The State of B2B Business Cases 2026

Why software purchases stall, what buying committees actually require, and how value engineering must adapt

A research synthesis published by ValueNova. April 2026.

Methodology note

This report is a synthesis of publicly available secondary research. ValueNova commissioned and produced the analysis to translate analyst-grade evidence from Gartner, Forrester, McKinsey, the European Spreadsheet Risks Interest Group, and other authoritative sources into operator-grade insight for B2B sales teams.

Research period reviewed: 2022–2026, with priority given to research published 2024 or later. All statistics cited are traceable to the source listed in the bibliography. No claim appears in this report without a verifiable third-party source. Where sources disagree or contradict, that disagreement is named explicitly in the Counter-evidence considered section.

ValueNova's editorial standard for this report is methodological neutrality. The product is mentioned only in the methodology note and the brief "about" paragraph at the end. The goal is to publish an artefact that is useful, citable, and defensible regardless of whether the reader has heard of ValueNova before.

Why this report exists

The research base on B2B buying has matured rapidly. Forrester's annual State of Business Buying survey now reaches more than 16,000 global buyers. Gartner's Future of Sales and The New B2B Buying Journey series have produced a decade of structural data on how purchasing decisions actually get made. Spreadsheet-risk research from Raymond Panko and EuSpRIG has, for over twenty years, quantified the reliability problems with the financial models most B2B sellers still rely on.

The gap is translation. Most operators on the front line of B2B sales — value engineers, RevOps leaders, sales directors, pre-sales teams — do not read 60-page analyst reports. They build business cases under deadline, often in spreadsheets, often without a clear methodology, and often hoping the model holds up when finance gets involved. By then, it is usually too late.

This report consolidates the most reliable findings from across that research base, organizes them through a single lens — why business cases succeed or fail in 2026 — and translates the implications into changes B2B sales teams can act on this quarter.

The buying group has gotten bigger and more conflicted

The single most important structural change in B2B buying over the past decade is the growth of the buying committee. Forrester's December 2024 State of Business Buying report, drawn from a survey of more than 16,000 global business buyers, finds that the typical B2B purchasing decision now involves 13 people inside the buying organization. 89% of purchases involve two or more departments.

Earlier research from Gartner placed the typical buying group at 6–10 stakeholders for complex B2B solutions. The trend across studies is unambiguous: the number is going up, not down. CEB research from 2015 (now part of Gartner) put the figure at 5.4 people. By 2017, Gartner's revised estimate was 6.8. Forrester's 2024 figure of 13 represents a near-doubling in less than a decade.

The mechanics of a larger committee matter more than the headcount itself. Each stakeholder brings their own evaluation criteria, their own success metrics, and their own accountability concerns. CFOs care about payback periods and assumption defensibility. CIOs care about technical fit and security posture. Procurement cares about commercial terms and risk transfer. End users care about adoption friction. Champions care about whether the deal will get them promoted or fired.

The result is structural conflict. SBI's 2024 research found that 74% of buyers faced too many competing options and paths in their last major purchase decision, and 70% said they worked with so many people on the supplier side that they were unsure who everyone was. High-friction buying environments — those with unresolved internal conflict, unclear evaluation criteria, or unaligned stakeholders — reduce the odds of a purchase by 43%.

This is the single most consequential finding for sales teams in 2026. Deals are not lost because the product fails the evaluation. They are lost because the buying committee cannot reach consensus on whether the product is worth the risk.

Buyers are doing more research, talking to sellers less

The second structural shift is that buying decisions are increasingly made without the seller in the room. Gartner's research finds that B2B buyers spend only 17% of their decision time in direct contact with potential suppliers. When the same buyer is comparing multiple vendors, the share of time allocated to any one supplier can drop to 5–6%.

The remaining 83% of decision time is spent on independent research, internal stakeholder alignment, and consensus building. Buyers read review sites, talk to peers in private Slack channels and communities, run their own ROI estimates, and form preliminary preferences before any seller is invited to a meeting. By the time a seller engages, much of the evaluation has already happened — often based on evidence the seller has never seen and cannot influence.

The preference is structural, not transient. Gartner research finds 67% of B2B buyers now prefer a rep-free experience, with the figure climbing to 44% specifically for millennial buyers. Forrester's 2024 work shows nearly 95% of buyers expect to use generative AI tools to inform their buying process within the next year. The direction of travel is unambiguous: buyers will continue to do more on their own, with better tools, before seller contact.

