Data processing & related hosting, London

Data that changes hands without changing meaning.

Datapasa designs and runs the transformations between one system, team or organisation and the next, and hosts the in-between stages so every handoff is clear about what was passed, in what shape, and when.

Illustration: colourful ribbons flowing through rings between lanes
Generated illustration

Where data usually goes wrong: between places.

Inside one system, data tends to behave. The trouble starts when it leaves. An export becomes an attachment, the attachment is opened and tidied by someone who knows the quirks, the tidied version is uploaded somewhere else, and a week later a number in a report does not match. Nobody did anything wrong. The handoff simply had no shape.

Our work sits in those gaps. We describe each movement of data as a flow with a start, a set of transformations and a receiver, and we agree a clear handoff at each boundary. Then we run the flow, host the stages in between, and keep a record of what passed through.

The anatomy of a flow

Four kinds of flow we build and run

All within data processing and the hosting that supports it. No apps, no dashboards for sale, no platform to subscribe to.

A handoff agreement fits on one page

The single most useful thing we produce is not code. It is a short description of each boundary, agreed by the people on both sides. It turns "the file" into something specific: these columns, this format for dates, this character set, delivered here by this time, with this person told if it is late or different.

When a receiving team says the data looks wrong, the agreement is the first thing to check. Either the flow broke it, or the agreement needs to change. Both are fixable, and neither involves guessing.

Draft one with the handoff tool

Illustrative handoff

Sales orders exportWarehouse picking list

Layout
11 columns, header row, UTF-8, comma-separated
Dates
YYYY-MM-DD
Schedule
Working days by 07:30
Excludes
Cancelled and on-hold orders
If late or changed
Warehouse lead notified by email
Illustration: coloured streams sorted into parallel bands by a prism
Generated illustration

Hosting for the in-between

A flow needs somewhere to put data while it is between owners. Our related hosting covers those staging areas: the landing place for an incoming file, the intermediate results of transformations, and the outgoing package waiting for the receiver.

Staging is temporary by design. Each area has a stated purpose, a list of who can reach it and a clear-out rule. We tell you which hosting provider and region are used before anything is staged, and we do not describe this as a data centre service, because it is not one.

How staging is arranged

How a new flow comes together

Trace it

We follow the data as it moves today, including the manual steps and the spreadsheet everybody forgets about.

Name the boundaries

Every point where data changes owner, system or shape becomes a handoff to agree.

Write the steps

Transformations are described in plain language first, then built.

Rehearse

Dummy or approved test data flows end to end. Receivers check the output looks right to them.

Run and record

The flow goes live on its schedule, with a record for every run.

The full method

What we are careful not to promise

Real-time streaming at any scale. Our flows run on schedules or on arrival, during UK working hours.

Access to systems we have not been given legitimate access to. Each source owner agrees the extract.

Legal advice about sharing data with partners. We implement what you and your advisers have decided.

Round-the-clock monitoring. If a run fails overnight, it is picked up the next working morning.

Try the three flow tools

What changes when handoffs have shape

Something not arriving?

If a flow we run is late, empty or shaped differently from its agreement, tell us which handoff and when. We check the run record and reply in working hours with what we found.

Go to help

Field notes

Questions people ask before the first flow

Do we have to change the systems we already use?

Usually not. Most flows work with the exports and imports your systems already offer. Where a system cannot produce what a handoff needs, we explain the gap and the options, and the decision stays with you and whoever supports that system.

Who owns the transformations once they are built?

You do. Every step is described in plain language and versioned alongside the handoff agreements, so another provider or your own team could take the flow over with the documentation alone.

What if a partner changes their file without warning?

The run notices that the incoming shape no longer matches the agreement and holds the handoff rather than passing on something broken. We tell the named owner, and the agreement is updated once both sides confirm the new layout.

Can you start with something small?

That is the best way to start. One boundary that causes regular friction, mapped, agreed and run for a few cycles, tells you quickly whether this way of working suits your organisation.

How is personal data handled inside a flow?

Only the fields a receiver genuinely needs are passed on, and written terms covering our processing are agreed before anything real moves. What you are permitted to share is a decision for you and your advisers; we implement it carefully.

Pick one handoff that annoys you.

Tell us where it starts, where it needs to end up and what goes wrong in between. That is enough for a first conversation.