Data ServicesProcessing & related hosting

Data processing & related hosting, England

Records in.
Checked.
Results out.

What we do
Run the routine processing jobs your team keeps doing by hand: cleansing, matching, converting and reconciling business records.
What you get
An output file in the agreed layout, plus a run note that says what came in, what changed and what was held back.
Where it lives
Related hosting for the processed files and the job itself, scoped to what the work actually needs.
Illustration: colour-tabbed batch folders, a sorting tray and stamps on an operations desk
Generated illustration

The work most firms would rather not think about, done the same way every time.

Every organisation has a handful of data chores that never quite earn a project. A supplier sends a price file in a slightly different layout each month. Two spreadsheets that should agree do not. An export from one system has to be reshaped before another system will accept it. Someone in the office knows the steps, does them late on a Thursday, and hopes nothing was missed.

FCR Data Services exists for exactly that kind of work. We take a defined processing job, write down how it should run, and then run it consistently: the same checks, the same rules, the same output layout, with a short note each time explaining what happened. When something unusual turns up, it is held and reported rather than quietly pushed through.

Processing jobs we take on

Each drawer is a kind of job. Open one to see what goes in, what comes out, and where the limits sit.

Cleansing and standardising

In: spreadsheet or delimited exports with inconsistent spacing, case, date formats and postcodes.

Out: the same records in one agreed format, with every changed field counted in the run note.

Limit: we standardise to rules you approve; we do not guess at missing values or invent records.

De-duplication and matching

In: one or more lists that overlap, such as contacts, products or account references.

Out: a merged list plus a review sheet of likely matches that need a human decision.

Limit: close matches are flagged for you to decide. Nothing is merged on a hunch.

Format conversion and reshaping

In: files in the layout one system produces.

Out: files in the layout another system or person needs, with column mapping written down and versioned.

Limit: we work from files you can legitimately export and share; we do not reverse-engineer closed systems.

Reconciliation runs

In: two sources that should agree, for example an order export and a dispatch list.

Out: a matched set, an exceptions list and totals for each side so differences are visible at a glance.

Limit: we report differences; deciding which side is right stays with your team.

Scheduled recurring processing

In: a file that arrives on a rhythm: weekly, monthly or at quarter end.

Out: the same processing applied each time, compared against the previous run so drift is noticed early.

Limit: schedules are agreed working-day windows, not round-the-clock monitoring.

Why a run note matters more than the output file

An output file on its own asks you to trust it. A run note lets you check it. Ours is deliberately short and plain: how many rows arrived, how many passed validation, how many were changed and by which rule, how many were held back and why, and how the totals compare with last time.

That one page is often the most useful thing we hand over. It gives a manager something to sign off, gives the person who used to do the job a way to spot-check it, and leaves a trail if anyone later asks how a figure was produced.

Run note illustrative example

  • Rows received4,318
  • Passed validation4,271
  • Standardised (date format)612
  • Standardised (postcode spacing)188
  • Held: missing reference39
  • Held: duplicate reference8
  • Change vs previous run+2.6%

Held rows returned in a separate sheet for your decision.

Illustration: punched index cards in a steel sorting mechanism
Generated illustration

Related hosting, kept in proportion

Processing does not stop at the output file. Results need somewhere to sit where the right people can collect them, earlier runs need to be kept for comparison, and the processing routine itself needs a stable home so it runs the same way next month.

Our hosting work is tied to those needs: arranging where processed outputs are held and for how long, separating incoming files from finished ones, and keeping the job configuration under change control. We are open about where that hosting sits and which provider it relies on before any data moves, and we do not claim facilities we do not run.

How related hosting is scoped

From first message to first run

Six steps, in this order, every time. Nothing is processed before step four.

  1. DescribeYou tell us about the job in words: what arrives, what you need back, how often. No data at this stage.
  2. Sample layoutWe look at column names and an anonymised or dummy sample so the rules can be drafted.
  3. Written rulesA processing specification you can read: every check, every transformation, every reason a row would be held.
  4. Transfer agreedWe agree how files move, who can see them and how long each copy is kept, then put the necessary data processing terms in place.
  5. Trial runThe first run is compared side by side with your current manual result. Differences are explained, not hidden.
  6. RoutineThe job runs to its rhythm. Changes to the rules are versioned and dated.

A good fit, and a poor one

Usually a good fit

  • A repeating file that someone reshapes by hand
  • Two lists that keep drifting apart
  • A backlog of exports that need tidying before a system change
  • A team that wants a written record of how figures were produced

Usually not for us

  • Live transaction processing that must run every minute of every day
  • Advice on what the law requires you to do with personal data
  • Building a new customer-facing application
  • Work that depends on access we would have to obtain without the system owner's agreement

Three bench tools you can use now

They run in your browser. Nothing you type is sent anywhere.

Standards we hold ourselves to

Nothing silent. Every row that is changed or held is counted and explained.

Rules before data. The specification is written and agreed before a real file is opened.

Least copies. Working files are deleted at the end of the agreed retention period, and we confirm when.

Versioned changes. If a rule changes, the date and reason are recorded.

Plain reports. Written for the person who signs off the figures, not for a technician.

Honest limits. If a job is outside what we do, we say so early.

Read the full delivery standards

When something looks wrong

If an output does not match what you expected, use the support form and name the job and the run date. We respond within working hours, Monday to Friday, 09:00 to 18:00 UK time. Every query is answered by comparing the run note with the input; you will hear what we found, not just that we are looking into it.

Support desk

Useful reading

Have a job that keeps coming back?

Describe it in a few sentences. We will reply with questions, a view on whether it fits, and what a first trial run would involve.