A Distribution Plan for a Product the Platforms Restrict: One Fintech Company, Ninety Days.

Abstract frameworks are easy to nod at. This is what the method looks like on one company.

The company is a composite: a consumer fintech app offering a yield-bearing account through a partner bank, six months after launch, with a product that converts well from word of mouth and a distribution problem. Two ad accounts rejected. A third under review for eight weeks. A marketing budget sitting unspent because there is nowhere approved to spend it.

Weeks one and two: the map and the sounding

The first job is to write down where the company cannot be seen. Meta's financial-products policy requires documentation the company can supply, but the review has no defined timeline. TikTok's restricted-industry rules place the product in a category requiring pre-approval in its main geography. Google will carry it with certification. Programmatic exchanges will carry it against the company's own audience definitions. Affiliate networks carry the category. Creators can carry it with disclosure and compliant claims.

So: two channels in the restricted band with paperwork filed and nothing forecast, one closed for now, and four open.

The second job is a one-week measurement across three of the open channels, each with a written question.

Affiliate: Can comparison and personal-finance publishers produce funded accounts below one hundred dollars each on a bounty model, with the company's own tracking confirming the funding event?

Creators: Can mid-tier personal-finance creators, on a cost-per-funded-account basis and with pre-approved claims, produce at least forty funded accounts in a week?

Programmatic: Against an audience the company defines, with a real-CPM standard and a rejection rule, can display and short video produce app installs that fund at a rate comparable to organic?

Fourteen thousand dollars across the three. Tracking in the company's own analytics. A decision written in advance for each.

Week two's result

Affiliate read. Eleven publishers recruited by hand, sixty-three funded accounts confirmed, a cost per funded account of seventy-one dollars, no coupon partners admitted. Creators read partially: strong cost per account from three of the eight, nothing from the rest. Programmatic did not read: installs were cheap, funding was poor, and the rejection rate on the initial buy was near zero, which explained the cheap installs.

The decision, per the written thresholds: build the affiliate program now, expand the three creators into a structured program, and hold programmatic until a proper desk with inventory standards can be stood up.

Weeks three through eight: build the thing that read

The affiliate program gets terms: a bounty on funded accounts above a minimum balance, a thirty-day cookie, a validation period long enough to catch reversals, no brand bidding, no scraped codes, coupon partners tiered at a lower rate and capped. Placed on one network that carries the category, with a direct platform for the hand-recruited publishers. Forty partners recruited in six weeks, each with creative, an FAQ, and a compliance-approved claims sheet.

The creator program gets a brief, disclosure language, per-creator tracking, and cost-per-funded-account terms. The three that converted recruit the next tier.

Owned channels get the neglected fix: every partner and creator link lands on a page that asks for an email, and the list gets a fortnightly message compliance has reviewed.

Weeks nine through thirteen: add the desk

With affiliate and creators producing, programmatic is revisited under proper terms: flat fee per account, an inclusion list, a verification standard, a rejection rule that discards inventory failing it, and a real-CPM report. The buy is optimized to funded accounts, not installs. It reads, at a higher cost per outcome than affiliate, and is kept at a size that does not crowd the channel that works.

The Meta approval arrives in week eleven. It is added as a fourth channel, capped, and its outcomes compared with the others on incrementality rather than on its own attribution.

What the numbers had to show

At ninety days: affiliate at roughly half of funded accounts, creators a quarter, programmatic and Meta the rest, no channel above half, no partner above a third of affiliate outcomes, a real cost per funded account the finance team can reconcile, an email list that did not exist in week one, and a written map of which channels are closed and why.

None of it required getting past a review. All of it required knowing where the answer was already yes.

lowob takeaway: A restricted-category distribution plan is a map, a week of cheap measurement, one channel built properly, and the rest added in proportion. The approvals are a bonus, never the foundation.