Custom CRM or off-the-shelf?
Start with an off-the-shelf CRM and stay with it while your work fits its shape. Build your own when the workarounds have become the system: the same thing typed into three tools, approvals happening in chat, and the number you need living in a spreadsheet beside it. We built our own, a.hub OS, for exactly those reasons, and we still tell most businesses to try what they have first.
Part of: Custom CRM and ERP
Every growing business reaches the point where its CRM, or its spreadsheet pretending to be one, starts to creak. The tempting answer is to build something of your own. Sometimes that is right. Often it is not yet.
When off-the-shelf is the right answer
Off-the-shelf CRMs are good at the common shape of selling: a contact, a deal, some stages, a reminder. If your process looks like that, a well-set-up tool you can switch on this week will beat anything you could build this year.
Before you decide it does not fit, spend an afternoon setting it up properly: your own stages, your own fields, the reports you actually look at. A surprising amount of "the CRM does not work for us" is a CRM nobody configured.
The signs you have outgrown it
- The same thing is typed twice. A lead goes into the CRM, then into a proposal document, then into the invoicing tool, by hand each time.
- The real work happens beside it. Approvals in WhatsApp, the job status on a whiteboard, the real number in someone's spreadsheet.
- Your process has steps the tool cannot name. A client review, a second approval, a hand-off between teams that the CRM has no stage for.
- You see the money too late. You only find out whether a client or a job made money after it is too late to change anything.
One of these is a nuisance. All four is a business running on workarounds.
What we did
Our own agency ran on a work board, three spreadsheets and a WhatsApp group, and we could see the profit on a job only once it was too late to fix. So we built the system we needed, a.hub OS: leads that become proposals, then clients, without retyping; work that moves through our own named steps and approvals; and the money next to the work. Other agencies wanted it, so it became a product, and we still run Artwing on it every day.
What to settle before you build
- The one job it must do first. Not a feature list. The single change that would make the biggest difference to how you work.
- Your process, written down. Every stage, every approval, who does what. If you cannot write it down, software will not fix it.
- What the AI may do. Reading, sorting and drafting save people time. In a.hub OS the AI drafts and a person approves; it never publishes or changes a record on its own.
- A scope and a timeline in writing, before anyone starts.
- Who looks after it afterwards. Your own software needs looking after: someone to fix it, update it and add to it.
A middle path
You rarely have to choose all or nothing. Many businesses keep a standard tool for the common parts and build only the missing piece: a client view, an approval step, a report that joins two systems. We start every conversation by trying to talk you out of the expensive version. You can see how on our custom CRM and ERP page.