BPO
Outsourced data processing
Data work at volume, with an agreed acceptance criterion and per-batch traceability.
In summary: capture, validation, cleansing and enrichment of data at scale, with sample-based quality control and traceability of every batch. The acceptance criterion is defined before the first record is processed.
Data processing looks like the simplest service to outsource and is where it is easiest to end up with an unusable result. The cause is almost always the same: nobody defined what a correct record means before starting.
So the first deliverable of any data project is the acceptance criterion, agreed against a real sample of your data rather than a theoretical description.
What is included
Tailored to data type and volume in the proposal.
- Capture and digitisation of data from documents or forms
- Validation against business rules defined with you
- Cleansing, normalisation and deduplication
- Enrichment from sources you provide or authorise
- Classification and labelling against your taxonomy
- Sample-based quality control with an agreed acceptance threshold
- Per-batch traceability with exception and rejection logging
What is not included
Worth delimiting so this is not confused with other services.
- Building pipelines and data architecture, which belongs to data engineering
- Analysis, statistical modelling or interpretation of results
- Acquiring third-party data on our own account
- Processing data obtained without a valid legal basis, which we do not accept
Where it is delivered from
From the Latin America corridor for volume, or from Poland where the dataset contains personal information that must stay inside the European Economic Area.
One measure that reduces risk considerably and is often overlooked: processing pseudonymised data where the work does not require identifying the person. If your case allows it, it is worth raising in the proposal.
Processing personal data
Where the data processed is personal, we act as processor under your documented instructions, with a processing agreement, role-based access control and access logging.
If processing takes place outside the European Economic Area, standard contractual clauses apply and, depending on sensitivity, a transfer impact assessment.
One point worth raising early: minimisation is not only a legal obligation, it is the measure that most reduces the project's operational risk. A dataset without unnecessary identifying fields is simpler to process, to transfer and to delete at the end.
Frequently asked questions
How do you guarantee quality at volume?
Statistical sampling per batch against an acceptance threshold agreed before starting, with exceptions logged so the rule gets corrected, not just the record.
What happens to batches that fail the threshold?
They are rejected and reprocessed. The cost of rework attributable to the provider is set in the contract, and it is worth writing down rather than arguing about later.
Can you work with pseudonymised data?
Yes, and where the work allows it, it is the better option. It reduces risk, simplifies the international transfer analysis and usually does not affect the result.
What happens to the data at the end?
It is returned or deleted as agreed in the processing agreement, and the deletion is documented. Worth settling before starting rather than at the end.
Have a volume of data to process?
With a real sample and your acceptance criterion we can size it precisely.