The consequence is that the evidence base buyers form their preferences from must exist independently of the sales team's ability to deliver it. Business cases that rely on a polished sales presentation to land are not business cases. They are pitches. The artefact that survives is one the buyer can build, defend, and circulate internally without needing the seller in the room.

Procurement and finance are now decision-makers, not gatekeepers

The third shift is the changed role of procurement and finance. Deloitte's 2025 Global Chief Procurement Officer Survey shows procurement teams are now significantly more strategic and digitally enabled than five years ago, with CPOs reporting greater involvement in earlier stages of the buying cycle and more influence over final supplier selection.

Forrester's 2024 work confirms the direction: procurement now exerts decisive influence in 53% of B2B technology purchases, up from a roughly 30–35% baseline a decade ago. This is not just a procurement-led negotiation at the end. It is procurement involvement throughout the cycle, including in the formation of evaluation criteria, the construction of the business case, and the calibration of risk.

Finance involvement has grown in parallel. McKinsey's Recalibrating Technology Budgets for the AI Era describes a sustained tightening of CFO scrutiny on enterprise technology spending, particularly in the post-2022 period when ZIRP-era growth assumptions collapsed. Forrester's research shows 87% of technology buyers reported adjusting their buying process to ensure they only bought mission-critical products. Translation: every business case now competes against a higher hurdle rate, and against finance teams more willing to say no.

The combined effect is that the modern B2B software business case is no longer a sales tool. It is a finance and procurement deliverable that the sales team is asked to assist with.

The CFO veto pattern

CFOs reject business cases for a small number of recurring reasons. The most common is that the assumptions cannot be defended. Forrester's Total Economic Impact methodology, which has framed analyst-grade ROI work for over twenty years, treats four components as table stakes for a credible business case: quantified benefits, quantified costs, flexibility (the value of optionality), and risk-adjusted returns.

Most vendor-built ROI calculators address one of these four. Some address two. Almost none address all four with traceable, source-cited assumptions. When a CFO asks where the 34% productivity uplift comes from, the answer is usually a slide that says "industry benchmark" or, worse, "based on customer testimonials." Neither survives a finance review. The Phillips ROI Methodology — the conservative-evaluation standard developed by Jack Phillips and the ROI Institute — makes this explicit: claimed benefits should be reduced by a confidence factor reflecting source quality, and the burden of proof sits with the proposer.

The second failure mode is the absence of sensitivity analysis. A point-estimate ROI of 287% is a marketing artefact. A range estimate that shows what happens to ROI under conservative, base case, and stretch assumptions, with the underlying drivers clearly labelled, is a finance artefact. The HM Treasury Green Book — the UK government's reference text on business case methodology and the basis of the Five Case Model used across British public-sector procurement — treats sensitivity analysis as non-negotiable. Most vendor-supplied business cases do not include one.

The third failure mode is the disconnect between the business case and post-purchase value tracking. The Project Management Institute's Business Case Best Practices for Success makes the case that a business case is not an approval document but an ongoing benefits-realization commitment. If the case promises a 24-month payback, the buying organization expects to be able to verify that payback after 24 months. When sellers cannot describe how realization will be measured, finance treats the original case as aspirational rather than evidentiary.

CFOs do not reject business cases because they are anti-purchase. They reject business cases because the cases do not survive standard finance scrutiny. The frameworks that exist to make cases defensible — TEI, Five Case Model, Phillips ROI, PMI's benefits-realization standards — have been public, free, and well-documented for years. The gap is execution, not knowledge.

Stakeholder requirements that contradict each other

The 13-stakeholder buying committee has a structural problem: each stakeholder evaluates the same purchase against a different set of criteria, and those criteria frequently conflict.

A procurement officer wants a multi-year commercial commitment with aggressive volume discounts and clear penalty clauses for non-performance. A CIO wants a flexible deployment model with the option to scale down. A line-of-business sponsor wants speed-to-value and minimum disruption. A CFO wants conservative assumptions and downside-bounded risk. An end user wants minimal change to their daily workflow.

These cannot all be optimized simultaneously. The vendor's job in a 13-stakeholder cycle is not to win each sub-evaluation. It is to help the buying committee navigate the trade-offs in a way that produces consensus. Forrester's research is explicit: the providers that win in modern B2B buying are those that help buyers make decisions, not those that try to sell harder.

