The 7 Decision Gates: why B2B deals die — and how to see it coming

Stage-based pipelines measure the seller, not the buyer

Open any CRM and you will find a pipeline: Discovery, Demo, Proposal, Negotiation, Closed. It looks like a map of the deal. It is actually a log of the seller's activity. A deal sits in "Proposal" because the seller sent a proposal — not because the buyer decided anything.

That gap is where deals die. Late-stage CRM data is self-reported by the rep, and nobody audits it. The rep had a good call, felt momentum, and dragged the card one column to the right. Meanwhile the buyer still has three unanswered questions and a CFO who has never heard of you. The forecast says 70% probability. The buyer's head says "maybe, eventually."

The fix is to track the thing that actually predicts revenue: the buyer's decision-making process. Every complex B2B purchase moves through the same sequence of internal commitments. We call them the 7 Decision Gates — seven buy-ins a buyer gives, in order, before a deal closes. Each gate answers one question in the buyer's head. Until that question is answered, no amount of seller activity moves the deal.

This model was built from patterns across 170+ B2B sales processes. It is the backbone of Cosa's methodology, and you can apply it with or without software.

The seven gates, one buyer question at a time

G0 — Pre-Qualification: "Is this transformation my priority right now?"

Before the first real conversation, the buyer decides whether the problem you solve ranks high enough to spend time on. If it does not, nothing downstream matters. This is the single most common point where deals die quietly — "not now, we'll build the team first, talk to us in Q3." Sellers tend to accept that deferral at face value. The better move is to make the cost of waiting concrete: the deals leaking from the pipeline while the buyer builds the team are exactly what the change would pay back on.

G1 — Qualification: "Do they understand my problem better than I do?"

The buyer needs to hear their current state, desired state, and blockers reflected back more precisely than they could state them. Buyers rarely contest this gate out loud — and sellers rarely confirm it out loud either. That is a free win left on the table: asking "did I capture your problem the way you would describe it?" costs nothing and locks the gate explicitly.

G2 — Thesis: "Does this approach get me to my goal faster than what I do today?"

This is belief, not features. The buyer compares your approach against their status quo — including the quiet alternative of "we'll get this clarity ourselves by reviewing deals manually." Two objection shapes dominate here: the do-it-yourself belief, and "this sounds similar to a tool we already have." Assertion does not win this gate. Evidence does — ideally a run on the buyer's own data, and a precise naming of what the current tool structurally cannot produce: where the deal actually is, why it is stuck, and what to do next.

G3 — Mechanism: "Do I understand how this actually works?"

The largest cluster of buyer questions in real sales calls lands here: how does data get in, how does it get out, how does it integrate with the CRM we already run? Buyers will re-ask a mechanism question until it is answered concretely. Deferring it to "we'll cover that in the first session" stalls the gate. Showing the mechanism live on the buyer's own stack passes it.

G4 — Onboarding: "Does my team have the resources to pull this off?"

Belief in the approach is not belief in the rollout. The buyer now weighs timeline, internal bandwidth, and what their side has to commit. "We don't record calls," "the tech team is maxed out," and "shift it to next quarter, it needs prep" all live here. The gate passes when the resource ask is explicit and small — a minimum-input path, a clear first-30-days picture, and a short list of what you need from them.

G5 — Feature Alignment: "Are these features enough to get me where I need to go?"

Not a feature tour — a sufficiency check. The buyer maps what you offer against the promised goal and probes the edges: "does this still work if we can't supply call recordings?" A live end-to-end example on one concrete case answers this gate better than any feature list.

G6 — Pricing: "Is this price fair for what was promised?"

In real calls, pricing objections are rarely about the number. They are about ROI defensibility ("we're pre-profit — what does this actually return?") and cashflow. The buyer who asks for justification is usually preparing to defend the spend internally. The gate passes when the buyer can take a number upstairs that survives scrutiny — which is why transparent unit economics land as fairness, not cost. See how this plays into Cosa's pricing.

After G6 — internal alignment

Even with all seven gates confirmed, one phase remains: the buyer sells the decision internally to stakeholders who were never in your calls. "I can't decide this alone" is not an objection — it is a request for forwardable material. The sellers who win here arm their champion with something the absent CFO can verify, rather than waiting for group consensus to form on its own. This is exactly what provenance-tracked deal intelligence is built for.

Three rules that make the gates useful

Gates are sequential, and skipping shows. A buyer discussing price before confirming the mechanism has not skipped ahead — the deal has a hole in it. Gate-skipping is one of the most reliable risk signals in a pipeline. When you see it, go back and close the open gate instead of pushing forward.

Track gates per stakeholder, not per deal. The Head of Sales may be at G5 while the CFO is at G0. A deal-level "stage" averages those two into a fiction. Complex deals close when every relevant stakeholder has passed the gates that matter for their role — which is why the decision maker's confirmations, not the champion's enthusiasm, advance the deal.

Confirmation must come from the buyer's mouth. A gate is confirmed when the buyer expresses it — not when the seller believes it. The honest standard distinguishes three states: confirmed in the conversation, assumed from context, or simply not mentioned. Most pipeline surprises are assumed gates that were never real. Cosa applies this standard mechanically: gate status is read from what was actually said in calls and emails, tagged with Provenance, so a seller cannot inflate a gate any more than they can inflate a transcript.

Using gates to qualify — and disqualify

The practical payoff is qualification you can defend. Instead of "this deal feels strong," you can state: G1 and G2 confirmed by the decision maker, G3 confirmed but only by the technical lead, G4 not mentioned, G6 raised once with no resolution. That sentence tells you the next call's agenda, who needs to be in the room, and whether the deal deserves the time at all.

It also tells you when to walk away. A buyer who will not confirm G0 — for whom this is genuinely not a priority — is not a deal to nurture with monthly check-ins. They are a disqualification that frees your calendar for the deals where the gates are actually moving. Pipeline reviews built on gate evidence take minutes, not meetings, because the question is no longer "what does the rep think?" but "what did the buyer say?"

That shift — from seller activity to buyer commitment — is the whole model. The companies that adopt it stop being surprised by their own pipelines. To see how the gates compress a 4–6 month cycle into a few calls, read how validated offers close complex deals in 2–3 calls.