In this article (7 sections)
Choosing a business management system is one of the most expensive decisions a small company makes, and it's almost always made with the worst possible method: watching three demos and picking the one that looked nicest.
The problem is that demos are designed for exactly that. Nobody is going to show you the module their product doesn't have.
This is a procedure for deciding with different information.
Before looking at a single product
The original mistake is starting with the catalog. If you start by asking "what systems are out there?", you end up comparing features you don't know whether you need.
Write down your process first. Not a wish list: the actual circuit, as it happens today. From the moment an order comes in to the moment it gets paid. Who does each step, what document gets generated, where it jams.
Two hours with the people who do the work is enough. What comes out of that is the standard you'll measure everything else against, and it also lets you ask for comparable quotes.
Separate the non-negotiable from the nice-to-have. Out of everything you wrote down, mark what's genuinely indispensable. The indispensable list is usually much shorter than it looked, and it's the only one that matters for ruling options out.
Identify your oddities. Every company has two or three things it does differently: a pricing logic of its own, a particular approval circuit, a requirement imposed by one big customer. Those oddities are what will decide this, because they're exactly what a standard product doesn't cover.
If you can't name your oddities, you probably don't have any, and that's good news: it means an off-the-shelf product will work for you.
The three real options
A vertical product for your industry. A system built for pharmacies, for auto shops, for distributors. It already covers the vocabulary and the typical circuits of the sector. It usually gives the best price-to-result ratio if your operation is standard within the industry.
A general ERP. It covers the cross-industry administrative side — invoicing, stock, customer accounts, accounting — without knowing your industry. More flexible, more configuration, more implementation time.
Custom software. Built on your process. Between USD 8,000 and 25,000 for a complete small-business management system, with timelines of two to four months.
The practical rule: if your oddities can be accommodated by changing your process without losing anything, buy. If changing them takes away what sets you apart, build.
The questions that separate a serious vendor
These are for the meeting, and the uncomfortable ones give you the most information.
About the product
- What that I need does your system not do today? "It does everything" is the worst possible answer.
- How much of what you showed me is standard and how much is configuration that has to be paid for?
- Can I see the system running at a company like mine, at their volume?
About the data
- How do I export all my information if I want to leave? In what format?
- Where is the data hosted and who owns the account?
- How often is there a backup and how does a restore work?
About the money
- What's the total cost of the first year, including implementation, migration, training and licenses?
- How does the price scale if I double users or volume?
- What gets charged separately that isn't in this quote today?
About afterwards
- Who do I talk to when something breaks, during what hours, and with what committed response time?
- What happens if my case needs specific development? Do you do it, how long does it take, who maintains it?
Number four is the most important on the list. A system you can't get your data out of isn't a vendor, it's a hostage situation.
How to ask for a demo that's worth something
A standard demo doesn't give you information: it gives you the salesperson's optimized route.
Send three of your own cases in advance and ask them to do those live. Pick the ones that represent your oddities, not the easy ones. An order with a special discount, a partial return, the invoice for that customer who demands a particular format.
What you'll see is far more useful than any feature list: whether they do it in three clicks, whether they need configuration that wasn't quoted, or whether they suggest changing your process. All three answers are valid. The one that isn't: "we'll look at that during implementation".
Ask that someone from your team drive it, not the salesperson. Ten minutes from a person who's going to use the system every day is worth more than an hour of presentation.
The warning signs
- The quote isn't itemized. A single number without licenses, implementation, migration and training separated out is going to grow.
- No verifiable references. Ask for two customers of a similar size and call them. If they won't give them to you, you already know.
- Data migration "isn't a problem". It's always a problem. Historical data is dirty and cleaning it takes time.
- Pressure to sign fast. A discount for closing this week, on a decision you'll live with for five years.
- Everything gets solved "with some custom development on top". If you need to build half of it, you're paying twice.
The cost that isn't in the quote
The biggest one of all doesn't get invoiced by anybody: your team's time.
Implementation takes hours from people who already have a job. Defining circuits, reviewing migrated data, testing, getting trained, and keeping the old system running in parallel through the transition.
Budget that as if it were money, because it is. And schedule the implementation away from your busy season, which is the most expensive and most common calendar mistake there is.
In one line
Write down your process, mark your oddities, and pick the option that respects them without making you pay twice. Everything else — the interface, the brand, the number of modules — is secondary next to that.