Already written down. Never read back.
Twelve months of them, in a format nobody enjoys reading. Send us one export. We read back every number it can answer, and name the ones it cannot.
These four need almost nothing. If your file opens, they come back. Everything after this depends on which columns are present, and we tell you which ones were missing.
Record completeness is a finding about the system, not about the people using it.
Put in three numbers you already know. Nothing is sent anywhere, and nothing is stored.
Example figures. Drag or type to use yours.
This is your three numbers multiplied together, and nothing else. No rejection rate, no shrinkage, no industry average. It is not money you have lost. It is money nobody has read back.
We read what your system already recorded, and tell you what it says.
Reading back what the system already recorded. Counts, totals, dates, gaps. No modelling, no prediction, nothing that is not already written down. This is most of the work and it is the part people are most surprised by, because the answers were always there.
We find the patterns you would not catch by eye, and show the working.
Used where reading it back is not enough: patterns across time, or across accounts, that you would not catch by eye. We show the working and name the assumptions. If a pattern is weak, you are told it is weak rather than sold it.
We stop your staff doing the same job twice.
Usually this is a report someone rebuilds by hand every month, or a reconciliation done with two screens open. It follows the reading rather than leading it, because automating a process nobody has examined just makes the wrong thing faster.
Who this is for
What comes back depends less on your industry than on which columns your system happens to write. Pick whichever is closest to the business you run.
Where the deadlines are real and written into regulation. A returned claim must be corrected and resubmitted within sixty days of being sent back.
The till has been keeping a record of every decision made at the counter, including the ones nobody reports on.
Every consignment leaves a record of how it was handled, and on instrumented equipment, of what it was exposed to.
How an engagement runs
Five steps, and two ways into the first. Start with an export and you can stop after we read it back and owe nothing.
If you run one system, you send one export: whatever it gives you by default. If you run several, or do not know what you have, we start with discovery.
Export is free. Discovery is scoped and priced before it starts.How many records it holds, what period it covers, how complete the fields are, and what it could answer if you went further. If the file is thin, this is where we say so.
In plain sentences, not a dashboardFrom what the file can actually support, not from a menu written before we saw it.
You get the number before any work startsEvery figure we hand back can be traced to columns in your own export. Where your data cannot support a question, you get that answer instead of a confident one.
Assumptions written next to the numberA once-off read ends here and your file is deleted. If you want the same questions answered every month, that is priced differently because it is different work.
Deleted within seven days if nothing followsWe are not a software vendor. There is nothing to install, no platform to log into, and nothing that replaces the system you already run. We read what it produces.
We are not a large firm, and the site does not pretend otherwise. The people who read your file are the people you deal with.
Questions
Then you get less, and we tell you what was missing and which column would have produced it.
Some things come back from any export: how many records there are, what period they cover, how complete the records are, and whether anything is recorded twice. Amounts add where your revenue is concentrated and what is going out in discounts. Payment dates add what is unpaid, how old it is, and how long payment usually takes. Item lines add what has stopped moving.
No names, no ID numbers, no contact details. Take those columns out before you send it.
What the numbers do need is a consistent code, so the same account reads as the same account across twelve months. Whether code 4471 belongs to a name stays on your system. Scheme codes, item codes, dates and amounts carry the rest.
If your system will not export without the name column, open the file, delete that column and save it before sending. Nothing we do needs it.
Some of what comes back is about how the system is used rather than what was sold. We report that as a finding about the system, not about a person.
It goes through an upload form hosted by Google, into storage behind it that may sit outside South Africa. Google holds a copy alongside us until we delete it. No analytics service and no advertiser ever sees it.
A few people have access to it. That is the whole company, so the list is short and it does not change.
You send it through the upload form, not by email. We keep it for as long as we are working with you, then delete it. If we read it and nothing comes of it, it is deleted within seven days and the trash is emptied. The full notice.
We do not publish a figure, because the work is not the same size twice.
Two things move it: how many systems we are reading, and whether you want them read once or kept running. Both are things you decide, not things we discover afterwards.
The first step costs nothing. Send one export and we will tell you how many records it holds, what period it covers, how complete it is, and what it could answer if you went further. If the answer is "not much", we will say so and that is where it ends.
Contact
If you run one system, send one export: the first read-back costs nothing. If you run several, or do not know what you have, book a discovery workshop.
Free
It costs nothing, and if your file cannot answer much we will tell you that instead of selling you something.
The upload form is hosted by Google, which holds the file alongside us until we delete it.
Some systems only export properly from the workstation on the premises. If that turns out to be yours, we come to you.
Sometimes the data is there but the column is not switched on in the export. If that is what we find, we tell you which one and where the setting sits.
Scoped and priced before it starts
Every engagement starts with a conversation. Tell us what you run and where the pressure is.
Write to consult@sinewcollective.co.za with your name, organisation and the systems you run.
Or write to consult@sinewcollective.co.za.
We do not publish a figure. We do not invent one either.
