The committee nobody mapped

Industrial technology vendors often lose deals they never knew were at risk. The demonstration goes well, the engineers are enthusiastic, and a pilot gets approved. Then the purchase order never arrives, and the pipeline report calls it a long sales cycle.

Sometimes it is one. More often, in our experience, the vendor has a persona problem that began in the first meeting: it built its evidence for the person who liked the product, and the purchase gets decided by people who were never in the room.

This paper is written for revenue leaders at technology vendors whose industrial pipeline fills and does not convert, and for OEMs launching digital products into an installed base. Operators approving a rollout can read it from the other side: the same seats decide whether a pilot reaches your other sites.

The argument starts with enthusiasm.

Engineers engage generously with good technology, which tells you the technology is credible and tells you nothing about the purchase.

Polite interest is weaker still, because a statement that can never turn out to be wrong cannot count as evidence; a commercial signal is specific enough that the buyer would pay something to fake it. The champion and the economic buyer, in a plant usually the plant manager or VP of operations, need different evidence, so a program that interviews only the champion leaves the purchase decision untested.

An industrial committee also has more seats than most sales motions brief. Operations, IT and operational technology (OT), environment, health, and safety (EHS), procurement, engineering, finance, and a plant manager with site-level veto can each stop a deal, and each wants its own evidence. In our experience, deals rarely die at evaluation. They die in procurement, when a specific budget line has to produce the money, or in integration, when the OT team is consulted for the first time.

All of this can be mapped before the sales motion is built. Primary research can find the seats, what each requires, and where comparable deals stalled, long before a sales team discovers them one account at a time. The appendix is a committee map template: each seat, the evidence it requires, and the point at which it can stop the deal.

Engineers at industrial companies are technically curious. They attend demonstrations willingly and give detailed, encouraging feedback, and that feedback is sincere. It confirms the technology is credible.

What it leaves open is whether the engineer controls a budget, whether the problem they described ranks in the top three for the plant manager who controls capital, and whether their enthusiasm will survive a multi-seat procurement review.

The pilot agreement deepens the illusion, because a pilot is cheap for the buyer's organization. It usually needs no capital approval, no infrastructure change, no safety review, and no procurement sign-off. A signed pilot shows that the champion found the value proposition credible enough to test, and nothing beyond that.

Technical interest arrives almost immediately. In our experience, where industrial procurement is involved, the purchase order commonly arrives many months later, if at all. Deals stall at the edge of the champion's authority, when the pricing model does not fit how the buyer allocates budget, procurement asks for evidence nobody gathered, or the economic buyer was never in a single conversation.

For a vendor, that misread reaches well past sales. The team runs the wrong motion, engineering answers OT integration questions nobody planned for, and finance forecasts on the wrong sales-cycle assumption.

Most vendors start scaling the sales motion when pilot results come back positive. That is also when the commercial hypothesis most needs testing, since everything after it spends real money.

Industrial buyers are courteous, and courtesy sounds like interest.

"That's an interesting approach." "We're always looking at ways to improve this." "Send me some information and we'll take a look." Each statement is sincere, costs the buyer nothing, and commits them to nothing.

The market makes this worse. In Redwood Software's 2026 manufacturing survey, 98% of manufacturers said they were exploring or considering AI-driven automation, and only 20% felt fully prepared to use it at scale. When nearly everyone is exploring, polite engagement with new technology is close to a professional norm and nearly worthless as evidence.

What a commercial signal sounds like

A commercial signal is specific, and it would cost the buyer something to fake. It includes:

• A named problem, ranked against the buyer's other priorities.
• A number attached to what the problem costs them today.
• A budget line they can point to.
• An honest account of who else would need to approve, and what that process looks like.
• A concrete next step that the buyer proposes.

The difference is falsifiability. A polite statement can never turn out to be wrong, so it cannot count as evidence.

The market also sets the price of the misread. In enterprise software, a pipeline built on politeness gets caught at the next quarterly forecast. In industrial markets the purchase order sits far behind the first enthusiastic conversation, so the same pipeline can carry a company through a full planning cycle, with hires made and a raise sized against it, before anyone admits nothing is converting. Scoring every conversation for specifics from the first interview is the cheap fix.

