The audience is trained to distrust
Every trading audience has been targeted by signal-selling scams for a decade. Enthusiasm reads as a red flag. The more excited your creative sounds, the faster they scroll.
Work Case 01 Fintech
$0 to $61,400 MRR
in 31 days.
PolyPick launched with a Domain Rating of 1 out of 100. No backlinks, no search traffic, no ad account history, no brand. In the single most sceptical vertical on the internet. Every dollar of revenue it has ever made came from creators.
Case 01 · Fintech · 31-day campaign
14Creators cast
1,240Videos shipped
6Hook families
11 / 3US / international
01 · Context
A product with no history, in a category built on distrust.
PolyPick is an AI tool for prediction markets. You screenshot a market on Polymarket, Kalshi or PredictIt, and it reads the news and data behind that question and returns a side, a target and its reasoning in under ten seconds.
Good product. Impossible launch conditions.
It went live in May 2026 with zero domain authority, zero brand recognition and no advertising history, into a category where the audience has been burned so many times that scepticism is the default posture. Prediction-market traders have seen a thousand “guaranteed picks” accounts. They assume you are one.
And the obvious channels were closed. SEO takes 8–14 months at DR 1. Paid social ad review is hostile to anything touching betting or financial outcomes. Which left exactly one option: other people’s credibility.
Where 921,600 sessions actually came from
Traffic mix across the 31-day primary window
Creator referral, link-in-bio & dark social
Direct + branded search
Non-branded organic search
Branded search is itself a creator artefact. People watched a video, then typed the name into Google. Treated conservatively as its own bucket rather than claimed as creator-attributed.
02 · The problem
Finance is the hardest vertical in creator marketing. Three reasons.
Platforms police financial claims
Any implied return triggers review or takedown on all three platforms. A hook that works for a skincare brand will get a finance video suppressed within hours.
Generic creators can’t carry it
A lifestyle creator holding a trading app has zero credibility with traders. The audience overlap is near nil and the tone is instantly wrong. Casting is the whole game here.
03 · The hypothesis
The belief we had to break.
Every HIVE campaign starts by naming the one belief standing between the viewer and the purchase. For PolyPick it was blunt:
“Anyone selling me predictions is running a scam. If it worked, they wouldn’t be telling me. They’d just be using it.” The objection, as it actually appears in trading-community comments
You cannot out-argue that belief. Claiming harder makes it worse, every additional promise is another data point that you’re the thing they’re afraid of.
So the hypothesis inverted the normal creative instinct:
Campaign hypothesis
In a distrustful category, showing losses builds more belief than showing wins. If a creator opens by showing the tool getting one wrong (and keeps filming anyway) the audience reclassifies them from “salesperson” to “someone actually using this.” Every claim made after that point inherits the credibility of the admission.
This also solved the compliance problem in one move. A video that documents a process (including its failures) makes no financial promise, so there is nothing for a platform to action. The thing that made the creative trustworthy was the same thing that made it publishable.
The belief-shift architecture
What each stage of the video is engineered to change
0.0 – 0.8s · Interrupt
“I lost $340 testing this.”
Opens on a loss. Breaks the sales pattern instantly. Nobody selling you something leads with a loss. Scroll-stop achieved through incongruity, not hype.
0.8 – 8s · Reframe
Screen on, no claims
Raw app usage. The viewer watches the mechanism instead of hearing about it. Belief moves from “is this real” to “how does it work”.
8 – 25s · Evidence
The seven-day log
Wins and losses both shown, with the net. Specific to the dollar. Precision is the trust signal, round numbers read as invented.
25 – 35s · Transfer
“Try it for a dollar.”
The $1 first month removes the last risk objection. Almost no ask, the offer does the closing, the creative did the believing.
Videos that opened with a loss held 2.9× the retention at 15 seconds compared with videos that opened with a win.
04 · Campaign design
Six hook families. 1,240 attempts to find out which one was right.
We never brief a single creative direction. We brief competing hypotheses and let the market vote, then move budget to the winner within 72 hours.
Hook family performance
Indexed to campaign average = 100. Conversion index weights signup rate, not views.
View index
Conversion index
View as table
| Hook family | Views idx | Conv. idx | Videos |
|---|---|---|---|
| Loss transparency | 236 | 410 | 402 |
| 7-day live test | 177 | 222 | 288 |
| Whale-tracking reveal | 216 | 103 | 214 |
| Beginner’s first week | 110 | 132 | 196 |
| Contrarian market take | 82 | 45 | 98 |
| Feature walkthrough | 39 | 29 | 42 |
Note the gap between the two bars on “whale-tracking reveal”: high views, mediocre conversion. Curiosity hooks travel but don’t sell. This is exactly why we never optimise a campaign on views, and why a views-only agency report should worry you.
The 72-hour kill decision
On day three, “feature walkthrough” and “contrarian take” were cut entirely: 140 videos of planned production reallocated. By day nine, 61% of all remaining production was loss-transparency variants. The campaign that ended is not the campaign that started, and that is the point.
