When you export a 400-contact Affinity base, you don't get a clean CSV. You get a flattened shadow of your relationship graph. A contact linked to six deals, three service partners, and two referral sources becomes three separate rows—each missing the others. Sixty percent of your relational depth evaporates because Affinity's many-to-many fields don't map to CSV rows: they collapse into single entries or vanish entirely. Your deal pipelines stay intact, but the why —the network of who knows whom, which referral partner introduced the deal, which account executive owns the relationship—gets orphaned. The cost is not just lost context. It's decision blindness. You can't see which contacts are actually hubs in your network. You can't trace deal lineage. You can't answer "which three people should be on this renewal conversation?" because you've lost the relationship thread. And if you're moving to a new CRM, you're building blind, reconnecting dots manually, or worse—leaving them disconnected and accepting a 20–30% drop in deal velocity for the first quarter. Here's the 9-step playbook to audit what you're about to lose, map it systematically, and rebuild it in a CRM that actually holds relational depth without collapse. Step 1: Inventory Your Relationship Field Types in Affinity Before you export, document what's at risk. Open Affinity's list configuration and list every field that stores relationships: Many-to-many links (Organization → Organizations, Contact → Contacts, Deal → Contacts) Organization hierarchy (parent/subsidiary, partner/vendor) Deal influence fields (decision-maker, influencer, champion, coach, economic buyer) Relationship type metadata (warm intro, cold, referral, inbound, existing customer) Cross-functional ownership (executive sponsor, technical lead, procurement owner) Count how many records have 3+ relationships. In a 400-contact base with active relationship mapping, you'll typically find: ~180 contacts linked to 2+ other contacts ~120 contacts bridging 3+ deals ~60 contacts with 4+ influence roles across different deals These are your highest-risk rows. They'll export as either (a) multiple copies with partial data, (b) a single row with only the "primary" relationship, or (c) completely missing link metadata. Step 2: Run a Dry Export and Map the Collapse Points Export a sample of 50 contacts that you know have multiple relationships. Open the CSV in a spreadsheet and spot the damage: Duplicated contacts. If Jane is linked to Deals A, B, and C, does she appear three times or once? Lost metadata. If Jane is the "economic buyer" on Deal A but an "influencer" on Deal B, does the export store both roles or just one? Broken organization links. If Jane works at Acme and Acme has a parent company, does the CSV include the parent relationship? Missing junction data. If Jane referred Tom, and Tom referred Sarah, does the export capture both the referral link and the chain, or only Jane → Tom? Create a simple table: Relationship Type Records Affected Export Behavior Recovery Method Contact-to-contact links ~120 Duplicated rows OR single row, link type lost Rebuild from de-duped CSV + manual reconciliation Deal influence roles ~200 Single role per row (if exported at all) Audit deal records separately; add custom contact field per role Referral chain ~85 Lost entirely or flat Manual source-of-truth audit; rebuild in CRM relationship fields Organization hierarchy ~40 Parent link only, no reverse lookup Re-link subsidiaries to parent; validate in new CRM Step 3: Audit Deal Ownership and Stakeholder Roles Export your Deals list with all influence fields visible. For each deal, you should have clarity on: Primary contact (deal owner) : Who is the main point of contact? Economic buyer : Who controls the budget? Technical decision-maker : Who evaluates the solution? Champion / coach : Who advocates for your solution internally? Executive sponsor : Who has political weight to unblock? Export these as a separate CSV: deal ID, deal name, primary contact name/ID, and each role with the contact name or ID. This is your source of truth for relationship recovery. You'll use it to rebuild influence mappings in your new CRM without guessing. Count how many deals have only one contact listed (no secondary stakeholders). If it's more than 40%, you've either lost stakeholder data on export or your Affinity setup was incomplete—either way, you'll need to manually audit your email threads and Slack history to recover the missing names. Step 4: Run a Lost Context Audit For your 50 highest-value deals (or top 10% by ACV), manually review: Affinity's relationship graph visualization (if available) for each deal The email thread history to identify who actually has influence Any Slack or email mentions of stakeholders not yet tagged in Affinity Create a de-duped list of "stakeholders we almost lost." These are contacts who appear in deal emails or Slack but aren't formalized in Affinity's influence fields. Flag them