Purchase Volume as the Single Variable

The reduction

Eric Glyman, the co-founder and CEO of Ramp, describes reasoning through the company's business model in the very first board deck, before Ramp had any revenue. His starting point is that most consumer and small-business products are consumption-based: the more a customer uses the product, the more the company monetizes, and if the company does its job, the more value the customer receives. Ramp's revenue equation, in his account, has only a few variables, roughly purchase volume multiplied by a take rate, minus the cost of funding. On inspection, he says, "there's only one thing that mattered, which is purchase volume."¹

What makes the reduction load-bearing, in Glyman's telling, is that optimizing purchase volume involves no trade-off, because it improves every other term at once. More volume raises revenue, retains more interchange, lowers the bank's cost of funding, and generates more signal for fraud and loss prevention. So instead of juggling competing levers, the whole company can optimize a single output. The strategic question then collapses to one thing: how do you get people to want to spend with you.

Value in, volume out

Glyman's hypothesis is that volume is earned rather than bought. "If you just make spending with us simply better than spending anywhere else, rationally people should want to spend more with you." He describes the product compounding around that: more spend produces simple cash back and constant spend-cutting suggestions, which produces more expense reports and accounting to automate, which produces better benchmarks. The internal product metric that follows is money and time saved per customer, which he says drives purchase volume fairly directly and, over time, expands margin, a large reason he gives for Ramp's bottom line growing faster than its top line.¹

He draws an operating conclusion from this: because the business fundamentally gets much better at scale, getting to scale as quickly as possible becomes the priority. The single-variable model is, in that sense, the economic argument underneath Ramp's emphasis on speed.

Why not seats

Glyman says he considered seat-based pricing, the model Concur used, and rejected it as the spine of the business. His reasoning is that a consumption model aligns price with value delivered while seats do not, so there was no orthogonal pricing mechanism fighting the value proposition. This is the structural reason he frames Ramp as set up from the start to sell outcomes rather than a tool, the go-to-market expression of the same idea described in selling outcomes not tools. He adds that the consumption model ages better under AI: when fewer people are needed to do a given job, seat counts mis-slope against reality, whereas paying on volume and outcomes does not.

Glyman's instinct is to reduce the business to a single output the whole organization can optimize. He is also candid about the limits of the elegance. The one-variable model, he notes, is a founding simplification: the mature profit and loss statement shows card spend plateauing as a business grows more complex, which is why Ramp added software, bill-pay, and treasury lines that are not pure purchase-volume plays. There is also a genuine tension in optimizing a customer to spend less while monetizing purchase volume, which he resolves by pointing at share of wallet rather than per-customer spend growth, since he estimates 98 percent of a customer's spend is still not on Ramp.¹ The same logic of driving one channel to its natural ceiling appears in saturate the winning channel.