05 · Creator sourcing
14 creators. All of them already traded.
We didn’t hire UGC actors. We hired people whose audiences already came to them for market takes, sports bettors, options traders, prediction-market regulars, finance-adjacent commentators.
Small accounts, deliberately. The best performer in the entire campaign had 8,400 followers and produced $9,180 of MRR on their own.
Creator mix
A 600,000-follower lifestyle creator holding a trading app converts worse than an 8,000-follower trader who has been posting losing screenshots for two years. Reach is not the asset. Standing is. HiveMind casting principle, Ingest stage
06 · Production & distribution
1,240 videos in 31 days.
Staggered on purpose. A wave that lands all at once reads as a paid campaign. A wave that builds over weeks reads as a trend.
Weekly publishing volume vs. MRR
Two separate scales are deliberately avoided, both series are indexed to their own peak = 100
Videos published (label = actual count)
MRR, indexed to peak
Volume peaked in week two; revenue peaked in week five. That roughly three-week lag between publishing and revenue is the most commonly misread number in creator campaigns. Teams that judge results at day 14 kill programmes right before they work.
The creative that carried it
Four representative posts from the winning hook families.
Post URLs are configured in js/config.ab5269f6.js. Tiles without a URL are non-clickable placeholders.
07 · Results
What 38.4 million views turned into.
Full-funnel conversion
31-day primary window · every stage measured, not modelled
View as table
| Funnel stage | Volume | Step rate |
|---|---|---|
| Organic views delivered | 38,400,000 | – |
| Sessions on site | 921,600 | 2.40% |
| Accounts created | 94,000 | 10.20% |
| Converted to paid | 1,574 | 1.67% |
| Peak MRR | $61,400 | – |
The $1-first-month offer is doing real work at the account-to-paid step. Removing price as an objection at the exact moment trust peaks is why the funnel holds together, creative earns the click, the offer closes it.
Before · Day 0
May 2026
- Domain Rating 1 / 100
- 0 monthly recurring revenue
- 0 registered users
- 0 creator relationships
- No advertising history on any platform
- No brand search volume
After · Day 31
Peak of the campaign
- Domain Rating 1 / 100, unchanged, and irrelevant
- $61,400 monthly recurring revenue
- 94,000 registered accounts
- 14 creators producing continuously
- 1,240 owned video assets, in perpetuity
- 4.9 / 5 average product rating
Third-party verification
PolyPick’s revenue is independently listed on TrustMRR, which verifies figures directly against the payment processor rather than taking a founder’s word for it. Current figures on that page reflect the present-day run-rate after the founder shifted focus to a separate B2B venture and listed the asset for sale, not the campaign peak reported here. We’ve kept the two clearly separated rather than blending them into one flattering number.
08 · Unit economics
The only slide a board actually reads.
Campaign economics, 31-day window
All spend figures are creator payouts plus production, no product or salary costs included
| Total creative spend | $23,424 |
| Videos produced | 1,240 |
| Effective cost per video | $18.89 |
| Views delivered | 38,400,000 |
| Cost per 1,000 views (CPM) | $0.61 |
| Paying customers acquired | 1,574 |
| Blended CAC | $14.88 |
| Subscription price | $39 / mo |
| CAC payback period | 12 days |
| Peak MRR | $61,400 |
| Annualised run-rate added | $736,800 |
| Return on creative spend | 31.5× |
A 12-day CAC payback means the channel self-funds inside a single billing cycle. That is the number that turns a marketing line item into a growth engine, and the reason this campaign could scale without a funding conversation.
09 · What it proves
Three transferable conclusions.
1. Domain authority is not a prerequisite for revenue.
PolyPick’s Domain Rating never moved off 1. It generated $736,800 of annualised run-rate anyway. If your growth plan is gated behind an SEO timeline, you are choosing to wait eight months for something creators can deliver in five weeks.
2. In sceptical categories, admitting weakness outperforms claiming strength.
The winning hook family opened with a loss and converted at 4.1× campaign average. Every instinct in a marketing team says to lead with the win. In finance, health, and anywhere else the audience has been burned, that instinct is actively costing you money.
3. Views and conversions are different businesses.
Whale-tracking hooks produced 216 on the view index and 103 on conversion. Loss-transparency hooks produced 236 and 410. If we had optimised on views (as most creator reporting does) we would have doubled down on the wrong thing and halved the revenue.
Where this transfers
Any product where trust is the primary barrier rather than awareness: fintech, insurance, health, legal, B2B tooling with a switching cost, and anything sold to an audience that has been burned by your category before.
Next case study
Everyone said UGC doesn’t work for B2B.
Writecream: $1.93M net-new ARR, at a CPM 42× cheaper than LinkedIn.
Free distribution audit
What would this look like for you?
Tell us your site and your bottleneck. We’ll come back with a written teardown, creator mix, realistic volume, projected CPM, and an honest view on whether this channel is right for your product at all.