Growth financing: the numbers have to be lendable first

RESEARCH LEAD | TRIPLE WHALE | JUNE TO JULY 2026

Triple Whale was considering lending to the brands on its platform, and the team wanted to know which of three structures to build. A financing product is expensive to half-build. It needs a source of funds, legal clearance, underwriting, and a way to collect, and none of that can be prototyped. So the phase answered a question one level up. Whether the need was real, whose it was, and whether brands would take money from the company that reads their numbers.

The director of fintech brought the three structures he was developing and sat in on the interviews. I designed and ran the program around what he had to resolve to make a build call. Phase one closed with a lead direction, a named target customer, and one condition ahead of the build. Reported numbers have to reconcile to what a brand treats as its source of truth and cover every channel they sell through, and today they often do neither. A discrepancy is a dashboard annoyance until the same number sizes a loan.

Approach

Three phases, each ending in a decision

Phase one proves demand and names the target. Phase two designs the offer and clears what a financing product depends on. Phase three runs a small pilot, funded and underwritten by hand, before anyone decides whether to automate it. Every phase ends in a build, refine, or hold call, so the company spends in the order the evidence arrives. Phase one ran about three weeks.

Seven decision questions as the spine

Everything the phase had to answer was written as seven questions before any instrument existed. Each interview question and survey item was tagged to the question it fed, and anything that fed nothing came out.

WHAT THEY DO TODAY

Interviews mapped what they borrow now, from whom, and at what cost

Recruiting the edges as well as the middle

Ten brands, not agencies, from roughly $1M to $44M, across supplements, footwear, pet, fitness equipment, lingerie, and feminine health. The sample deliberately included brands with cheap capital already, including one funded by its parent at a rate we could not beat. Their reasons for not needing us drew the outer edge of the target.

Qual and quant in the same window

A survey went to the customer list while interviews were still running, testing the same structures. The guide opened on behavior, how they fund the business and what they pay for it now, and held the concepts until the back half so reactions had something to be weighed against. Interviews carried a moderator who could explain a concept. The survey carried none, and the two came back with opposite readings on cohort financing.

A hub instead of a readout

Findings went into a live hub. Every session got a same-day landing, one takeaway tied to a decision question plus what it changed. Stakeholders could see the phase plan, the open questions, the interview record, and the survey results in one place. By the phase-one review nobody was seeing the evidence for the first time, so the meeting was about the decision.

What we found

Three of the six findings arrive at the same asset from different directions. A trusted read on performance is what converts the brands sitting on cash without conviction, what a loan gets sized from, and what wins the operator the internal yes.

DEMAND

Operator-led physical-product brands with seasonal or volatile demand hit a hard cap. Cash goes into stock, they sell out, and the reorder takes three months. One footwear brand had been roughly flat for two years on that pattern alone. This is the one constraint in the data where a brand can point at revenue it lost, and several already borrow to fix it, so the demand is proven and the spend is going to other lenders today.

Ad spend is mostly not the gap. Brands running profitable ads fund them from returns or float them on a card, and one general manager's ads pay back in about 72 hours, often before the platform bills him.

A second group has the cash and lacks the conviction

SEGMENT

Six of eleven survey respondents named confidence in the return as what holds them back, the most common answer. One growth director's whole constraint was proving the money would be well spent, since his parent company lends him all he needs. Another had raised ad budget by half and watched nothing move.

What converts this group is proof, and capital without it converts no one. That layer is worth building whether or not lending ships, and it underwrites every shape that follows.

Every concept read as something they had already been pitched

POSITIONING

Eight of eleven rated a Triple Whale option about the same as how they fund growth today. In interviews the concepts mapped straight onto Shopify Capital, Wayflyer, PayPal, or a business card with a new name. Naming and positioning cannot create preference here.

Preference comes from beating incumbents where brands say they fail. Cost was the top walk-away at eight of eleven. Underwriting friction came next, and one founder described paying her accountant to assemble statements because a lender needs an accounting sync she cannot keep working. Repayment fit was third, and the open text on it was blunt about daily remittance dragging cash on strong days.

One concept sold itself live and stalled cold

METHOD

Cohort financing, where the brand repays out of what a funded group of customers spends, drew the strongest reactions of anything in the set when a moderator explained it. Brands valued the shared downside and the chance to fund acquisition against payback they can already predict. It was also the only concept nobody mapped to an existing lender.

The survey ran the other way. Cohort scored 3.3 of 7 on clarity against 5.5 to 5.7 for every other concept, and usefulness followed clarity down. One respondent asked whether it meant Triple Whale would run campaigns and take a cut.

Two reads fit that gap. Either comprehension is the blocker, or live explanation inflated the response in a way the real offer moment will never reproduce. The offer moment looks like the survey, with no moderator in it. Both reads stayed open in the recommendation.

Lending will not cost the relationship. Lending on today's numbers would

REFRAME

The team's working worry was that becoming a lender would damage the analytics relationship. It did not show up. Nobody said financing would lower their trust in the product, and two said it would raise it. The condition brands set was on the numbers, and it has two independent failure modes.

One CFO had stopped using Triple Whale for her core number because weekly sales read £20K to £30K above Shopify, which reconciles to her bank. That is accuracy, on a business where nearly all revenue is visible to us. Separately, Triple Whale sees roughly a tenth of one retail-heavy brand's revenue, so anything sized off what we can see would misjudge the business. Two brands, two different ways for the same number to fail, and a lending product has to clear both.

The operator who wants the capital rarely signs for it

BUYER

At larger brands the handoff is clean. Marketing proposes, a finance director or board approves at annual budget. At one $5M brand the path ran five steps from sales forecast to CEO to board, about six weeks end to end, and she expected it to double as they move toward institutional capital.

Operators asked, unprompted, for something to take into that conversation. A recommendation finance can read, a forecast, a scenario case. Two said they would use a Triple Whale version of the model they build by hand today.

From finding to recommendation

The recommendation named a lead, split the one ad-spend direction into four products, held one bet, and put the data work ahead of the build.

• Lead with working capital anchored on inventory. The broadest and most durable shape, well understood in the survey and mapped to the sharpest pain. It looks like everyone else's product, so it wins on the four things brands named. Speed from data we already hold, repayment that flexes without daily remittance, transparent cost, and help making the internal case.

• Treat ad spend as four products. The card is the first move, since nearly every brand already funds ads on a card or platform terms, and one founder named the exact switching floor that becomes the first pricing stimulus in phase two. Channel-testing credits are the second move, for brands sitting on a channel they will not fund out of inventory cash. Financing to scale proven winners waits on the reconciliation work, though the analysis underneath it is worth building now. The advance against ad history folds into the card as a backstop for brands pushed off cards onto wire.

• Hold cohort financing as the differentiated bet. Scope it to repeat-revenue brands and settle it with a one-line, self-serve explanation test. Interview-level pull means comprehension was the blocker. A flat response means the interview enthusiasm was the artifact, and the concept gets rescoped or cut.

• Start reconciliation and the integrations brands named. Every shape sits on it, and the accounts that surfaced the gaps are where it starts.

The limits went in the same document. At n=11 the survey gives directional counts rather than rates. The interview sample skews toward inventory-heavy brands, which is part of why inventory reads as the dominant need. And concept reactions proved sensitive to how the concept was presented, which is a finding and a caution for how anything gets tested next.

Where it landed

Design is underway on the offer experience. The capital source, legal clearance, and underwriting run alongside it.

The pilot follows once designs and approvals are in. Seven of eleven survey respondents volunteered as early testers and one interview participant offered to pilot directly, so recruitment is in hand.