Score the conversation

We score every buyer interview on five dimensions for that reason, each from one to five:

• Urgency: is the problem costing the buyer something now?
• Priority: where does it rank against everything else competing for the same attention and budget?
• Budget: does a budget line exist, or could one be created?
• Authority: can this person approve the spend, or must they sell it upward?
• Openness: would the buyer engage a new vendor at all?

A buyer can score five on urgency and one on authority, which describes a passionate advocate with no signature. Ask a single "would you buy this?" and those two scores average into a misleading yes. Scored separately, the profile shows what the deal requires: this person champions, and someone else approves.

Hedged and generic answers score low on whatever dimension they hedge. We set a strong signal at an average of 4.0 or above, declared before any interviews take place, so two enthusiastic conversations cannot carry a sprint past eight polite ones.

A shared rubric also makes interviews comparable. When every conversation in a sprint is scored on the same five dimensions and the same scale, the results line up regardless of who ran each one, and patterns appear that no single interviewer would report: budget scores clustering low in one segment, for example, or authority sitting one level above the persona being interviewed in account after account. The aggregated scores become the evidence base for the verdict, through the Decision Matrix our paper on pilot purgatory describes.

The technical champion evaluates whether the solution works: its capability, its integration, the vendor's credibility. The economic buyer evaluates whether the organization should pay for it, which covers the budget line, the approval process, competing priorities, and how the spend will be justified internally. In most industrial organizations these are different people. The engineer who found the demonstration impressive is rarely the plant manager, the vice president of operations, or the chief financial officer who signs the capital request.

Pipelines built on champion evidence follow a recognizable course. They fill, win evaluations, get pilots approved, and then stall at the approval step the champion cannot carry.

Two signatures, two commitments

A letter of intent (LOI) shows the same gap on paper. A technical champion can sign an LOI at almost no organizational cost: no capital approval, no procurement review, no safety sign-off, no budget allocation. It expresses real interest and commits nothing.

A purchase order requires the buyer's organization to run its full approval process, and under our evidence standard that is the boundary between Benchmarked and Verified. An LOI from a champion rates Benchmarked, as real support for the hypothesis without confirmation. A purchase order that went through procurement rates Verified.

Many vendors approaching a board review hold a portfolio of Benchmarked claims and call it traction. They find out the difference when an investor or a first enterprise customer asks for the evidence behind the pipeline slide.

Investors who know industrial markets price that difference. In our experience they now examine commercial evidence as closely as they once examined the technology, and a list of pilot partners who liked the product no longer answers their question. They want to know whether a named economic buyer confirmed willingness to pay at a specific price, through a procurement path the buyer's own organization described. More letters of intent will not close that gap. Scored interviews with economic buyers in the target customer profile will, and the same evidence carries into the first enterprise contract negotiation, which asks the identical question.

Confirm the decision structure first

Every downstream hypothesis, including pricing, channel, timeline, and contract structure, is calibrated against whoever answered the interview questions. If that person cannot approve the purchase, every finding is calibrated against the wrong answers.

So we treat access to the decision maker as its own scored dimension. Reaching the economic buyer is the research plan, and in many sprints the most important finding is the distance between what the champion believes and what the economic buyer requires.

Complex business purchases involve more people than most sales motions plan for. Gartner's widely cited research puts a typical buying group for a complex B2B solution at six to ten decision makers, and its 2025 sales survey found buying groups of five to 16 people across as many as four functions.

In the industrial purchases we assess, the seats with real power to stop a deal usually include:

Seat	What they need to see
Operations (VP or Director of Operations or Manufacturing)	Proof the system survives contact with the production schedule
IT and OT (controls or automation engineers)	Integration and cybersecurity answers specific to their own stack
Environment, health, and safety (EHS manager; HSE in energy)	The safety and liability model, in writing
Procurement	A contract structure and pricing model that fits an approvable budget category
Engineering	Technical depth
Finance (the plant controller checks the value case)	A return case that survives their own assumptions, from a budget someone owns
Plant manager	What happens to throughput, and to their people

The technical champion usually holds one of these seats. Winning the champion leaves procurement undecided, and winning procurement still leaves the EHS review. A presentation that answers the champion's questions answers a small part of the room.

