The short definition
ERP stands for Enterprise Resource Planning, a term inherited from 1990s manufacturing software that describes the category poorly. Ignore the name. An ERP is the single system that records what you own, what you owe, what you have promised to customers, and what all of it cost you.
That sounds modest until you notice how many separate places those four facts currently live in your business. The stock figure is in a spreadsheet. What you owe is in accounting software. What you promised is in an inbox and a rep's notebook. What it cost is an estimate somebody made when the container landed. An ERP's only real claim is that all four come from one place and agree with each other.
Everything else: the modules, the dashboards and the workflow engines exist to keep that single record accurate while dozens of people transact against it at once.
ERP versus accounting software with add-ons
This is the comparison that matters for most growing businesses, because the alternative to an ERP is rarely "nothing". It is QuickBooks plus an inventory app plus a shipping portal plus a storefront admin plus the workbook that reconciles them.
Accounting software records financial transactions after they happen. An ERP runs the operation that produces them (the purchase order, the receipt, the putaway, the allocation, the pick, the shipment and the invoice) and posts the financial result as a consequence. The distinction is not academic:
- Costing. Accounting software takes a cost. An ERP calculates one, apportioning freight, duty and brokerage across the units that actually arrived.
- Commitment. Accounting software knows what you have. An ERP knows what you have left after everything already promised.
- Sequence. Accounting software records the invoice. An ERP knows the pick was short before the invoice exists.
If your accounting package is currently where inventory lives, you are already running an ERP, just without the operational half, and with people filling the gap by hand.
The honest signals that you need one
Size is a weak predictor. Complexity is the real one. Two or more of the following, sustained for a couple of quarters, is when spreadsheets start costing more than software:
- Stock is held in more than one location, including consignment, third-party logistics or a vehicle.
- You sell through more than one channel and each channel holds its own idea of availability.
- Landed cost matters to your margin and is currently estimated.
- You track lots, batches, serials or expiry — or a customer or regulator could ask you to.
- More than one person needs the same number at the same time and they disagree.
- Month-end requires a workbook only one person can operate.
- You cannot state gross margin by customer or by SKU this week without building it.
Why cloud, specifically
A cloud ERP is delivered as a hosted service and used through a browser. No servers in a cupboard, no client software to install and update on every machine, no VPN before a rep can check availability.
The practical consequences are more interesting than the infrastructure story. Warehouse handhelds become ordinary phones. A rep in a customer meeting sees live stock. A second warehouse is configuration rather than hardware. And upgrades stop being projects, which matters, because on-premise ERPs frequently sit five years behind their own release, unpatched, because upgrading is a quarter of work nobody has scheduled.
What it costs, in structure if not in number
ERP cost has three components, and vendors who publish only the first are being economical with the truth:
- Licence or subscription. Usually per named user, often tiered by role.
- Implementation. Configuration, data migration, integration and training. On mid-market projects this frequently exceeds the first year of licence. It is also where fixed scope matters enormously as an hourly implementation has no natural end.
- Internal cost. Your team's time during migration, training and the first month live. Real, routinely omitted from business cases, and the reason short implementations beat long ones on total cost even at a higher day rate.
Ask every vendor for all three as one number with a defined scope. The ones who cannot produce it are telling you something.
Why ERP projects fail, and what to do about it
The failure causes are almost never technical:
- Scope expanded mid-project. Every "while we are in here" adds a week and moves the go-live.
- Nobody cleaned the data. Migrating twenty years of duplicate item records reproduces the mess at higher cost.
- Training was an afterthought. A system used incorrectly for a month generates data nobody trusts, and trust does not come back easily.
- No single decision-maker. Process questions need someone who can settle them in a day, not a committee that meets fortnightly.
- It took too long. Duration makes every other risk worse. Sponsors leave, priorities shift, and enthusiasm is a depleting resource.
Which is why we treat timeline as the primary design constraint rather than a scheduling detail. A four-to-six week deployment is not a marketing claim — it is a structural defence against the five failure modes above.