The practical consequence is that stakeholder-tailored business cases must be internally consistent. A CFO version of the case that emphasizes conservative ROI cannot contradict the line-of-business version that emphasizes growth upside. The numbers must reconcile. The assumptions must be the same. The methodology must be the same. What changes between stakeholder views is the framing, not the underlying model.

This is exactly the operational problem that ad-hoc spreadsheet-based business cases are worst at solving. Different reps build different models for different stakeholders. The numbers don't reconcile because no one is enforcing methodological consistency. By the time the buying committee compares notes, the contradictions are visible — and the deal stalls.

The spreadsheet governance problem

The reliability problem with spreadsheet-based business cases is not a hypothesis. It is a 25-year research literature with consistent findings.

Raymond Panko's spreadsheet-error research, summarized across multiple peer-reviewed publications and the EuSpRIG conference proceedings, establishes three durable findings:

First, spreadsheet developers are 95–99% accurate at the cell level, which sounds high. But large operational spreadsheets contain hundreds or thousands of formula cells. With a 1–6% cell error rate, the probability that a large spreadsheet contains at least one material error approaches certainty.

Second, recent field audits — those conducted with rigorous methodology rather than self-reporting — found errors in at least 86% of operational spreadsheets examined. Most of those errors were not minor. The audits typically reported only substantive errors.

Third, error detection is much weaker than error commission. Individuals working alone find about 63% of errors during code inspection. Teams of three find 83%, but only with structured inspection methodologies that almost no commercial sales team uses.

These findings have direct implications for vendor-supplied business cases. The typical sales-built ROI model is constructed under deadline by a single rep, often without peer review, often using formulas copy-pasted from older models with their original errors intact. The probability that any given vendor-supplied business case contains at least one material error is, on the EuSpRIG evidence, very high.

CFOs know this. ICAEW guidance, professional financial-modelling standards from the FAST Standard Organisation, and the broader F1F9 modelling community have all converged on a clear position over the past decade: spreadsheet models that drive material business decisions need to be governed, version-controlled, peer-reviewed, and built to a documented standard. Most sales-supplied business cases meet none of these criteria.

The consequence is that finance teams treat vendor-supplied ROI models as opening positions to be discounted, not as evidence to be relied on. Phillips ROI Methodology recommends reducing claimed benefits by a confidence factor that reflects the rigor of the underlying model. For a typical vendor-supplied case, that factor can cut claimed ROI in half or more. When the discounted ROI no longer clears the finance hurdle rate, the deal stalls.

The hidden cost of inconsistent ROI claims

There is a second-order problem that compounds the first. When the same vendor's ROI methodology produces different numbers in different deals — because different reps use different spreadsheets with different assumptions — buyers eventually notice.

A CFO who sees one vendor claim 287% ROI for a $200K deal and another vendor claim 156% ROI for a similar deal in similar circumstances does not conclude that vendor A is better. They conclude that one or both numbers is unreliable. Both deals get treated with skepticism.

This is what value engineering done at scale, with governance, is designed to prevent. The methodology is encoded once. The drivers, benchmarks, and calculation logic are consistent across deals. The version a buyer receives in one deal is methodologically identical to the version a different buyer receives in a different deal. The numbers may differ — different customers have different value drivers — but the framework that produced them is the same.

Methodological rigor: TEI, Five Case Model, Phillips ROI

The frameworks that produce defensible business cases share a structural pattern. They all decompose the case into separable components, each of which can be evaluated, challenged, and updated independently.

Forrester's Total Economic Impact methodology is the most widely used framework in B2B technology, having been applied in hundreds of vendor-commissioned studies over two decades. Its discipline is in the four-factor decomposition: quantified benefits, quantified costs, flexibility (option value), and risk (uncertainty discounting). Each factor is sourced separately, defended separately, and updated separately.

The Five Case Model, codified in HM Treasury's Green Book and used across UK public-sector procurement, decomposes the case into five separate dimensions: strategic, economic, commercial, financial, and management. The strategic case asks whether the investment fits the broader plan. The economic case asks whether the option chosen represents the best value for money. The commercial case asks whether the procurement arrangement is sound. The financial case asks whether the investment is affordable. The management case asks whether delivery is achievable.

The Phillips ROI Methodology brings a fifth element: post-investment evaluation. Phillips' approach is built around five evaluation levels — reaction, learning, application, business impact, and ROI — each more rigorous than the last, and each requiring different evidence. The discipline is in matching the evidence requirement to the claim being made.

