became the reason to stop scaling.

WeekSpendNC CPA
Week 1Aug 6–12$5,284$45.44
Week 2Aug 13–19$20,332$24.40
Week 3Aug 20–26$36,305$24.35

Week 3 ran at 6.9× the Week 1 budget with no efficiency give-back — and 91% of attributed revenue came from new customers.

Aug 6–26, 2026. Spend from AppLovin advertiser reporting; new-customer figures from Triple Whale paid-linear attribution. AppLovin’s own platform reporting reads $23.83 new-customer CPA and 1.81× new-customer ROAS over the same window — the two land within $1.54 of each other on new-customer cost. Totals: 1,415,880 impressions, 102,482 clicks, 7.24% CTR, $0.60 CPC. Week 1 carries the learning-period cost and is not representative of steady state.

SoFyre

$78K across a 90-day run — and 98% of it bought a first-time customer.

Spend, 90 days
$78,07230 May → 27 Aug 2026
Conversions
1,68128-day matured
New customers
98%share of checkouts at day 0
ROAS
143%$46.44 cost per purchase

The account

A pest-control DTC brand running repellent pouches across several pest categories and seasons. The constraint is seasonality: an angle that carries one month is out of season the next, so the account needs a standing supply of new concepts rather than one hero ad held for as long as possible.

What we did

A continuous mix of 3D animation, creator-led UGC and straight product demo, plus playable end cards. Ninety days in, almost every purchase is still a first-time customer — the account is buying new demand, not re-buying its own list.

ReadROASCPP
7-dayat 27 Aug140.6%$47.23
28-day maturedsame window143.2%$46.44
Current 7-dayto 5 Sep155.5%$42.33

New-customer ROAS at day 7 was 136.0% on a $48.80 new-customer cost. The account is reading better now than across the 90-day average.

30 May – 27 Aug, 2026. Spend and conversions from AppLovin advertiser reporting, cohorted attribution. Unlike Healora, this account has no Triple Whale connection, so these are platform-reported only and not cross-checked against a second source. Cohorted numbers mature: the 7-day read of 1,653 conversions settles to 1,681 at 28 days. The seven days to 27 Aug softened to 116.8% ROAS and $56.22 CPP, driven mostly by a single weak day.

My Derma Dream

$1,510 to $30,236 a day in thirty-six days — and CPA fell 39% on the way up.

Daily spend growth
20×$1,510 → $30,236 per day
Cost per acquisition
−39%base period → ramp
Orders
1,209$219,866 spend, 36 days
Blended ROAS
1.19×first-order revenue only

The account

A skincare device brand already spending on the channel. The question was not whether it could work but whether it could hold together at twenty times the budget — most accounts lose efficiency once spend more than doubles.

What happened

The account held a $1.5k–$4k base through Aug 25, then stepped hard once the funnel proved it could absorb volume. Efficiency improved through the ramp rather than degrading: cost per acquisition fell 39% and return on ad spend rose 49% against the base period.

PeriodSpendDays
BaseAug 1–25$68,92925
RampAug 26–Sep 5$150,93711

The ramp spent 2.2× the base period in under half the days, and did it at a materially better cost per order.

Aug 1 – Sep 5, 2026, 36 days, AppLovin Axon web. Orders, CPA and ROAS are measured in the client’s own Checkout Champ data across Axon funnels, not in AppLovin platform reporting. These two sources are known to disagree on this account and we have not reconciled them — the figures here are the client’s own and are not cross-checked against a second independent source. ROAS counts first-order revenue only and excludes any repeat purchase.

SP Nutrition

Spend quadrupled across eleven weeks and ROAS held within 6% of where it started.

Daily spend growth
4.0×$1,258 → $5,036 per day
New-customer CPA
$52.24blended, full window
New customers
3,609D0 new-customer checkouts
D7 ROAS
1.80×cohorted, still maturing

The account

A supplements brand with an established base on the channel. The test was whether budget could go up fourfold without the efficiency read falling apart underneath it.

What we did

Budget moved in deliberate steps, each one taken only after the prior level held its efficiency read — never a single jump. Total spend across the window was $188,541.

ReadD7 ROASNote
First weekof the window2.01×base
Last weekof the window1.88×at 4× spend
Full windowblended1.80×cohorted

New-customer volume per week tripled across the window. Cost per new customer did rise as spend scaled — the trade was volume, not free efficiency.

Jun 15 – Aug 30, 2026, 77 days, AppLovin Axon web. AppLovin advertiser reporting, cohorted attribution. This account has no second attribution source connected, so these figures are platform-reported only and are not cross-checked. Cohorted D7 ROAS is still maturing and will move as later cohorts settle.

Super Greens Hemp Co.

Zero to $4,318 a day in nine weeks — cost per order moved 1%.

Daily spend growth
10.6×$409 → $4,318 per day
Cost per order
$51.60blended, full window
Orders
2,426D7 cohorted
D7 ROAS
2.00×cohorted

The problem

No history on the channel at all. Jun 27 is the account’s first day of spend — no baseline, no account structure, no proven creative to lean on. Everything here was built from a standing start.

What we did

Budget moved through five deliberate steps, each held long enough to read before the next increase. Total spend across the window was $125,184.

PhaseDaily spendCost per order
Phase 2early ramp1.0×$53.24
Phase 5steepest ramp6.6×$53.82

Daily spend rose 6.6× between those phases while cost per order moved from $53.24 to $53.82 — a change of 1%. Acquisition efficiency did not degrade under scale.

Jun 27 – Aug 27, 2026, 62 days from launch, AppLovin Axon web. AppLovin advertiser reporting, D7 cohorted attribution. Zero pre-GetConversions baseline on this channel. No second attribution source is connected to this account, so these are platform-reported figures only.

Tillman’s Tranquils

Budget doubled and held — then efficiency came back past where it started.

Daily spend
2.0×$1,007 → $2,058 per day
New-customer CPA
−29%trough → current
New customers
1,088D0 new-customer checkouts
D7 ROAS
1.30×against a 1.14× baseline

The account

Unlike the others here, this is not a scale story. Budget doubled once in late July and has held at roughly $2,000 a day since. What changed is what that budget returns.

What happened

The account dipped when spend doubled and bottomed out in early August — the dip is real and worth showing, because the recovery only means something against it. It then recovered past its original efficiency while running at twice the budget.

StageD7 ROASNC CPA
Baselinebefore the step1.14×—
Troughearly August—peak cost
CurrentAug 19–Sep 51.30×−29%

Same account, twice the daily budget, and a better return than it produced at half the spend.

Jun 27 – Sep 5, 2026, 71 days, AppLovin Axon web. Total spend $120,170. AppLovin advertiser reporting. The current D7 ROAS reads the Aug 19 – Sep 5 window; the 1.14× baseline is the pre-step period. New-customer CPA improvement is measured from the early-August trough to current, not from the pre-step baseline. No second attribution source is connected to this account.

Platforms

AI ads built for each platform's rules.

Every platform rewards something different. Each ad is built to that platform's best practices: the aspect ratio, the length, how fast the hook lands and the metric it's judged on.

See what your data
already knows.

Send us your last 30 days. We'll come back with the concepts worth building next — before you commit to anything.

Free. Yours to keep either way.