Pipedrive exports cleanly until they don't. You'll get your contacts, their email addresses, your open deals, and a neat CSV of activity timestamps. Then you'll log into your new CRM and realize your forecast has aged five years in five minutes. The problem isn't Pipedrive's export—it's what gets lost in translation. Deal stages collapse into free text. Custom fields either vanish or import as empty columns. The probabilistic logic that told you 'this deal closes 60% of the time in this stage' evaporates. Your forecast accuracy regresses 15–30 percentage points in month one, and rebuilding it takes longer than you budgeted. This guide maps what actually survives a Pipedrive exit, where the gaps appear, and how to recover forecast accuracy inside 30 days. What moves cleanly from Pipedrive (and why) Pipedrive's CSV export is well-structured. These elements port without loss: Contacts and companies: Names, email addresses, phone numbers, postal addresses. Zero friction. A contact is a contact. Activities (calls, emails, notes): Timestamps, type (email/call/meeting), duration, and free-text notes all export intact. The history is there; the context stays. Open deals (with caveats): Deal name, owner, value, expected close date, and current stage export as columns. But the stage label is only metadata; its meaning —the win probability baked into your forecast—is not. Deal products and line items: If you use Pipedrive's products module, those export too, albeit as a secondary table that requires remapping in most target platforms. Basic deal custom fields (text, dropdowns, numbers): Simple custom fields export as columns. The data is there—the field definition might not be. The reason these move so well: they're flat, non-relational, and don't encode business logic. A deal value of $50k is a deal value of $50k in any CRM. An email sent on March 15 was sent then, regardless of platform. What breaks during migration (the forecast reckoning) This is where forecast accuracy dies. Pipedrive stores deal stage probability in the stage definition, not in the deal record itself. When you export, you get the stage name —'Discovery', 'Proposal', 'Negotiation'—but not the 25%, 50%, 75% probabilities you configured against each stage. Your new CRM doesn't know that your 'Proposal' stage is worth 65% confidence; it has to ask you again. Deal stage probability mapping is manual: You have 47 deals in 'Proposal'. Your old Pipedrive said that stage is 65% likely to close. Your new CRM (HubSpot, Orin, Zoho, Salesforce) sees them as 0% confident until you reconfigure the stage and reassign probability. If you skip this step, your forecast shows $500k in deals with no weight—it's phantom revenue. Custom field loss or misalignment: Pipedrive's custom fields export, but most platforms don't auto-map them on import. A field called 'Decision Timeline' in Pipedrive arrives as 'Decision Timeline (1)' or gets lost entirely if the target platform enforces field types differently. Checkboxes become text. Date pickers become strings. Deal stage workflow and routing breaks: Pipedrive stages are linear (you move down the funnel). If you've built logic that says 'only sales managers can move deals out of Negotiation' or 'deals older than 60 days in Discovery auto-escalate', that doesn't export. You're rebuilding permissions and automation from scratch. Lost deal history and stage transition data: Pipedrive doesn't export the history of a deal—which stages it passed through, when, who moved it, how long it spent in each stage. You see the current state only. If you need to analyze 'deals close in 45 days on average', you can't—you only have today's snapshot. Forecast logic is binary after migration: Pipedrive's forecast sums deal value by stage, multiplied by stage probability. When you export and re-import, you often lose that multi-stage logic. Some platforms (like Orin) rebuild it in setup; others (like Salesforce) require you to hand-code custom probability fields. The forecast becomes a dumb sum until you fix it. Data survival by destination: HubSpot vs. Orin vs. Zoho vs. Salesforce The destination matters. Not all CRMs rebuild forecast accuracy the same way. HubSpot HubSpot's strength: the import wizard is forgiving. Contacts, companies, deals, and activities import with minimal friction. You get a visual mapping step—pick your Pipedrive 'Company' column and match it to HubSpot's company field. Done. HubSpot's gap: forecast logic is user-configurable but not auto-restored. You'll import deal stages (labeled exactly as they are in Pipedrive), but you must manually assign probability to each stage. If you had ten custom fields in Pipedrive, HubSpot will attempt to match them to its standard fields (if your 'Decision Timeline' looks like a date, HubSpot guesses 'Close Date'). This saves time but creates misalignment—you'll spend week two auditing and remapping. Forecast accuracy hit: 20–25% regression in month one. HubSpot recovers it faster tha