Enterprise MDM vendors like Informatica and SAP MDG release one, maybe two major updates per year. Planning committees. Feature councils. Staged rollouts. By the time something ships, the workflow that triggered the request has already been patched around with spreadsheets and manual workarounds.
We took a different approach. Since Primentra launched on February 14, we have shipped ten major releases in sixteen days. Not hotfixes. Not patch notes. Full features — staging tables, automated scheduling, approval workflows, cascading lookups, an import wizard — each one live and production-ready the day it ships.
But speed alone is worthless. Shipping fast and shipping garbage is easy. The hard part is shipping fast while only building things that actually improve your day.
Release timeline — February 14 to March 2
16 days10 releases • 50+ features
Every feature starts with a question
Before anything gets built, we ask: does this make your job easier? Not “would this look good on a feature comparison chart.” Not “does our competitor have this.” Does it solve a problem you actually run into, today, in your daily workflow?
If the answer is a clear yes, it gets built. If it is a maybe — if it requires a paragraph to explain why someone might need it — it stays on the shelf.
That filter is the reason we ship fast. No features nobody asked for. No enterprise checkboxes to justify a price point. Build what matters, build it well, move on.
What we shipped — and why
Here is what went live in the past three weeks, and the real-world problem behind each one.
Staging Tablesv0.2026.3.1 The problem: your SSIS package, ETL pipeline, or data warehouse process produces updated reference data. Getting that into your MDM system used to mean manual uploads or fragile direct-SQL scripts with no validation.
What we built: each entity gets a dedicated staging table in the stg schema. External systems write data there. You review it, click process, and the engine validates types, resolves domain codes, and merges into production with configurable rules — overwrite, fill empty, ignore, or sentinel-based NULL handling. Bad rows get flagged with specific error messages. Good rows go live.
Staging Schedulerv0.2026.3.2 The problem: staging tables solved the loading problem. But someone still had to log in, check for new rows, and click “Process.” That is babysitting, not automation.
What we built: three processing modes. Scheduled: run every N minutes, daily at a fixed time, weekly, or monthly. Trigger-based: fire when a row count threshold is reached or when rows have been idle for N minutes. Manual: the existing click-to-process flow. Overlap guards prevent double-processing. Zombie timeout recovery auto-releases batches stuck for over thirty minutes. A full scheduler log tracks every event.
Approval Workflowv0.2026.2.25 The problem: not every data change should go straight to production. Some entities — cost centers, product hierarchies, customer segments — need a second pair of eyes before anything goes live.
What we built: per-entity approval gating. Changes to protected entities get staged for review instead of hitting live data. Approvers receive email notifications, see the full change summary, and approve, return for rework, or reject. Conflict detection catches concurrent edits and duplicate inserts automatically.
Cascading Dropdowns & Derived Columnsv0.2026.2.27 The problem: your Branches entity has a Country field and a City field. Without cascading, the City dropdown shows every city in every country. Users scroll through hundreds of irrelevant options.
What we built: link a child domain attribute to a parent. Select a Country and the City dropdown filters to that country’s cities. Works through any number of levels. Derived columns pull related attribute values into the grid as read-only fields — no extra data entry, no stale copies.
Excel & CSV Import Wizardv0.2026.2.26 The problem: someone hands you an Excel file and says “these are the updated product codes.” Getting that data in used to mean converting to CSV, matching columns manually, and hoping nothing breaks.
What we built: a four-step wizard. Upload your .xlsx, .xls, or .csv. Auto column mapping matches headers to attributes. Choose how to handle conflicts — skip, overwrite, or replace all. Validation runs before anything touches production, and errors are downloadable as a CSV for review.
Why speed matters in MDM
Master data management is not a deploy-once-and-forget system. Your business changes. New regions get added. Product hierarchies get restructured. Regulatory requirements shift. The tool that manages this data needs to keep up with the business, not the other way around. If you want context on what good MDM actually looks like, what is master data management covers the core concepts.
When your MDM vendor ships one update per year, every gap in functionality becomes a workaround you maintain yourself. Spreadsheets fill in for missing import features. Manual SQL scripts replace absent scheduling. Email threads become the approval workflow.
Those workarounds are where data quality goes to die. And once they are in place, they are surprisingly hard to remove.
What comes next
The roadmap runs on the same filter. Data retention management, bulk operations, and API-based entity provisioning are already in progress. If something would save you thirty minutes a week, we are already building it. The approval workflows deep-dive covers the change review system in detail if you want to understand how the workflow layer operates.
You can see every release in our changelog. No marketing spin — just a list of what shipped and when.
Try every feature we just described
60-day free trial. No credit card. Everything included.
Start your free trial