From FR Advisory.
Your forecast has a category of deal in it that is already dead.
Not at risk. Dead. The champion is still responsive, the next call is in the diary, the stage is still Negotiation, and the thing that will kill it has already happened in a meeting nobody invited you to.
That is Stage Two. It is the stage where your deal stops being about your product and starts being about somebody's career, and almost every sales methodology ever written is silent on it.
The stage, defined
What starts it. The account has a category and a candidate set, and someone inside it decides to try to get money. That is the handover from Stage One.
What happens in it. A document gets written. It has a different name everywhere, business case, paper, submission, capital request, but it is always the same object: a written argument that this spend is worth more than the other things the money could do, and that nobody will be blamed if it goes wrong.
What ends it. Either a funded commitment to proceed, or nothing. In 38% of major initiatives it ends in nothing at all, and nothing is not a competitor.
Who owns it today. Your champion. Alone. Unmanaged, unsupported, usually unqualified to do it, and certainly unpaid for it.
What done looks like from your side. The document that reaches the decision contains your numbers, in their format, answering the question that will actually be asked, which is not the question you have been answering.
How long it lasts. An average of 5.35 additional weeks of internal alignment after the buyer has already chosen a preferred supplier. Five weeks, after they picked you. Look at your average stage duration and tell me honestly whether your pipeline knows that.
The numbers that should end a few arguments
G2 surveyed more than a thousand B2B software buyers and decision makers this year, alongside interviews with over fifty sales and marketing leaders.
49% said their chief financial officer reversed a deal the buying team had already approved, in the last twelve months.
Read it again. Not deals that stalled. Not deals that lost. Deals the buying team approved, then had taken away from them.
Finance involvement in software decisions rose from 31% to 46% in a single year. Not influence. Presence. The person who kills your deal is now in the room for nearly half of them, and they arrived after you built your entire account plan.
Then the number that breaks the usual consolation. The veto rate is 54% at companies with a dedicated token or large language model budget, against 29% at companies without one.
The more serious the buyer is about AI, the more likely you are to be killed late.
Everyone selling agentic AI has been told to find the sophisticated buyer, the one with budget already allocated, the one who gets it. That buyer vetoes you nearly twice as often. Sophistication does not mean permission. It means scrutiny.
And the detail that should genuinely unsettle anyone who runs a forecast: three in four buyers whose deal was vetoed late still expected positive return within six months of signing.
They believed. They were right about the product. They lost anyway.
Your champion's conviction is not evidence. Your forecast treats it as evidence. That is the gap this stage lives in.
What the buyer is actually optimising for
Separate research, VML's study of 750 B2B buyers across eight markets, puts two numbers next to each other that explain more about enterprise selling than any methodology I have read.
71% will choose a safe, well-known supplier over a rival they believe offers better value.
Believe. Not suspect. They have assessed you as better and will buy the other one anyway.
80% consider how a purchase will affect their professional reputation.
There it is. The actual decision criterion, stated plainly by the people making the decision, and absent from every business case template ever circulated.
The business case is not an argument about value. It is a document that transfers risk away from the person signing it. Return on investment is in there because the format demands it, not because it decides anything. What decides it is whether the signer can survive being wrong.
So when your champion asks for another case study, they are not stalling. They are collecting armour.
And when you send more product detail, you are arming them for the wrong fight.
What kills these deals, in order
From the same study, the reasons major initiatives stall:
Risk and compliance, 39%. Budget or return uncertainty, 34%. Shifts in corporate strategy, 31%. Procurement complexity, 27%. Misalignment inside the buying group, 26%.
Notice what is not on that list. Your competitor. Your price. Your product.
Four of those five are internal conditions inside the customer's business. You cannot sell your way out of a strategy shift. You can only know it is coming, and most sellers do not, because nobody asks.
Meanwhile only 40% of major decisions give the business unit leader meaningful sway. In most deals, the person whose problem it is does not control whether it gets solved. Your champion may be the most senior person you have met and still be a spectator in the room that matters.
What the seller actually does
This is the hardest behaviour change in the whole framework, because it requires a seller to stop doing the thing they are best at.
You cannot attend this meeting. Ever. No amount of relationship gets you into the capital allocation conversation. So the job is not persuasion. The job is to write the document, hand it to your champion, and accept that someone else gets the credit for it.
