← Field notes

Signal series · Part 3 of 5 · July 24, 2026 · 5 min read

Ramp: when appetite outruns income.

This is the signal that separates a synthetic from your best growing member, and it is the subtlest. Both borrow more over time. Only one of them borrows more than its real income can explain.

Part 3 of a five-part series on the signals a retrospective review reads. Earlier: Coherence. Next: Inflection.

Growth is not suspicious. A real member who is doing well borrows more as they earn more. They ask for a limit increase, they carry a larger balance, they take a car loan, and every bit of it is backed by rising income you can see arriving. If you flagged accounts simply for wanting more credit over time, you would flag your healthiest relationships. So the signal cannot be "credit appetite is rising." It has to be "credit appetite is rising faster than real income," and staying that way.

The two lines to watch

Picture two lines drawn across an account's life. The first is credit exposure sought: limits requested, balances carried, new tradelines opened, the total amount of borrowing power the account is reaching for. The second is genuine inflow: the real money actually coming in, payroll and the other funding that provenance taught us to look for.

For a real growing member, those two lines rise together, roughly. The appetite tracks the income, because the income is what services the appetite. For a cultivated synthetic, the exposure line pulls away from the income line and keeps climbing. The account wants more borrowing power than any real money coming in could ever support. That widening gap is the ramp.

The move that hides the gap

There is a complication, and it is the reason this signal has to be computed carefully rather than eyeballed. Inflow can be faked, at least on the surface, by cycling money. Funds get pushed into the account and pulled straight back out to the same place, or moved in a loop between related accounts, so that the raw deposit total looks healthy. On a naive read, the income line looks high enough to justify the borrowing.

So the honest version of the signal removes that first. You net out money that arrives and departs to the same counterparty in a short window, transfers that round-trip rather than settle, deposits that never actually fund anything. What is left is genuine inflow: money that came in and stayed to be spent on a life. When you measure appetite against that number instead of the gross one, the ramp that cycling was hiding tends to reappear.

Why this reasoning, not a threshold

Notice that none of this is a fixed cutoff. There is no universal ratio of credit-to-income that means "synthetic," because the right ratio depends entirely on who the account is. A high earner can carry a lot of exposure against a lot of income and be completely ordinary. The signal is not the level; it is the divergence, read against the account's own income and its own past. Appetite that outruns real, non-cycling inflow, and keeps outrunning it, is the shape we mean. A single month above trend is noise. A persistent, widening gap after you strip out the round-tripping is the signal.

Where it sits in the lifecycle

Ramp is the middle of the story. Provenance was the origin; coherence was the texture of the spending; ramp is the account straining against the limits of what its real economics can support. It is louder than the first two signals and earlier than the last one. By the time the ramp is unmistakable, the account has usually not yet done the dramatic thing, the sudden run to the limit, that everyone eventually notices. That dramatic thing is the next note in the series: the inflection, when the account finally breaks from its own pattern all at once.

Sources & notes

This is a conceptual, operator-facing description of one behavioral signal. It makes no factual claims about any specific institution, portfolio, or case, cites none, and describes no Delegate engagement or result. The mechanics it names (credit exposure sought, net versus gross inflow, transfer round-tripping) are ordinary properties of account data. You can see the signal families run on a simulated book in the Portfolio Explorer.