Every business has at least one weekly task that lives in a spreadsheet. Someone opens it, reads supplier values, applies a calculation, copies the result into another system, saves the file in a specific format. Hours. Every week. Forever. We build a small custom processor tuned to that exact input and output. Drop the file, click once, done.
Hours saved per week
3-6
Time per task after
~5 min
Live processors shipped
3+
Output formats supported
Any
The bottleneck
Weekly tasks that eat staff time forever.
Every business has them. The bookkeeper who spends three hours every Monday processing a supplier price list. The manager who reformats a time sheet into payroll-ready columns. The warehouse lead who reconciles a stock report from a portal into Shopify, line by line.
Different files. Different industries. Different output systems. But the shape of the work is the same: a person sits in front of a spreadsheet for hours doing the same calculation, the same lookup, the same reformat.
It's the kind of task that's too bespoke for an off-the-shelf tool, too small to justify a full software project, and so it just keeps consuming staff hours week after week.
The solution
A processor tuned to your task.
We build a small custom processor for that exact task. Not a generic tool — a specific one. Drag the messy input file onto a dashboard. One click. Out comes a clean output file in the exact format your next system needs.
If your input is a supplier CSV, the processor parses it. If it's a multi-tab Excel file, it parses every tab. If it's a PDF or a portal export, the processor handles that too. The output is whatever your downstream system wants — Shopify CSV, ServiceM8 import format, MYOB columns, a direct API push.
The hour-long weekly grind becomes a five-minute upload. Every week. Same processor, same output format, every time correct.
Margin Manager · drop the supplier price list, click Calculate, get a ServiceM8-ready CSV back — 463 items recalculated, run finished in seconds
The outcome
Same task. A fraction of the time.
Each processor we ship is tuned to one specific input/output pairing — so the result is predictable, repeatable, and immediately usable. No human reformatting on either end.
Staff stop dreading the weekly task. The output goes straight into the next system without re-editing. New product lines or supplier additions are handled by updating the processor once, not retraining everyone who touches the file.
Metric
Before
After
Weekly time spent
1-6 hours of manual work
~5 minutes (upload + click)
Output formatting
Hand-typed, error-prone
Always correct, every run
Adding new products / suppliers
Retrain the staff member
Update the processor once
Cost of staff turnover
Whole task knowledge walks out
Knowledge is in the processor
Three live processors
Same pattern, different industries.
Trades · ServiceM8
Margin Manager
Jentec Electrical. Supplier sends a price file weekly. The processor applies a margin calculation per product line and outputs a ServiceM8-ready import. What was a multi-hour Friday job is now a 5-minute upload before knock-off.
Retail · Shopify
Discount Manager
Raw Blend. Supplier price list comes in, the processor applies the negotiated discount and outputs an updated Shopify PurchaseCost file. Margins stay correct without anyone touching every product in admin.
Retail · Shopify Inventory
Supplier Stock Feeds
Raw Blend. Three suppliers send stock data three different ways (portal CSV, Google Sheet, weekly Excel). One unified processor handles all three formats and keeps ~470 SKUs in sync daily, untouched.
How we price
Priced on your saving, not our effort.
Every build is quoted at one-fifth of what it saves your team in man-hours per year. The other four-fifths stay with your business, every year, after the build.
Most builds in this tier land between $2,500 and $25,000, depending on the size of the saving — a small single-task processor sits at the lower end, a multi-system platform at the upper.
Local
Lives on your computer
Desktop app or local automation that runs on a PC you already own. One-off payment.
Hosted
Lives on our infrastructure
We host, run and maintain the build. Monthly subscription, sized so cost lands as the saving lands.
Discovery is free. We spend 30 minutes mapping the saving on a call and quote a specific number inside 48 hours.
Frequently asked
Questions worth getting out of the way.
Who owns the build after we ship it?
+
You do. Source code, deployment guide, and a runbook — all yours. There's no per-seat licence to keep paying and no vendor lock-in. If you decide to maintain it in-house later, you can. We're happy to stay on as a retainer; we don't make you.
How do you work out the saving?
+
On the discovery call we ask what your team does today — name the task, the frequency, the hours, the wage rate. The estimate is conservative. If the maths doesn't land somewhere we both think is worth it, we'll tell you on the call rather than sell you a build.
How long does it take from scope to live?
+
Most builds run one to three weeks from confirmed scope to production. We work in working prototypes — you'll see something running by the end of week one, refine through week two, hand over and go to production around week three.
What if our stack is unusual?
+
Most are. The framework we ship against is the same; the connectors differ. We've shipped against Shopify, ServiceM8, QuickBooks Online, MYOB, Xero, ShipStation, custom supplier portals, Google Sheets and Excel-by-email. If your system has an API or a file export, it's in scope.
What if we want to extend it later?
+
Easy and cheap. The platform pattern means a second or third automation slots in for hours of work, not weeks. After the first build the marginal cost drops sharply — most clients return for a second within a few months.
What if our supplier sends a PDF, not a CSV?
+
Handled. PDF parsing — including multi-page tabular PDFs, scanned PDFs with OCR, and odd supplier formats — is part of the toolkit. The processor cares about the data, not the file format.
What if the file format changes from one week to the next?
+
We build in defensive parsing — column names matched flexibly, missing columns surfaced rather than silently dropped, unexpected rows flagged. When a supplier does change format, the dashboard turns red and the run halts before bad data lands in your downstream system.
Can multiple staff drop files?
+
Yes — the dashboard isn't single-user. Anyone with access to the URL can upload, run, and download results. Audit log shows who did what.
Want the technical detail — APIs, code, the gotchas we hit, the actual log output?
It's all in the Raw Blend case study.
Related Builds
Other angles on the same platform.
Every page on this list is drawn from the same Raw Blend build — different problems, same operating model.
Send us the file your team processes every week (or describe the workflow) and tell us what system the output needs to feed into. We'll quote a specific processor in 48 hours.
