# How B2B sales teams make choices in complex enterprise selling China B2B Organizational Sales Simulation Whitepaper Collaborations open · Whitepaper not yet published Canonical: https://www.longarena.ai/research/b2b-sales-whitepaper Updated: 2026-10-09 Frozen source: b2b-whitebook-78-v1.1 Not survey answers, but the steps teams actually take when facing a limited budget, a full customer decision network and unexpected change. ## Evidence - 78: training classrooms - 361: active teams - 1,776: classroom attendances - 19,733: approved sales actions - 1,281: AI customer stakeholders Real enterprise training classrooms from 2024–2026. Attendances are not unique people; action counts refer to approved, effective sales actions. On the SalesNail simulation platform, teams represent a selling company and interact with AI customer stakeholders with distinct roles, positions and needs. They choose whom to contact, when to propose a solution and which purchasing opportunity deserves their limited budget. Across three rounds, records capture each action’s sequence, target, cost and approval. ## Questions ### Customer decision networks How do teams find, prioritize and engage key stakeholders? Do differences emerge in sequence, depth or resource allocation? ### Exploration and focus When do teams shift from broad exploration to selected opportunities? How does resource concentration change across rounds? ### Proposal timing Before a first proposal, has the team engaged the relevant key decision-maker? How are discovery and prescription connected? ### Coordination and adaptation How do teams with shared decisions and a designated operator act? How do they adapt to designed shocks such as leadership changes? ## Boundaries - A convenience sample of training classrooms, not a nationally representative survey. Simulation behavior is not real-world sales performance. - Message bodies and win/loss outcomes are not currently available. Roughly one in ten actions lacks a cost value, which remains unknown. - This page and the invitations contain aggregate information only. Company identities, participant identities and raw classroom records are not public. - Public strata require at least 10 verified organizations and 30 teams. No individual, team or company rankings. - Individual-level research data is not available for public download. Access requires a data-use, anonymization and publication-review agreement. ## Collaboration ### Faculty & independent researchers Bring a question from selling, organizational behavior, management, education or process mining. Align on one or two questions and conduct reproducible analysis. Academic papers and the business whitepaper proceed in parallel; authorship follows academic contribution. The first collaboration window closes November 30, 2026. Later collaborators remain welcome. Early findings may support subsequent research-funding preparation. English invitation: https://www.longarena.ai/research-assets/20261009/LongArena-Academic-Research-Collaboration-Invitation-20261009-EN.pdf ### Training institutions & participating enterprises Verify the hosting relationship, participant organization type and training context for your classrooms. Verification concerns facts; it gives no institution a veto over research conclusions. Verification completed by mid-December 2026 enters the first edition; later verification enters subsequent editions. Anonymous aggregate comparisons follow publication, not rankings. English invitation: https://www.longarena.ai/research-assets/20261009/LongArena-Classroom-Verification-Collaboration-Invitation-20261009-EN.pdf ## Timeline - October–November 2026: Research conversations, question alignment and classroom verification. - December 2026: Begin descriptive analysis; mid-December is the first-edition verification deadline. - Q1 2027 (planned): Proceed toward publication subject to statistical and privacy review. Strata below the threshold will not be published. - Later editions: Verification remains open. New annual samples are frozen separately and do not overwrite the existing source. Contact: bill.bai@long-arena.com This is a project overview, not a published final whitepaper or a public raw dataset.