Every vendor below appears in published DataKitchen material because a real commercial platform used it or a real launch team evaluated it. We don’t score them. We don’t have their pricing either.
The map
| Layer | Names that appear in real platforms | Source |
|---|---|---|
| Syndicated prescription and prescriber data | IQVIA (Xponent), ICON/Symphony Health | Launch case study, the data most likely to be wrong |
| Claims | Komodo Health and Fingertip, explored by one launch team | Launch case study |
| Formulary and market access | MMIT, explored by the same launch team | Launch case study |
| Specialty pharmacy and distribution | The specialty pharmacy network; 3PL; EDI 852, 844, and 867 | Launch case study, data sources |
| Patient services | Hub, copay, and REMS program data | Data sources |
| CRM and field | Veeva CRM; Veeva Vault for medical communications | Rare-disease biotech story |
| Marketing and patient engagement | Salesforce Marketing Cloud and Health Cloud at the biotech; Constant Contact, HubSpot, and Salesforce Marketing Cloud weighed for CRM by one launch team | Biotech story, launch case study |
| Prescriber master | Veeva Network, IQVIA OneKey, Reltio | Launch case study, platform page |
| Call planning and sales force design | Beghou, ZS, KMK, as contenders | Launch case study |
| Content and events | Hybrid Health, CVENT, as contenders | Launch case study |
| Finance | NetSuite | Biotech story |
| Warehouse | Snowflake, Databricks; Redshift at Celgene | Platform page, Celgene launch |
| BI and AI on top | Tableau, Power BI; Snowflake Cortex Analyst, Databricks Genie | Platform page |
| Testing, orchestration, monitoring | DataOps TestGen, DataOps Automation, DataOps Observability | Platform page |
The data sources reference breaks the feed layers into 37 named feeds and 10 identifiers, with how each one arrives and how it breaks.
What a small team actually buys
The list above is a union, not a shopping list. You will not need all of it.
The rare-disease biotech in our Cadillac and Corolla story had a commercial team of about five and about 20 datasets across 13 source systems. Its specialty pharmacy sent four files a day: patient status, dispense, inventory, and demographics. It reached prescribers through HCP websites, targeted display, and third-party medical email, not a field force. And it ran with no claims, no IQVIA or Symphony, and no formulary feed, because a handful of patients per prescriber made a claims universe less useful than knowing each patient’s status.
A launch team building for a broad brand did the opposite. It started with over a decade of claims data, then tried IQVIA and Symphony Health for NRx and TRx, explored Komodo Health and Fingertip for claims and MMIT for formulary and market access, and added sales force alignment, zip-to-territory mappings, and EDI data. That is in the launch case study, along with the systems it weighed for MDM, CRM, and call planning.
So buy by question. The launch timeline lists the 24 reports a launch analytics team owns and the data each one joins.
Where the platforms sit
On DataKitchen’s commercial data platform, the warehouse, the raw landing zone, the analyst-ready stars, TestGen, and the Git repository all live in your cloud account. DataOps Automation and Observability run in DataKitchen’s cloud and drive your environment without holding your data. Your BI tool, Veeva Network, and Jira and Confluence stay where they are.
That split is the one to settle first in a security review. It is also why a later handover is a normal day rather than a migration project.
Where the vendors meet
Every layer joins on a handful of identifiers: NPI, DEA, NDC, ICD, HCPCS, and each vendor’s own provider ID. Vendor provider IDs are stable within a vendor and meaningless across vendors, so the prescriber master has to link on several keys at once. Those joins are where the expensive errors live, because a malformed identifier does not error; it fails the join and the row disappears. See finding duplicate HCP records and why generic tools miss pharma identifiers.