A sales motion built for fewer seats fails without much noise. It generates pipeline and wins evaluations, then stalls at stages the customer relationship management (CRM) system labels "procurement" or "security review." Often those stages are stakeholders meeting the product for the first time, with nobody assigned to their questions.

The forecast breaks for the same reason. Rep capacity, lead volume, and conversion rates borrowed from enterprise software assume a shorter cycle than industrial procurement runs, and every unmanaged seat adds cost across the longer one. Finance ends up managing a forecast built on pilot timing when it needed procurement timing.

A successful pilot adds seats

The committee also grows as a deal moves from pilot to production. A pilot usually impresses the people closest to it: the field team, the site engineer, and the operations lead at one closely managed site. Production brings in approvers who were never part of it.

Safety and legal arrive once the system will run without the vendor watching. Multi-site operations leadership arrives when the question becomes whether results carry across plants with different equipment and different crews, and finance arrives when the spend moves from an innovation budget to a capital request. Each brings a bar the pilot was never tested against: liability, generalization across sites, maintenance capacity, or an approval process with its own evidence requirements.

A pilot that succeeded on its own terms can stall indefinitely at that gate. The innovation budget that approved the pilot does not approve the fleet, and the champion who ran the evaluation rarely controls the capital.

Caught early, this is cheap to fix. While the pilot is still running, identify who will approve production, find out what evidence each of them will require, and start gathering it. A team that learns the list only after submitting the production request has turned a successful pilot into a long one.

The most common place an industrial deal dies is a budget line.

A USD 50,000 annual subscription and a USD 150,000 capital purchase can have the same three-year cost and still be approved by different people, at different thresholds, through different processes. Capital purchases usually run through approval chains with committee review and formal justification. Operating spend runs against a budget someone owns and defends every year, with renewal scrutiny each cycle. Neither path is easier in general; each is easier for a purchase shaped to fit it.

Technology vendors often default to subscription pricing because that is how enterprise software sells, while, in our experience, industrial buyers frequently approve technology as capital spend (capex), not operating spend (opex). A subscription that reaches an organization used to capital approval can end up in a budget nobody owns.

This mismatch almost never loses at evaluation, which is why it goes undetected. The champion likes the product and the committee accepts the case, and then the deal stalls at the step where a specific budget line has to produce the money. The CRM records a procurement delay, and the sales team runs the same motion into the same wall at the next account.

Outcome models raise the stakes

The mismatch gets sharper as commercial models change. Equipment-as-a-Service and outcome-based contracts move the service, and sometimes the machine, into operating budgets. Most capital-heavy buyers keep the asset and buy the maintenance by the hour; The servitization bet sets out why. For OEMs the margin case is real: Deloitte's 2026 manufacturing outlook puts aftermarket service margins at more than twice those of equipment sales alone.

Each of those contracts, though, asks the buyer's finance organization to approve a kind of spending it may have no process for. An outcome contract also has to survive its first dispute.

For the OEM, the model also moves the performance risk for the life of the contract, which makes the renewal a stronger test than the signature. What the OEM takes on, and how to test it before launch, is covered in The servitization bet.

Ask the buyer which budget

Which budget would pay for this purchase, who owns that budget, and what evidence does the owner require? Ask the buyer directly, in research, before the pricing model is committed. Willingness-to-pay evidence that does not name the budget and its owner rates Benchmarked at best. Design the commercial model around the confirmed answer, even when a different model would be easier to invoice.

IT and OT are converging. IDC predicts that by 2027, 40% of operational data will be integrated across applications and platforms autonomously. None of that changes what a vendor meets at a specific plant this year: controllers and historians of mixed age, supervisory control systems under strict change control, cybersecurity requirements with no cloud-native equivalent, and an OT team whose first duty is uptime.

