CRM for small business: how to introduce one without the rollout failing

Do these five things

1. Import your contacts. One CSV, mapped, previewed, committed. Do not clean it perfectly first — you will clean as you go, on the records you actually touch.
2. Check the pipeline stages against how you really sell. Most businesses need four or five. Rename them to the words your team already uses out loud.
3. Enter every live opportunity, with a value and an expected close date. Live means you would be annoyed to lose it this quarter.
4. Book a next action, with a date, on every single one. This is non-negotiable and it is the step that creates the return.
5. Set an owner on everything. If a record has no owner, its follow-ups belong to nobody.

Do not do these

• Do not create custom fields. You do not yet know which you need, and every unnecessary field slows down logging forever.
• Do not import historical closed deals. They add no operational value in week one and they make every early report look strange.
• Do not build reports. Use what ships. Build reporting needs after you have three months of real data.
• Do not invite people who will not use it weekly. A dormant account is a permission risk and a licence cost.

The only thing that matters in this period is that logging becomes automatic. Two habits carry the whole system.

Habit one: log immediately after, not later

Ten seconds straight after the call, while you are still in the car park. "Later" means the end of the day, which means Friday, which means never. If logging takes long enough that you would rather do it later, something is wrong with the tool or with what you are asking people to record.

Habit two: never close a record without booking the next action

This is the discipline that turns a CRM from a filing cabinet into a system. If there genuinely is no next action, the deal is not open — close it out honestly.

Make the return visible early

In week two, run one meeting entirely from the pipeline board instead of a spreadsheet. People need to see the thing they are feeding actually feed them back, and this is the fastest way to demonstrate it.

Adoption in a business of three to thirty people is a social problem more than a training problem.

• Move one recurring meeting into the CRM. Whoever runs the pipeline review must run it from the board. If leadership still asks for a spreadsheet, you now have two systems and the CRM will lose.
• Do not ask for data you never look at. Every field you require and never reference teaches people that logging is pointless.
• Reduce, then reduce again. If someone is not logging, the first question is what is slow, not why they are not complying.
• Make the win-rate and follow-up numbers visible to the people doing the work, not just to management.
• Let people fix their own data. A CRM maintained only by one "data person" decays the moment that person is busy.

Do not schedule a data-cleaning exercise. They are miserable and they do not hold. Clean opportunistically instead.

1. Fix a record when you touch it. If you are opening it to make a call anyway, correcting the job title costs five seconds.
2. Deal with duplicates at creation, not afterwards. Use whatever duplicate warning your CRM offers rather than deduplicating in bulk later.
3. Archive rather than delete. A lead that went nowhere is useful history; a deleted one teaches you nothing.
4. Set one convention and write it down. Usually just how you name organisations — "Nguyen & Associates Advisory", not "nguyen advisory" and "Nguyen and Assoc".

Three months in, you have enough data to answer real questions. Resist the urge to build a dashboard of twenty metrics; four are enough to run a small business.

• Open pipeline value. The total of your open deals. Watch the direction more than the number.
• Win rate. Won divided by closed, over a period long enough to mean something — for most small businesses, a quarter rather than a month.
• Average sales cycle. How long a deal takes from creation to won. This is the number that tells you whether a quiet month was actually a slow month three months ago.
• Deals with no next action. The best leading indicator in the entire system, and the only one you can fix this afternoon.

If you also record why deals are lost, the pattern in those reasons will be the single most useful report you have — usually more useful than the win rate itself.

• Someone answers a customer question by opening the record rather than asking a colleague.
• Your pipeline review is about decisions rather than about establishing facts.
• A person going on leave hands over by sharing a filter, not by having a meeting.
• You stop maintaining the spreadsheet without having consciously decided to.
• Somebody complains that a record is out of date — which means they have started relying on it.

That last one is the real milestone. People only complain about the accuracy of systems they depend on.