What unites these frameworks, beyond their analytical structure, is that they are public, free, and well-documented. There is no proprietary insight that distinguishes a good business case from a bad one. There is only execution discipline.

Explainability as the new floor

A separate strand of research is becoming increasingly important to B2B business case work: AI explainability and model governance.

The US National Institute of Standards and Technology's AI Risk Management Framework 1.0 identifies seven characteristics of trustworthy AI systems: validity and reliability, safety, security and resilience, accountability and transparency, explainability and interpretability, privacy enhancement, and fairness with harmful bias managed. NIST's Four Principles of Explainable Artificial Intelligence goes further: an explainable AI system must produce explanations that are meaningful to the audience, accurate to the underlying decision logic, and bounded by the system's actual knowledge limits.

This matters for business case work because generative AI is now widely used to produce ROI claims, customer-impact narratives, and value calculations. Business cases produced by AI systems that cannot explain their reasoning, source their assumptions, or bound their confidence are increasingly visible as such — and increasingly rejected on sight by finance teams that have read the same NIST guidance.

The category implication is that "AI for business cases" is not the differentiated product. "Explainable AI for business cases" is. This is a distinction the market is just beginning to make, and it will sharpen quickly.

The AI authenticity crisis

The third structural change in 2026 is that buyers can now detect AI-generated content with reasonable accuracy, and they have started to discount it. TrustRadius's 2025 buyer research shows that traditional analyst reports — long considered a credibility floor for vendor claims — are now used by only 14% of buyers, a 60% decrease since 2022. Buyers increasingly trust their own experience, hands-on free trials, product demos, and reviews from real users.

The implication for business case work is that AI-generated ROI claims, AI-written executive summaries, and AI-produced value narratives are losing credibility faster than vendors are adopting them. The competitive moat is not "we use AI to build business cases faster." It is "we use AI to build business cases that are more rigorous than what humans alone produce, and we can show why."

That requires governance. It requires methodology. It requires a system that pre-commits to assumption sourcing, calculation traceability, and confidence bounding before AI is applied to the work. AI is a force multiplier on a governed methodology. It is a credibility destroyer on an ungoverned one.

Post-sale value proof loops

The final differentiator is whether the business case connects to post-sale value tracking. PMI's research is clear that benefits realization is a discipline distinct from initial business case construction, but the two are linked: the value claims made in the original case become the measurement targets for post-purchase tracking.

Forrester's 2024 finding that 81% of buyers are dissatisfied with the provider they ultimately choose is, partly, a measurement problem. Vendors make value claims. Customers buy. Vendors do not measure realization. Customers conclude, eighteen months later, that the value never materialized — even when, in many cases, it has, but no one has tracked it.

Vendors that close the loop — that revisit the original business case at six, twelve, and twenty-four months and document realization against the original claims — turn renewal conversations from defensive (was the value real?) to expansive (here's what we delivered, here's the next phase). Vendors that don't, leave their renewals exposed.

The implications of the research above translate into specific operational changes for B2B sales organizations. Five shifts matter most.

First, the business case is a deliverable, not a conversation. The artefact that survives is the one a buyer can build, defend, and circulate without the seller in the room. That means business cases must be standalone, traceable, and defensible against challenges the seller will not be present to answer.

Second, methodology must be encoded once and reused. The spreadsheet-governance evidence is overwhelming. Every business case built fresh from a blank spreadsheet is an unreliable artefact. The leaders are those who encode their value methodology — drivers, benchmarks, calculation logic, sensitivity ranges — once, and require every business case to start from that foundation.

Third, stakeholder-specific outputs must reconcile to a single underlying model. A CFO version of the case and a CIO version of the case must produce the same numbers from the same assumptions. What differs is framing, not substance. The buying committee will compare notes. Inconsistencies are visible and corrosive.

Fourth, sensitivity analysis is non-negotiable. A point-estimate ROI is a marketing artefact. A range estimate with conservative, base, and stretch scenarios — and clear labelling of which assumption drives which sensitivity — is a finance artefact. Buying committees that cannot see the range cannot defend the recommendation internally.

Fifth, the case must connect to post-sale measurement. A business case that promises a 24-month payback without a measurement plan is a credibility liability, not an asset. Increasingly, finance teams will require the measurement plan as a precondition for approval.