Stop doing this. Sending the deck again. Offering another demo for the wider team. Asking whether they need anything else from you. That last one is the most common and the most useless question in enterprise selling, because the champion does not know what they need and will not admit it.
Start asking these.
"Who signs this, and what is the last thing they rejected?"
"What format does a request like this have to be in?"
"Which other projects is this competing against for the same money?"
"If this goes badly in six months, who gets asked about it?"
"What is the thing someone will say in that meeting to kill it?"
That last question is the single most valuable question in enterprise sales and almost nobody asks it, because it invites a bad answer. Ask it anyway. The champion almost always knows. They have usually already heard it.
Then write the document. Not a deck. A one to two page argument in their language, with the risk section first and the return second, naming the objection you were told about and answering it before it is raised. If your company cannot produce that inside 48 hours, you do not have a sales enablement function, you have a slide library.
The diagnostic
Score each 0, 1 or 2. Ten minutes, no tooling.
1. In your last ten losses, do you know which were no-decision rather than competitive losses? 0 we record them all as lost to competitor, 1 roughly, 2 they are separate outcomes in the system.
2. Can you name the person who signs, on every deal above your average size, in your current pipeline? 0 no, 1 on the big ones, 2 on all of them.
3. Have you ever read one of your customers' actual internal business cases? 0 never, 1 one or two, by accident, 2 we routinely ask for them and often get them.
4. Does your forecast distinguish champion confidence from committed budget? 0 no, 1 the manager does it in their head, 2 it is a field and it changes the weighting.
5. Do you know your own veto rate, meaning deals approved by the buying team and then reversed above them? 0 we do not measure it, 1 anecdotally, 2 a number with a trend.
6. Can your sellers state, in a live deal, what the objection in the approval meeting will be? 0 the question would confuse them, 1 the good ones can, 2 it is required to advance the stage.
0 to 3. You are forecasting hope. Roughly half your approved deals are exposed to a reversal you cannot see and do not record, and your win-loss analysis is attributing those losses to competitors who were never there.
4 to 7. You can see it, you cannot act on it. The information exists in individual sellers' heads and leaves when they do.
8 to 12. Instrumented. The remaining question is whether anyone changes a forecast because of it.
What to instrument
Separate no-decision from competitive loss. Right now these are almost certainly the same field. They are completely different failures with completely different fixes, and merging them makes your win-loss analysis actively misleading. If 38% of major initiatives end in no decision and your system says you lost them to rivals, you are optimising against an opponent who was never in the room.
Add Economic Approver, and require it to advance. Not the champion. Not the user. The name of the person who can reverse the decision. If the seller cannot name them, the deal does not move to the next stage. This will shrink your pipeline on the day you introduce it, which is the point.
Add Veto Flag, and derive Veto Rate. Any deal approved by the buying team and then reversed above them. You need the number, by segment and by deal size. If it is anywhere near G2's 49%, your forecast is not a forecast.
Derive Alignment Lag. Days between the buyer naming you as preferred and a signed commitment. The benchmark to beat is about five and a half weeks. Most pipelines assume days. That delta is where forecast accuracy goes to die every quarter.
Record the objection. One free-text field, captured in discovery: what will be said in the approval meeting to kill this. Read them monthly. You will find three objections account for most of your losses, and you will discover that none of your marketing material addresses any of them.
Stop treating champion enthusiasm as a stage gate. It correlates with nothing. Three in four vetoed buyers were enthusiastic and expected returns within six months.
What changes underneath
Sales strategy and enablement. Qualification gains a hard gate on the economic approver. Enablement's job changes from producing material that helps sellers sell, to producing material that helps champions argue when the seller is not there. Those are different artefacts. One is a deck. The other is two pages with the risk section first.
Pricing and packaging. This stage is where pricing structure stops being a commercial preference and becomes a political tool. A smaller first commitment, a shorter term or a usage-linked floor reduces the career risk of the person signing, which is the actual objection. 70% of buyers say the pace of AI change is pushing them toward shorter contracts. 40% of buyers who have been vetoed push for terms under twelve months, against 18% of everyone else. Your twelve-month minimum is not protecting revenue. It is manufacturing no-decisions.
Revenue architecture. Preference for outcome-based pricing doubled in a year, from 11% to 23%. Half of buyers have been offered variable pricing and another 42% expect it, and 52% said variable pricing improved their opinion of the vendor. The structure that survives a chief financial officer's scrutiny is the one where the customer's exposure scales with what they get. That is a revenue architecture decision taken quarters before the deal, not a discount decided in the room.
