You're moving from Pipedrive to HubSpot or Salesforce. The export looks clean on day one, but three weeks in, your forecast is $300K off. Deals that were marked 60% probability import as 0%. Shared deals appear as duplicates. Custom fields collapse into plain text. Your sales manager rebuilds the entire pipeline by hand. This is not unusual. It's the norm. And it's preventable—if you know what to audit before you leave. The core problem: Pipedrive's deal structure, probability logic, and field mapping don't translate cleanly to HubSpot's stages or Salesforce's record types. A 30-minute manual export doesn't expose these gaps. By the time your team discovers them, you've lost weeks of forecast visibility and deal context. Here's what actually breaks, why, and how to catch it before migration day. Shared deals: the forecast multiplier that vanishes In Pipedrive, a single deal can be owned by multiple people. Each rep sees it in their pipeline. Each rep's forecast includes the deal at full value. It's collaborative—and it's a forecast time bomb. When you export and import that deal to HubSpot, the system force-assigns it to one owner. The duplicate you created for the second rep is now a separate record. You have two deals of the same value in the pipeline—one real, one ghost. Or Salesforce imports it under the deal's primary contact, stripping the secondary owner entirely. Your AE thinks the deal is gone. What to audit: Pull a Pipedrive export and count deals with multiple participants listed in the Participant field. For each shared deal, document the owners, their expected forecast contribution, and whether the deal is split or full-value for each rep. Test your target system's import logic: does HubSpot/Salesforce create one record or many? Who gets ownership? What happens to the secondary rep's visibility? If both systems claim the deal, your forecast just inflated by 40–60%. The fix: Before migration, decide on deal ownership rules. One owner per deal is simplest but kills collaboration. Alternatively, use a role-based assignment: Account Owner + Closer, with split forecast weighting. Codify this in your target system before you import a single deal. Probability scoring: custom logic that doesn't survive the export Pipedrive lets you set deal probability as a fixed percentage or tie it to the pipeline stage. Many teams create custom rules: if the deal has a signed contract, auto-set to 80%; if two meetings are logged, jump to 50%. When you export, the probability field exports as a simple number. The logic dies. The rules don't follow. Worse: if your target system uses stage-based probability (HubSpot does this natively), Pipedrive's freeform percentages often conflict with the new stage map. A deal in Pipedrive's "Qualified" stage was 45% probable in your old system; HubSpot's "Qualified" stage defaults to 20%. Your forecast just dropped $200K on import. What to audit: In Pipedrive, extract the deal probability distribution by stage. What percentage of deals in each stage have custom probability overrides? Document your current probability logic: which stages map to which percentages? Are there rule-based exceptions? Pull a sample of 20–30 deals across all stages. Export them and re-import them into your target system. Compare probability before and after. Calculate the forecast delta: if you re-imported your entire pipeline today, how much would forecast change? The fix: Map your Pipedrive stage-probability pairs to your target system's stage model before import. If they don't align, choose: recalibrate stages in the source system before export, or accept a one-time forecast recalculation in the target system and document why. Do not let probability drift silently. Custom fields: the collapsed data trap You've built custom fields in Pipedrive: decision-maker confirmed (yes/no), budget range (dropdown), solution fit score (number), competitor mentioned (text). They live in your deal record and drive your forecast logic. When you export to CSV and import to HubSpot, custom fields often map incorrectly or vanish. A Pipedrive checkbox becomes a text field. Your dropdown loses its options and imports as a string. A number field becomes text. Formulas break. Even when fields map correctly, the values don't. Your "Competitor Mentioned" field had a consistent set of values (Salesforce, Zoho, HubSpot, etc.). The export includes typos ("Salesfodce", "salesforce ", "SALESFORCE"). The import sees three versions of the same value. What to audit: In Pipedrive, list every custom field on your deal object. For each field, document: field type (text, number, dropdown, checkbox), whether it's used in workflows or reports, and how critical it is to forecast accuracy. Map each Pipedrive field to its target system equivalent. Flag fields with no direct match. Export 50 deals and spot-check custom field values in the CSV. Look for inconsistency, whitespace, capitalization, and out-of-range values. Re-import those 50 deals i