There is a specific look people get when you ask how they use their CRM. A pause, then: "we're not really using it properly."
They almost always are using it properly, in the sense that they type things into it. What they mean is that typing things into it has not made anything better. Which is fair, because a CRM that only stores is a database with a subscription, and a spreadsheet does that for free.
The difference between a CRM that pays for itself and one that does not comes down to whether it does work on its own. Here are the four changes that make that happen, in the order they matter.
1. Nothing arrives by hand
The first and largest problem is that records get created by a person deciding to create them.
Enquiries come in on six channels (phone, web form, WhatsApp, an email to the general inbox, a message on a social page, and someone stopping a van in the street. Some of those get entered. The ones that arrive during a busy afternoon do not, or get entered on Friday with half the detail missing.
So the pipeline is not a picture of your business. It is a picture of what somebody had time to type. Every report built on it is wrong in the same direction, and nobody trusts it, which reduces the incentive to type things in, which makes it worse.
The change: every channel writes automatically, with source, timestamp and whatever structured detail can be captured. Web forms are trivial. Phone requires the phone system to talk to the CRM, which is nearly always possible and nearly never configured. WhatsApp and social need a proper inbox integration. The van in the street stays manual, and that is fine, because it is genuinely rare.
Once this is done, one thing changes immediately: the numbers become real. Everything else on this list depends on that.
2. The stages mean something and moving between them does something
Open most CRMs and you will find the default pipeline stages, lightly renamed. New. Contacted. Qualified. Proposal. Won. Lost.
Ask what has to be true for something to be Qualified rather than Contacted, and you get an answer that varies by whoever you ask. Ask what happens when a deal moves into Proposal, and the answer is "we send a proposal", performed by a human who remembered.
Both of those are the same failure: stages describing a feeling rather than encoding a rule.
The change: each stage gets an entry condition written down, and each transition triggers something. Moving to Qualified sets a follow-up task with an owner and a date. Moving to Proposal starts a reminder sequence that stops when they reply. Moving to Lost captures a reason from a short fixed list, because "lost) no reason recorded" is how businesses stay ignorant about their pricing for years.
The written entry conditions matter as much as the automation. Two salespeople with different definitions of Qualified produce a forecast that is arithmetic performed on opinions.
3. Follow-up runs whether or not anyone is thinking about it
The single most valuable thing a CRM does is remember on your behalf.
Most enquiries that go cold do not go cold because the customer chose a competitor. They go cold because the second contact never happened. The first one always happens (that is the exciting part. The fourth one, eleven days later, is where the deals actually are, and it is exactly the one a busy person forgets.
The change: sequences that are defined once, run automatically, stop the moment a human replies, and escalate to a person when the rules run out. The stopping condition is not optional. A sequence that keeps sending after someone has replied is worse than no sequence, and it is the most common way this gets implemented badly.
This is also where AI genuinely earns its place) not in deciding whether to follow up, which is a rule, but in writing a follow-up that references what the person actually asked about rather than a template with their first name merged into it.
4. Reporting answers a question someone actually asked
The last change is the one people skip, and it is the one that keeps the system alive.
Default dashboards show activity: calls logged, emails sent, deals created. Activity metrics measure whether people are using the CRM, which is a question that only interests the person who bought the CRM.
The change: pick three numbers that a decision depends on, and report only those to start with. For most businesses they are:
- Enquiries by source, so you know where to spend.
- Time to first response as a distribution, not an average, so you can see the tail where the losses are.
- Conversion by source and by job type, so you find out that the channel producing the most enquiries is producing the worst ones.
Three real numbers beat forty vanity ones. Add more when someone asks a question the existing three cannot answer.
Which CRM, then?
Less than people think. We work in GoHighLevel, HubSpot and others, and we build Nexus One ourselves, and the honest position is that the platform matters far less than whether the four things above are true.
A well-configured cheap CRM beats a badly-configured expensive one comfortably. Migration is expensive, disruptive, and frequently reproduces the same problems in a nicer interface (because the problem was never the software.
So the sequence we recommend is: fix the four things in the CRM you already have. If, after that, it genuinely cannot do what you need, migrate with a clear list of the specific capabilities that forced the move. That list will be short and defensible, and the migration will go better because you now know what good looks like.
The test
Here is the one-question version. Close your CRM for a week and change nothing else.
If nothing breaks) if no follow-up is missed, no report goes stale, no enquiry falls through: then it is a spreadsheet with extra clicks, and you should either fix it or stop paying for it.
If things break, it is doing work. That is the whole distinction.
If you want a second opinion on which of the four is missing in your setup, the free AI Opportunity Assessment covers it, and the answer is frequently cheaper than the migration someone has already proposed.