Commercial systems. Three new fields, two derived numbers, one outcome type split in two. All of it cheap. None of it will happen unless the forecast review starts asking who signs.
The first ninety days
Week one. Pull your last twenty losses and re-code them by hand into competitive loss and no decision. Do not automate this, do not delegate it, and do not skip it. The ratio is the entire business case for everything else on this list.
Weeks two to four. Add the economic approver field. Brief the team on the five questions, especially the last one. Make naming the approver a requirement to advance, and hold the line the first time a big deal fails the gate, because that is the only moment the policy is actually decided.
Weeks five to eight. Write the two-page document for your three most common deal shapes. Risk first, return second, the known objection answered explicitly. Give it to champions, not to sellers.
Weeks nine to twelve. Measure alignment lag for the first time. Expect it to be roughly double what your stage durations assume. Then take that number to whoever builds the forecast model and watch the conversation get difficult.
What not to do. Do not try to get into the approval meeting. Do not escalate to the chief financial officer yourself, which converts your champion from an ally into someone who has been embarrassed. Do not add a stage to the pipeline. You are not adding a step, you are admitting one already exists.
The position
You won the champion. You did not win the deal, and in this market those have stopped being the same thing.
Half of buying teams had a decision reversed above them last year. Finance went from a third of decisions to nearly half in twelve months. The more AI budget the buyer has, the more likely the veto. Seven in ten will take the safe supplier over the one they think is better, and eight in ten are thinking about their own reputation while they do it.
None of that is in your pipeline. All of it is in your results.
The business case is not written to prove value. It is written so that somebody can survive having signed it.
Write that document, or watch someone lose the argument on your behalf, in a room you were never going to enter.
Next: Stage Three, Compares. The shortlist you are on was assembled by something that has never used your product, and the comparison is already running without you.
The Agentic Commercial Model. agenticcommercialmodel.com
Sources
- G2 2026 Buyer Behavior Report, "The Evaluation Maze", survey of more than 1,000 B2B software buyers and decision makers plus interviews with more than 50 B2B sales and marketing leaders, reported July and September 2026: 49% of buyers said their CFO reversed a deal the buying team had already approved in the last 12 months; finance involvement in software decisions rose from 31% to 46% year on year; CFO veto rate 54% at organisations with a dedicated token or LLM budget against 29% without; three in four buyers who experienced a late-stage veto still expected positive ROI within six months of signing; 40% of vetoed buyers push for contract terms under 12 months against 18% of other buyers; 70% say the pace of AI innovation is pushing them toward shorter contracts; preference for outcome-based pricing rose from 11% to 23% since 2025; half of buyers have been offered variable-cost pricing with a further 42% expecting it, and 52% said variable pricing improved their view of a vendor; internal resistance to AI adoption rose from 16% to 29% in a year; IT security review is the largest source of post-selection delay at 39% overall and 50% among enterprise buyers, followed by budget approval at 32%; evaluation has overtaken research as the longest stage of the buying journey; 80% of organisations give developers or technical teams a dedicated token or LLM budget. G2 does not publish fieldwork dates, geographic breakdown or margin of error for this report
- VML Future Buyer Report 2026, survey of 750 B2B buyers across eight international markets spanning manufacturing, energy, transportation, finance, professional services, telecommunications and healthcare, reported 14 November 2026: 38% of major B2B initiatives end without a decision; an average of 6.8 stakeholders is involved in a significant purchase; buyers need an average of 5.35 additional weeks of internal alignment after selecting a preferred supplier; finance is the most influential function in final supplier approval; business unit leaders have meaningful sway in only 40% of major decisions; 71% would choose a safe, well-known supplier over a rival they believe offers better value; 80% consider how a purchase affects their professional reputation; stall reasons are risk and compliance 39%, budget or ROI uncertainty 34%, shifts in corporate strategy 31%, procurement complexity 27%, misalignment within the buying group 26%; abandonment highest in telecommunications at 44.1% and finance at 40.8%. Fieldwork dates and methodology beyond sample size are not stated, and the reporting itself cautions that the B2B sample is small enough that sector splits should be read with care
- Both studies are vendor-published research rather than peer-reviewed work, and neither states a margin of error. They are used here because they agree with each other on the direction of every shared finding, not because either is precise