Information systems · Scenario 2 · Retail SME
Your shop already produces the data. It just isn't keeping it.
A practical guide for Lerato's Hardware & Home on moving from spreadsheets, paper and memory to a system that captures a fact once and reuses it everywhere — written for the people who run the shop, not for an IT department.
tied up in stock that has not sold a single unit in twelve months
owed to the shop on account, across ten customers
written off or lost to stock-take differences, with no explanation recorded for most of it
of turnover comes from just three contractor accounts
Figures from the demonstration data set. Not one of them is visible in a spreadsheet, yet every one is the owner's own money.
The problem, stated plainly
A spreadsheet stores a value. It does not store what that value means, where it came from, when it was last checked, whether anybody verified it, or what else depends on it. So every time the shop needs a quote, an invoice, a stock take or a supplier order, somebody retypes the same facts — and every retype is a fresh chance to be wrong.
An ordinary day at Lerato's would look like this, sales being written in a duplicate book at the till, debtors' details are in a separate book in Lerato's handwriting. Sipho is asking for stock on the WhatsApp group and Thabo is trying to answer from memory. The one spreadsheet with the stock levels and prices is on a laptop and only updated when one of the staff has time. Each part has their own process but are not synced together and is prone to errors.
For most days this works but the moment it doesn't it's usually a small issue. A contractor disputes an invoice, and nobody can recall what the agreed price was or a customer was promised a geyser element that is not in stock now. Month end takes up a full weekend of adding up books that should have been added up.
Just buying software on its own will not help the situation. The problem is not which program the shop needs, rather how facts are captured and kept, checked by nobody and stored with no record of where they came from. Putting these habits into new software leads to wasting an investment and the same problem occurring.
The demonstration shows the data quality screen that lists every broken record by name which includes a customer with R12 000 of credit whose phone number is incomplete. A system does not repair that record by itself but makes sure someone sees it before it costs money.
The whole premise of this resource is capture once, use many times. Facts entered properly at the counter can give an invoice, stock count, debtors list, and a POPIA request without being typed again or looking around multiple books. Data is the raw fact, information is data that has been given meaning, and knowledge is knowing what to do with it.
The idea in one picture
Everything in this resource is a variation on one diagram. A fact enters the business once, is checked while the person who knows it is still standing there, and is then stored in a form that lets it be used again without being typed again.
Scroll sideways to see the whole diagram.
The win is not that the computer is faster at typing. The win is that a fact captured properly once can be reused — on an invoice, in a reorder calculation, in a POPIA request, in the margin report — and all of those stay in agreement with each other because they are all looking at the same record.
What this resource covers
Five areas, in the order they matter to a business of this size. Each one is written to be read on its own, in about ten minutes.
Data & information management
The difference between a row and a record. What metadata is, why a spreadsheet cannot hold it, what information quality means in a shop, and how dashboards turn trading into decisions.
Read this first →Data ethics, privacy & POPIA
What the Protection of Personal Information Act actually requires of a shop this size, the four questions you must be able to answer, and the difference between being compliant and being decent.
Read →AI & decision support
Which decisions a small shop should hand to a system, which it should keep, and how to tell a useful forecast from an expensive guess.
Read →Enterprise systems & integration
What POS, inventory, accounting and CRM each do, why they stop talking to each other, and what integration costs when nobody planned for it.
Read →Adapting systems to fit
Buy, configure or build. Where to bend the system to the shop and where to bend the shop to the system — and why that choice decides whether staff use it.
Read →What to actually do
A staged plan with real South African costs, what to do first, what to leave alone for now, and the honest trade-offs at each step.
Read →Don't take our word for it — use it
The demonstration is a working system, not a video. Capture a sale, break the rules on purpose, generate the invoice, then look at what was actually stored.
Capture a sale
Type a quantity. Everything else on the invoice comes from records that already exist.
Open the till →Try to break it
Sell stock you do not have, price below cost, put a cash customer on account. Watch what happens.
Try it →Answer a POPIA request
Pick a customer, press one button, read everything the shop holds about them.
Try it →Take your data back
Export the lot to a file you can open and read. Check this before signing up for anything.
Try it →The demonstration stores everything in your own browser. No account, no sign-up, no data sent anywhere. Every person, supplier and transaction in it is invented.
What we are not going to tell you
The honest trade-offs are that it is going to cost you every month and it is going to take some time. The staff must learn the new system so the first few weeks will be slower, and the actual costs are on the roadmap page.
There are three things that would most likely happen, the staff going back to using the old manual system on busy days, migration issues will occur if old data is invalid, and vendors will be difficult to leave once you have multiple years of data with them. Each is manageable only if someone is there to monitor it.
Some things should not be digitised yet such as the automated ordering, the reorder suggestion should stay a suggestion that Lerato approves. Do not try to predict demand with the old data until there is at least 6 months of verified sales data to use. Keeping the books in case of loadshedding is a must because the day you need it is the day there is no power for the system. The paper book stays only as a fallback, the real answer to loadshedding is a system that works offline and has backup power. Lerato's judgement on pricing and credit remains top priority so the system should record what she decides on.
A system does not fix a business that is not making money. It makes the numbers visible so the owner can act on them. If the margins are small, it will show it and the rest of the work is on the team.
How to use this resource
Read the data page first
Everything else assumes the capture-once idea. It is the longest page and the one worth your time.
Open the demonstration beside it
Each page links to the screen that shows the thing being described. Reading about a reorder calculation is not the same as looking at one for your own stock.
Argue with the roadmap
The staging is a proposal, not a prescription. The order should change to fit what is hurting the shop most right now.
Keep the glossary open if a word is unfamiliar
Every term is explained in plain language, with the shop's own example beside it. Open the glossary →