Most integration projects quietly rest on one habit: copy the data somewhere central and work on the copy. It feels safe. It is also the single biggest reason your reports disagree - and it is the habit boTec designed out more than 25 years ago.
In Brief
- Copying data into a lake, warehouse or spreadsheet creates a second version that starts ageing immediately.
- Copies cost three ways: staleness between batch runs, endless reconciliation, and multiplied interfaces (n×m) plus storage.
- boTec uses referential integration - live links to the source, resolved in real time, never duplicated.
- The single source of truth stays in the system that owns it, so KPIs and reports stop disagreeing.
Part of the series , Data mesh & the boTec DataBridge
- Part 1 - Data mesh, since 1998 - 20 years before it had a name
- Part 2 - Link, don’t copy - why referential integration beats copying - you are here
- Part 3 - Harmonise, don’t standardise - Business Integration Models
- Part 4 - From data maturity to AI - getting the sequence right
The hidden cost of a copy
The moment you copy a number into a lake, a warehouse or a spreadsheet, two versions of the truth exist. From that instant they drift apart. Batch jobs run overnight; between runs, the copy is wrong by definition. Nobody decided to make decisions on yesterday’s numbers - the architecture decided it for them.
A copy carries three costs at once:
- Staleness. Between batch runs the copy is out of date - and you rarely know by how much.
- Reconciliation. When the copy and the source disagree, someone has to work out which number is right. That work never ends.
- Multiplication. Every new copy adds interfaces to maintain (the n×m problem), storage to pay for, and one more place for the truth to diverge.
A copy isn’t your data. It’s a snapshot that’s already going out of date.
→ Want to know which copies are costing you most? Book a free 30-minute check
The boTec principle
Link, don’t copy - referential integration
Instead of copying data into an intermediate store, the boTec DataBridge maintains live references to the source systems and resolves queries in real time. Product structures stay in PLM, inventory stays in ERP, process values stay in the historian - and the DataBridge reads them where they live. This one architectural choice removes an entire category of problems:
- No synchronisation lag between source and integration layer.
- No reconciliation - there is only ever one number, in its authoritative system.
- No stale copies eroding trust in KPIs and analysis.
- No duplicate storage cost for replicated data sets.
- The single source of truth stays where it belongs - in the system that owns it.
“But doesn’t everyone copy?”
Yes - and that’s exactly the problem
- Data lakes and warehouses are copy-by-design: everything is pulled into one store where it immediately begins to age.
- Point-to-point and iPaaS tools copy and multiply - each new interface is another copy to keep in sync.
- Even “copy integration” - a central copy kept broadly in sync - is always slightly wrong, and slightly wrong is enough to lose trust.
Others move your data around. boTec points to it where it lives.
What it saves you
Trustworthy numbers, lower cost, faster decisions
Linking instead of copying is not an architectural nicety - it shows up on the balance sheet and in the boardroom. Reports that agree because there is only one number. Lower storage and interface-maintenance bills. New data sources onboarded in a fraction of the time. And decisions made on what is true now, not on last night’s batch.