What is a CRM? A plain-English guide for small businesses

Most CRMs, including bizzMate, are built around a small number of record types. Understanding them is most of understanding the category.

Contacts

The individual people you deal with: name, role, email, phone. The important part is not storing the details — your phone does that. It is that a contact record can carry an owner (who in your business is responsible for this relationship) and a history.

Organisations (or companies, or accounts)

The businesses those people work for. This matters more than it first appears. If you sell to businesses, you are rarely dealing with one person — and when your main contact leaves, an organisation record is what stops the relationship restarting from zero.

Deals (or opportunities)

A specific thing you are trying to sell to a specific customer, with a value, a stage and an expected close date. The collection of open deals is your pipeline. Deals are what turn a contact list into a sales forecast.

Activities

Calls, meetings, tasks and notes — the actual work. Attached to a contact, an organisation or a deal, with a due date and an owner. Activities are how a CRM answers "what happens next", which is the question a contact list can never answer.

The timeline

The chronological record of everything above, per customer. Good CRMs make this the centre of the product rather than an audit trail, because it is what someone actually reads before a conversation.

Vendor websites tend to describe CRMs in terms of capability. It is more useful to picture the day.

1. You open the CRM and see what is overdue and what is due today: three follow-up calls and a quote to send.
2. You click the first call. The customer record opens with the last four interactions, the open deal and its value, and a note from a colleague about who signs off on budget.
3. You make the call. Afterwards you log it in about ten seconds — type, subject, one line of what was said — and book the next action while you remember it.
4. The deal has moved on, so you drag it to the next stage. The pipeline total updates.
5. At the end of the week you look at the deals nothing has happened on, and make a decision about each one: chase it, or close it out honestly.

That is it. The value comes almost entirely from step 3 being fast and step 2 being complete. Every feature that makes logging slower makes the CRM less accurate, which is why heavily-configured CRMs so often end up abandoned.

Some clarity here saves a lot of money and disappointment.

• It is not an email marketing tool. Sending campaigns to a list is a different job, and many CRMs either do it badly or charge separately for it.
• It is not an accounting or invoicing system. A CRM tracks the opportunity; your accounting software tracks the money once the opportunity becomes a customer.
• It is not a project or delivery tool. Managing the work after the sale is a different shape of problem.
• It is not a replacement for talking to customers. It is a record of having done so.
• It will not make an unclear sales process clear. If nobody can say what the stages are, the CRM will simply reflect that back at you.

There is a reasonably reliable set of signals. You probably do not need a CRM yet if none of these are true, and you are probably already losing money if several are.

• More than one person needs to know the state of a customer relationship.
• You have lost work because a follow-up did not happen, and you know it.
• Someone going on leave means their customers effectively go on leave too.
• You cannot say what the total value of your open opportunities is without building a spreadsheet first.
• You are being asked "how did that quote go?" and having to search your inbox to answer.
• You have more open opportunities than you can comfortably hold in your head — for most people that is somewhere between fifteen and thirty.

Most abandoned CRMs fail for one of four reasons, and all four are avoidable.

1. Logging costs more than it returns

If recording a call takes two minutes and eleven fields, people stop doing it during a busy week — and a CRM that is 60% current is worse than no CRM, because you cannot trust it.

2. It was configured before it was used

Custom fields, workflow rules and a bespoke stage model designed on day one, before anyone knew what they actually needed. Start with the defaults and change them after a month of real use.

3. Nobody owns anything

If records have no owner, follow-ups are everyone's job, which means they are nobody's. Ownership is the single most important field in a shared CRM.

4. It was bought for reporting, not for working

If the CRM primarily serves management reporting, the people entering data get nothing back and it decays. The reporting has to be a by-product of a tool that makes the daily work easier.

1. Export your existing contacts to a CSV. Whatever you have — a spreadsheet, your email contacts, another system.
2. Import them and accept the default pipeline stages for now.
3. Enter only your live opportunities. Not the ones from two years ago. The ones you would be annoyed to lose this quarter.
4. For every one of them, book a next action with a date. This is the step people skip and it is the step that pays.
5. Work from the CRM for two weeks. Log as you go. Do not configure anything.
6. After two weeks, change what genuinely got in your way — usually the stage names, and nothing else.

That process takes an afternoon plus two weeks of habit. It is considerably cheaper than a migration project, and it tells you whether the tool fits before you have invested anything into it.

Almost every CRM now advertises AI. It is worth being specific about what is useful and what is marketing.

Genuinely useful today: summarising a long relationship history before a call, drafting a follow-up message you then edit, and extracting details from a document for you to confirm. All of these save real minutes and none of them require you to trust the output blindly, because you read it before it goes anywhere.

Treat with caution: predicted close probabilities, automated lead scoring, and anything described as autonomous. A prediction built on a few dozen deals is not a prediction. And any system that sends messages to your customers without you reading them first is a risk you are taking on behalf of your reputation.

The useful question to ask a vendor is not "do you have AI?" but "what can your AI do without a human clicking approve?" The narrower the answer, the more carefully the product has been thought about.