Security is where this seat most often says no. In Cisco's 2026 State of Industrial AI research, 40% of the more than 1,000 operational technology decision-makers surveyed named cybersecurity as the biggest obstacle to scaling AI [7].

"We integrate with your OT stack" is therefore a claim, and only the buyer's OT team can say whether it is true at their sites. The question often goes unasked until after the pilot is scheduled, because the commercial conversation ran through IT or operations leadership.

Data access is a commercial term

Most industrial initiatives depend on equipment or operational data. Who controls that data, what quality it has at the target installations, and on what terms it will be shared are regularly unresolved at launch. In our assessments, OEMs assume access because they installed the equipment, then learn that plant operations controls the data. Vendors assume modern interfaces and then meet historians.

KPMG's 2026 survey of 258 industrial manufacturing technology leaders captures the gap: 83% believe their organization is building strong AI data foundations, while 76% cite unreliable data as a top AI risk.

Closing it costs two interviews. Talk to the OT team and the data owner during validation, before the pilot and before pricing, and the largest silent risk in the deal becomes a scored finding: integration fit at a commercially viable cost gets a Verified, Benchmarked, or Assumption rating like any other claim. A product that cannot integrate with the buyer's OT stack at a viable cost and timeline is not viable for that buyer, however well it performs elsewhere.

Original equipment manufacturers (OEMs) launching a digital product face a version of this problem that their own strengths hide.

Decades of installed base, trusted relationships, and a channel that reliably reaches the plant manager who approves capital equipment all validate the hardware business. The digital product borrows that confidence without having built any of its own. Its buyer is often an IT or OT leader or a digital transformation executive where the hardware buyer was the plant manager. Its budget is operating spend, renewed annually and justified by ongoing value, in place of a one-time capital approval. And the distributor who carries hardware to the plant manager was never set up to sell subscriptions to IT leaders, and often has no commercial reason to try.

There is also a stakeholder inside the OEM. In our experience, field service teams are often paid on service calls, parts, and time-and-materials work, and a subscription or outcome model can cut what they earn. Unless those incentives change, the team closest to the customer resists without saying so. Confirm that field service leadership is paid toward the new model before the external rollout, or the transition stalls in deployment after the sale has closed.

Validate the digital hypothesis separately, even at accounts the OEM knows well. Build the digital buyer profile independently of the hardware one, and interview the digital buyer instead of the hardware contact. Test the pricing against the budget that would pay it. Confirm the channel's willingness and its capability with the partners themselves, because those are separate questions.

Skipping this separate validation moves the discovery to the worst possible point, mid-transition, with the hardware revenue model winding down and the digital one not yet converting. The cost lands on the functions with the deepest installed-base relationships. Sales meets buyers who trust the brand for hardware and judge software on other terms, and channel partners have the wrong conversation with the wrong person.

Each seat also has a step. Between "the committee accepts the case" and "the purchase order arrives" sits the paper process: the procurement review and vendor onboarding, the IT and OT security review, the safety review, and legal's pass over liability, data and termination terms. Sales methods that track complex deals treat this as its own stage, which the MEDDPICC method calls the paper process [8], because it is where many deals that were already won go quiet. Each step has an owner, a queue and evidence it requires. A plan that names the steps, their owners and the evidence each needs before the first request goes in is shorter than one that discovers them in sequence.

The other outcome the forecast rarely shows is no decision at all. Matthew Dixon and Ted McKenna analyzed more than 2.5 million recorded sales conversations for The JOLT Effect and found that 40 to 60% of deals end in no decision rather than a loss to a competitor. About 56% of those losses come from indecision, the fear of making the wrong choice, and 44% from preferring the status quo [6].

A committee multiplies that fear. With six to ten people who can each stop the purchase, every one of them is asking a private question: if this goes wrong, can I defend having said yes? An industrial buyer has good reason to ask it. A failed rollout is visible on the shop floor, in the capital report and to the board.

