At ₹50M ARR with 100+ sales reps, Pipedrive's infrastructure cracks in three specific places: API response times degrade, custom field logic becomes unmaintainable, and forecasting math diverges from reality. You'll notice it first in your reporting—weekly dashboards that used to refresh in 30 seconds now take 2 minutes, or your pipeline forecast swings 15% unexpectedly when two reps update their stage logic differently. The hard truth: you won't wake up one morning unable to use Pipedrive. Instead, you'll spend six months patching around its edges—custom Zapier workflows, spreadsheet reconciliation, manual forecast adjustments—before admitting the tool no longer fits your operation. That's the worst time to migrate. This playbook maps what data moves cleanly, what to rebuild, and how to plan a 90-day handoff that doesn't orphan deals mid-cycle. Where Pipedrive holds steady (and what to export first) Pipedrive's core data structures—leads, deals, activities, and basic contact fields—export reliably. Its API for these objects is well-documented, and the export format is clean enough that most platforms ingest it without friction. If you're moving to a CRM built for scale , this is your foundation. Leads and basic contact data: Name, email, phone, company, source. Exports cleanly via CSV or API. No data loss. Migration time: hours. Deal records: Title, value, stage, owner, close date, created date. The core deal structure is portable. Caveat: custom deal fields often have logic baked in (see below). Activity logs: Calls, emails, meetings, notes. These are timestamped and unambiguous. Export the full log; it's your audit trail. Pipedrive's API handles 500+ activities per deal without complaint. Basic pipeline structure: Your stage names, stage order, probability weighting. Portable as configuration, not data. Export these first—they're your baseline. Use Pipedrive's native CSV export for leads and deals, not Zapier (Zapier adds 30 seconds per 100 records and costs ₹5K+/month for a data pull you only need once). Allocate 2 weeks: 1 week to identify and document which leads/deals are actually live versus dead, and 1 week to reconcile with your accounting system (what Pipedrive closed at ₹5M might only have invoiced ₹4.2M). What breaks at scale: forecasts, custom fields, and team hierarchies The three failure points that force migration: Forecasting diverges from reality Pipedrive's forecast assumes each rep updates stage consistently and that probability weighting is accurate. By 100 reps, neither assumption holds. You'll see one team using stages to track buyer sentiment, another tracking internal approvals, a third using them as time-based buckets. Pipedrive's single probability-per-stage model can't accommodate this—your forecast becomes a floor wax, shined up but bearing no relation to what actually closes. Additionally, Pipedrive's historical forecast is not queryable—you can't ask "what did we forecast for this pipeline 60 days ago?" without manual inspection. At ₹50M ARR, your finance team needs 60-day forecast drift analysis. You won't get it from Pipedrive. What to rebuild: A forecast model with deal-level custom fields (likelihood %, decision stakeholder count, budget confirmed Y/N) that sits alongside Pipedrive's stage. Use your CRM's native custom field logic or a lightweight BI tool (Looker, Tableau, even structured Sheets) that queries your CRM API daily and stores snapshots. Budget 2 weeks for model design, 1 week for backfilling 90 days of historical data. Custom fields become unmaintainable By ₹50M ARR, you've likely accumulated 40–80 custom fields across leads, deals, and contacts—some actively used, most orphaned. Pipedrive doesn't let you bulk-delete or archive unused fields; it shows all of them in every form, making data entry slower and error rates higher. Worse, custom field dependencies are implicit (a "decision criteria" field that should only appear if "deal type" equals "RFP") require Zapier workflows that break silently. Pipedrive's custom field permissions are also crude—you either allow reps to edit a field or you don't, with no role-based granularity. Your sales leadership team needs to see forecast probability, but reps shouldn't edit it. You'll end up enforcing this through trust and spreadsheet spot-checks, which doesn't scale. What to rebuild: Audit every custom field: which are actively used, which are read-only reference data, which have conditional logic. Delete the orphaned 30%. For fields with conditional logic, move that logic into your new CRM's native field dependencies or a workflow automations layer. If you're moving to Orin , custom fields can be hidden/shown by role, and conditional field logic lives in no-code automations . Budget 3 weeks: 1 week audit, 1 week deletion and consolidation, 1 week rebuilding conditional logic. Team hierarchies collapse under shared deals Sales teams organized by region, vertical, or account size often split responsibility for large d