You've built a relationship graph in Affinity that represents years of deal work: who introduced whom, which companies connect to which contacts, what conversations happened and when. The moment you hit export, 40% of that depth vanishes. Deal participation records flatten. Interaction logs lose their temporal cardinality. Company-to-company connections evaporate. Most teams don't realize what they've lost until they're three months into their new platform and chasing emails to reconstruct who-knows-whom. What Affinity's export actually removes Affinity stores relationship data in three interconnected layers. When you export, only the top layer survives as static records. Deal participation and routing context Affinity tracks which contacts participated in which deals, the sequence of their involvement, and the strength of each connection (decision-maker, introducer, champion). Your export gives you a CSV with contact name, company, and deal title. You lose: Who introduced whom in the deal process Whether a contact was decision-maker, influencer, or implementer The sequence in which they joined the deal How many deals they shared with each other contact This matters because deal routing in your new platform will default to first-contact-owns-it or deal-team-split logic, not relationship-based routing. A contact who has introduced five successful deals for you will appear identical to one who bought once. Interaction log density and timing Affinity's strength is its interaction timeline: emails, calls, meetings, and other touchpoints, timestamped and threaded to specific contacts and deals. Your export captures the raw count. It loses: Email thread clustering (which emails belong to the same conversation thread) Call or meeting duration metadata Sentiment or outcome tagging (if you tagged interactions as positive, negative, stalled, or warm) Multi-party interaction context (who else was on the call, who was CC'd) A contact with 47 email interactions in your export might look equally engaged as one with 47 interactions across seven years, with no clustering to show whether those interactions were sustained or sporadic. Company-to-company relationship mapping Affinity allows you to map relationships between companies—partner networks, acquisition targets, referral sources. These are rarely exportable as structured data. You'll export contacts and companies as separate tables, and the company-level edges disappear entirely. You lose context like: Which companies have referred business to you Which companies are strategic partners or competitors to each other Which contacts are known to hop between related companies Measuring the actual damage: 60-day audit framework Before migration, run a forensic export and measure what you're about to leave behind. This audit will inform your rebuild priority. Step 1: Export and count by cardinality Export your contact, company, and deal tables. Count unique values per row, then compare against your live Affinity view. Contacts per deal: In Affinity, pull a mid-size deal (3–8 contacts). Count how many are listed in the deal view. Export the same deal and count how many contacts appear in your export linked to that deal. The difference is your cardinality loss. Interactions per contact: Pick a warm contact (50+ interactions). In Affinity, pull the full timeline. Export the same contact. Count interaction records in both. The loss percentage here—usually 5–15%—reflects your email sync or meeting log incompleteness, but also tagging loss. Deal overlap per contact: In Affinity, find a contact in 4+ deals. Export and count how many deal linkages survive for that contact in the export. Most exports give you contact ID, deal ID, role—but lose the semantic weight of the role and the introduction chain. Step 2: Sample three use cases and map what you lose Take three real deals and walk through the export to see what a salesperson actually loses: Deal 1: A warm champion at a prospect company who has also referred two other deals. In Affinity, you see the referral chain. In your export: one contact, three deals, no relationship semantic. Can you tell which deal they referred vs. participated in? Probably not without hunting emails. Deal 2: A deal with five participating contacts from two companies, one of which is a warm partner. In Affinity, the partner relationship and the contact roles are visible. In your export: five contacts, two companies, deal title. No partner link, no role hierarchy. Deal 3: A contact with 120 interactions over five years, who you last spoke to 18 months ago. In Affinity, the timeline shows the density cliff. In your export: interaction count, maybe interaction dates. No pattern, no warmth signal. The 40% loss statistic reflects missing roles, interaction clustering, company-level relationships, and temporal density metadata—not missing rows. Your contact count will be accurate. Your relationship understanding will be hollow. Rebuild strategy: 60-day phased a