That is why the evidence matters more than the enthusiasm. A champion can argue for the purchase. A sponsor who can show which claims the buyers confirmed, which rest on benchmarks and which are still assumed, graded by someone with no stake in the answer, can defend the decision in either direction. Defensibility is what moves a frozen committee, and it is also what lets a sponsor stop a bad purchase without it reading as a personal failure.

Primary research can uncover the buying committee before the sales motion is built. For a given product category and account type, it can establish who sits on the committee, what each seat requires, and where comparable deals have stalled. A sprint that confirms the decision structure turns the committee from a series of surprises into a briefing plan.

Skip that step and you find the fifth seat in month eleven of a deal the forecast counted in month six.

Diagnosing a pipeline that does not convert

For a vendor already in market, the committee usually shows up first as conversion below forecast. Three causes produce that single symptom:

1. The wrong ideal customer profile (ICP), which fills the pipeline with accounts that were never going to close.
2. A value proposition that persuades the champion and misses the economic buyer.
3. A commercial model that does not fit how the buyer approves spend.

More sales coaching fixes none of them, and more sales capacity spends money against the same broken assumption. Structured interviews with the accounts that stalled, as opposed to the ones that closed, show which cause is at work.

When a vendor scales on an unvalidated motion, the cost spreads beyond sales. Support, finance, and customer success each inherit a different symptom of the same misread, a pattern our paper on pilot purgatory sets out in full. In our experience, validating the committee before scaling costs a small fraction of one quarter of misdirected sales capacity.

Complete one row per seat, for each target account type, during validation. Every cell should cite the interview that supports it. A blank "evidence required" cell is an Assumption.

Seat	Named person or role	Evidence this seat requires	Point at which this seat can stop the deal	Interviewed?	Score (1 to 5)
Technical champion
Economic buyer (plant manager or VP of operations)
Operations
IT
OT (controls or automation engineers) and the data owner
Environment, health, and safety (EHS)
Procurement
Finance (plant controller)
Plant manager
Channel partner, if any
OEM field service, if an OEM

Questions to complete the map

1. Who approves this purchase, and have they been interviewed separately from the champion?
2. Which budget pays for it, who owns that budget, and does the pricing model fit how that budget approves spend?
3. What would the procurement team need to see, and has anyone asked them?
4. Has the OT team confirmed integration at a viable cost at this buyer's sites?
5. Who controls the operational data, and on what terms will it be shared?
6. What would the EHS function need in writing?
7. For an outcome contract, what happens the first time the outcome is disputed?
8. For a service or outcome model, what would a customer approaching renewal do if the service were unbundled and quoted from scratch?
9. Where did comparable deals stall, and at which seat? Which seats join only when a pilot moves to production?
10. For an OEM, is field service paid toward the new commercial model?
11. For a pipeline already in market, what did the accounts that stalled say?
12. What are the steps of the paper process (procurement, IT and OT security review, safety, legal), who owns each, and what evidence does each need?
13. What would each seat need to see to defend a yes, and to defend a no, if the decision is questioned later?

