The Data Model: The Bowtie

A board member once asked me a question that should have taken thirty seconds to answer: of the new logos we closed eighteen months ago, how many had expanded? I could not answer it from the CRM. The opportunity object stopped at Closed Won. Everything after the signature lived in a separate success tool with its own stages, its own record IDs, and no reliable key back to the deal that started it. Renewals sat in a spreadsheet a CSM updated by hand. We had three systems describing one customer and no single model that connected them, so the honest answer was that I would get back to the board next week.

That gap is not a tooling problem. It is a data-model problem, and it is the one this chapter fixes. In the last chapter, the Revenue Model, we chose a monetization strategy and saw how it sets price, contract term, and who carries the delivery risk. Those choices produce data: an ACV, a term length, a renewal date, a promise of impact. The question now is where that data lives and whether the structure holding it describes the whole customer or only the sale. Revenue Architecture’s answer is the Data Model, and its shape is a bowtie.

The funnel only measures half the business

The funnel we all inherited traces back to AIDA (Awareness, Interest, Desire, Action), which Elias St. Elmo Lewis sketched in 1898. AIDA ends at the purchase, because in 1898 the purchase was the end. The classic marketing-and-sales funnel keeps that shape: it narrows from a wide band of prospects down to a thin band of closed deals, and then it stops. Revenue Architecture lists five things wrong with that funnel, and the one that matters most for a data model is structural. Two of the three growth engines, retention and expansion, live entirely past the point where the funnel ends. We established the reason in the Recurring-Revenue Factory: roughly 83% of a five-year contract’s value arrives after the first year, and only if the customer keeps getting impact. A data model built on the funnel is blind to the majority of lifetime revenue by construction, not by accident. It was designed to describe a sale, and you are running a lifecycle.

The Bowtie fixes this the simplest way possible. Keep the funnel as the left half. Add a mirror image on the right for everything that happens after the customer commits. Draw the two halves meeting at a central knot, and the shape you get is a bowtie.

Eight sub-systems, one knot

Walk it left to right. The left side is the acquisition funnel you already know, with one addition. Awareness turns prospects into leads. Education turns leads into opportunities. Then comes a stage the old funnel never had a name for: Prioritization. Prioritization is a mutual-commitment checkpoint inserted before the most expensive resource in the entire model, a human being, is deployed. Before an account executive spends a discovery hour, both sides confirm there is enough pain, enough priority, and enough fit to justify the cost of that hour. In a high-touch motion that checkpoint is a qualification call. In a product-led motion it takes seconds and happens inside the product itself. Same stage in the model, wildly different cost to run it. After Prioritization comes Selection, the stage most people just call the sales process, ending in a win.

At the center sits the knot: Mutual Commit. This is where a deal count becomes a dollar amount. Everything to the left of the knot is measured in units (leads, opportunities, wins), and everything to the right is measured in revenue. The transition is a single multiplication: VM6 = VM5 × ACV, the number of wins times the average contract value the Revenue Model set in the last chapter. That equation is the seam between the two chapters. Whatever monetization strategy you chose flows into the model right here.

The right side mirrors the left. Onboarding converts the commitment into a live, using customer. Retention is the adoption engine seen from the seller’s side: the stage where recurring impact either lands or leaks. Expansion grows the account. The whole thing accumulates into Lifetime Value, the long tail of revenue that made the factory worth building. Notice that Revenue Architecture relabels the knot “Mutual Commit” instead of “Closed Won” on purpose. The word you use for the center of the model quietly decides whether your team treats the signature as a finish line or a starting line, and for a recurring-revenue business it is a starting line.

Two domains, inverse math

The two halves of the Bowtie are not just two ends of one line. They are inverse domains that obey opposite rules, which is why an operator has to think about them differently. Revenue Architecture draws the contrast sharply, and it is worth keeping the table in front of you whenever you make a decision that touches one side or the other.

We will do the actual algebra in the Mathematical Model, but the intuition matters now because it changes how you read the shape. The left side is an open loop, which is why acquisition is volatile: a small decline in each stage conversion rate compounds into a large drop in wins, and nothing in the system pulls it back. The right side is a closed loop, which is why retention and expansion are the durable, compounding engines: a satisfied customer feeds the next renewal and the next expansion. That asymmetry is the reason the same headcount cut lands so differently depending on which side it hits. Firing an AE removes some units from an open loop that will keep running. Firing a CSM removes a feedback mechanism from the closed loop that compounds all your future revenue.

Where the money actually is: the four growth areas

The right side is not one motion. It is four distinct plays, and confusing them is one of the most common ways expansion revenue gets left on the table. Revenue Architecture organizes them on a two-by-two: is the impact new or already realized, and is the person receiving it an existing beneficiary or a new one? A beneficiary here is the person or team that actually benefits from the impact, not the person who signs the check.

Renew and Resell are Retention: same impact, kept alive with either the same person or a new one (the resell exists because champions change jobs, and the model treats winning over the replacement as its own play). Upsell and Cross-sell are Expansion: new impact, sold to either an existing beneficiary or a brand-new one. Cross-sell, the bottom-right cell, is the hardest of the four because you are landing new impact with a new beneficiary at the same time. Revenue Architecture is blunt that cross-sell should never sit with a CSM, because it is a genuine sale and not an account-management task.

Tool-agnostic on purpose

Here is the property that makes the Bowtie an Operations asset rather than a whiteboard drawing: it is standardized and tool-agnostic. The volume spine runs VM1 through VM9 regardless of motion. In a product-led motion, VM3 is a product-qualified lead; in an enterprise motion, VM3 is a marketing-qualified account. Different names, different tools collecting them, but both are VM3, which means a self-serve motion and a field motion can be measured against the same structure and compared honestly. That comparability is also what makes benchmarking possible at all: the 868-company benchmark set only exists because everyone in it maps to one shared spine.

Standardization has a cost, and the cost is discipline. Revenue Architecture is explicit that once the model is agreed, adherence has to be enforced through a change-request process, because moving a single stage definition silently rewrites every downstream metric and, eventually, the valuation. I have watched a team quietly redefine what “qualified” meant and spend a quarter arguing about a conversion rate that had not actually changed. The model was fine; the governance was missing. We will build that governance out in Module 5, where the Bowtie stops being a diagram and becomes an object-and-field model in the CRM.

What you can do now

You can draw the whole customer lifecycle as one shape and put every part of the business on it. You can say where units become dollars (the knot, VM6 = VM5 × ACV), which side is an open loop and which is closed, and which of the four growth areas a given expansion actually is. You can also answer the board member’s question, because the Bowtie gives you one model where the closed deal and the expansion eighteen months later are the same customer moving through connected sub-systems.

What the Bowtie does not have yet is numbers. Right now it is a shape with eight labeled segments and a knot. The next chapter, the Metrics Spine, hangs the measurements on it: a volume metric, a conversion rate, and a time metric at every segment (VM, CR, and Δt), plus the distinction between the leading indicators on the left that predict the quarter and the lagging ones on the right that confirm it. The shape tells you what to measure and where. The spine tells you the actual state of the factory.