1. Gartner research on B2B buying groups, as widely cited: a typical buying group for a complex B2B solution involves six to ten decision makers. Reported, for example, in Madison Logic, "Navigating the Rise of the Buying Committee in B2B Sales." https://www.madisonlogic.com/blog/navigating-the-fall-of-the-individual-buyer-and-the-rise-of-the-buying-committee/ The 2025 range is from Gartner, "Gartner Sales Survey Finds 74 Percent of B2B Buyer Teams Demonstrate Unhealthy Conflict During the Decision Process," press release, 7 May 2025. https://www.gartner.com/en/newsroom/press-releases/2025-05-07-gartner-sales-survey-finds-74-percent-of-b2b-buyer-teams-demonstrate-unhealthy-conflict-during-the-decision-process
2. Redwood Software, "Manufacturing AI and Automation Outlook 2026," survey of 300 manufacturing professionals, press release, 20 January 2026. https://www.redwood.com/press-releases/manufacturing-ai-and-automation-outlook-2026-98-of-manufacturers-exploring-ai-but-only-20-fully-prepared/
3. Deloitte, "2026 Manufacturing Industry Outlook," section on aftermarket services, which describes margins "more than two times higher than equipment sales alone." https://www.deloitte.com/us/en/insights/industry/manufacturing-industrial-products/manufacturing-industry-outlook.html
4. IDC, Jeffrey Hojlo, "Charting the AI-driven future of manufacturing," 12 November 2025, 2026 manufacturing industry predictions: "By 2027, 40% of all operational data will be integrated across applications and platforms autonomously due to increased standardization and the use of AI agents purpose-built for specific data." https://www.idc.com/resource-center/blog/charting-the-ai-driven-future-of-manufacturing/
5. KPMG, "Global tech report 2026: Industrial manufacturing," 258 industrial manufacturing technology leaders across 22 countries and territories. https://kpmg.com/xx/en/our-insights/ai-and-technology/global-tech-report/industrial-manufacturing.html
6. Matthew Dixon and Ted McKenna, "Stop Losing Sales to Customer Indecision," Harvard Business Review, June 2022, drawn from their book The JOLT Effect: How High Performers Overcome Customer Indecision (Portfolio, 2022); the 56% and 44% split as stated on the authors' site, "What Is the JOLT Effect?" https://hbr.org/2022/06/stop-losing-sales-to-customer-indecision https://www.jolteffect.com/blog/what-is-the-jolt-effect
7. Cisco, "Cisco Research: Industrial AI Moves into Physical Operations, Readiness Gaps Determine Scale," press release, 7 April 2026 (also on Cisco's newsroom from March 2026), State of Industrial AI Report: a double-blind survey of more than 1,000 operational technology decision-makers in 19 countries and 21 industry sectors, conducted with Sapio Research; "40% cite cybersecurity as the biggest obstacle to scaling AI." https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m03/state-of-industrial-ai-report-2026.html https://s21.q4cdn.com/812015656/files/doc_news/Cisco-Research-Industrial-AI-Moves-into-Physical-Operations-Readiness-Gaps-Determine-Scale-2026.pdf
8. MEDDPICC, the sales qualification method whose letters include Metrics, Economic Buyer, Decision Criteria, Decision Process, Paper Process, Identify Pain, Champion and Competition. https://meddpicc.net/meddpicc-sales-definition/

What is not yet sourced

• The Gartner six-to-ten figure (reference 1) is cited through secondary reporting; Gartner's original publication has not been reviewed. The 2025 five-to-sixteen figure is from Gartner's own press release.
• The seats listed in "Count the seats" and what each requires come from Thing Company's engagement experience, not a published study.
• "Many months" between technical interest and a purchase order is engagement experience. No published measure of that gap for industrial purchases has been found, so the paper gives no range.
• How industrial investors now examine commercial evidence, and why enterprise-software forecasts break against industrial cycles, are observations from engagement experience.
• The USD 50,000 subscription and USD 150,000 capital purchase are an illustration, not figures from an engagement.
• The five-dimension interview rubric and the 4.0 threshold are Thing Company methodology, not an external benchmark.
• The JOLT figures (reference 6) come from enterprise sales conversations across industries, cited to the authors' own article and site; the study does not report an industrial-only figure. The Cisco figure (reference 7) is cited to Cisco's release. Cisco's 2024 industrial networking report ranked cybersecurity third among external obstacles to growth and first among internal ones, which is a different question from the 2026 one about scaling AI, so the paper makes no comparison between editions.
• How often industrial buyers approve technology as capital spend, not operating spend, has no published measure we could find; the paper describes it from engagement experience. So do how field service teams are paid and how OEMs and vendors assume data access.
• The claim that deals rarely die at evaluation, and that validating the committee costs a small fraction of a quarter of misdirected sales capacity, are observations from engagement experience, not measured figures.

Thing Company is an independent market validation practice for industrial technology. We test whether a buyer exists at a price that works.

Most of that work is finding the people who sign, and asking them directly. How we work sets out the evidence standard, and the Sprint is the instrument that applies it.

Pipeline full and not converting? Start with a Sprint Assessment.

Not ready to talk? Run the free five buyer questions check self-check from the toolkit.