Do we need a separate CRM and inbox?

Not if the inbox writes to the customer record. The thing that makes a CRM useful at this size is that the history is already there when somebody picks up, and that only happens if conversations land on the contact without anybody copying them across. Two systems means somebody is the copying, and at five people that somebody is whoever is least busy, which means it stops in the first busy week. See CRM vs inbox for the longer version of this argument.

When should we add more structure?

When the same question keeps arriving that the current setup cannot answer, and not before. Structure added in advance of a question is a guess, and the cost of a wrong guess is not the setup time, it is that five people now maintain a field nobody reads. The useful trigger is hearing yourself say I wish I knew how many of these were X for the third time in a month.

One person, usually the office manager, and the job is smaller than it sounds: keep the tag list short, decide when a new field is justified, and close stale leads. Without a named owner the tag list grows a synonym a month until nobody can filter on anything, which is the specific way a small team's CRM dies.

Should a part-time helper get a login?

Yes, their own, on the lowest role that lets them work. Shared logins are the common shortcut at five people and they cost you the two things the record is for: you cannot tell who replied to a customer, and you cannot remove one person's access when they leave without changing everybody's password. A viewer login costs nothing.

What is the difference between a field, a tag and a note?

A field is a fact with one value per customer that you want to filter or act on, like the plan they are on. A tag is a label a customer either has or does not, useful when you want a list of everybody like this. A note is a sentence for a human to read later. Most teams reach for a field when they mean a note, which is how you end up with twenty fields that are each filled in for nine customers.

A CRM for a team of five: what to use and what to skip

Five people need a customer record, a lead list where every lead has an owner, and follow-up tasks. What to set up in week one, the four things to refuse, and how to give each person the least access that lets them work.

Sohail Safdar · November 9, 2026 · 5 min read

Five people, one phone line, and a shared inbox that everybody can see and nobody owns. Somebody asks on Thursday whether the lady about the extraction was ever called back, and three people check three different places: a sticky note, a text thread on a personal phone, and the inbox, where the message is sitting read and unanswered because each of them assumed one of the others had it.

That is the problem a CRM solves at this size. Not pipeline visibility. Not forecasting. Knowing who has the ball.

The three things that earn their keep

One record per customer, with every conversation on it. The test is whether somebody who has never met this customer can pick up the thread and sound informed. If the history lives in one person's phone, the business has five front desks rather than one.

A lead list where every lead has an owner. A name next to each row, visible, assigned when it arrives rather than at a weekly meeting. An unowned lead is not a lead that everybody is working, it is a lead that nobody is working, and the failure is silent because each person has a reason to think it is handled.

Follow-up tasks with due dates. A callback is a task, not an intention. The value is the overdue list: it is the only artefact in a small business that reliably says what slipped, because people do not volunteer it.

The arithmetic that justifies the hour of setup

Say the five of you take 25 new enquiries a week. Say that without owners, two of them get no reply at all, and another three get one slowly enough that the customer has gone elsewhere. That is five lost a week. If a customer is worth $600 over the time they stay with you, that is $3,000 a week, which is a number that makes a weekly fifteen minutes look like the cheapest thing on the list.

Your numbers will differ. Count your own for two weeks before deciding this is overstated: the count is the point, because "we are pretty good at getting back to people" is exactly the belief the overdue list exists to test.

What to skip, and why it is not laziness

Multi-stage pipelines with probabilities. A probability on a stage is a forecast, and a forecast needs volume to mean anything. At 25 enquiries a week the noise is larger than the signal.
Revenue forecasting. Same reason, plus it invites arguments about stage definitions that produce no bookings.
Lead scoring you cannot explain in a sentence. If nobody can say why this lead is a 72, nobody will act differently because of it, and an unexplainable number that changes no behaviour is a cost with no benefit.
Custom objects and deep integrations you will not maintain. The question is not whether it would be nice. It is who fixes it in March.
Dashboards nobody opens. One number somebody looks at beats nine nobody does.

The common thread: at five people, every piece of structure has a maintenance cost paid by somebody who also answers the phone.

Roles, and giving each person the least that works

Person	Role	Sees
Owner	Owner	Everything, including billing
Office manager	Admin	Everything except ownership
Front desk	Agent	Conversations, contacts, leads, bookings
Practitioner	Agent or viewer	Their own schedule and the customers on it
Part-time helper	Viewer	Read only

Frontiva's roles are owner, admin, manager, agent and viewer, and there is exactly one owner per business, because "who can close the account" is not a question that benefits from two answers.

The manager role is the one worth understanding before you use it, because it is narrower than it sounds: a manager belongs to exactly one location. The routes that create one require a single location to be named and refuse a second active manager for the same branch, and what a manager reads is narrowed to that branch rather than filtered in the interface. For a single-site team of five, admin and agent are usually the whole answer, and manager starts mattering on the day you open the second place.

The weekly fifteen minutes

Open the lead list together, once, at a fixed time.

Anything without an owner gets one, now, out loud. Anything overdue gets a decision rather than a sympathetic noise. Anything stale gets closed as won or lost with a reason, because a lead list that only grows stops being read within about six weeks, and then the overdue count is meaningless.

What not to do with this meeting: review every lead. Five people reading 25 rows is an hour, and it will be cancelled by the third week. Read the exceptions.

Tags, fields and the mess they become

Keep a short tag list with written meanings, agreed once. The failure mode is not complexity, it is synonyms: one person tags "VIP", another "vip", a third "priority", and the filter that was supposed to produce a list produces a third of one.

Use a field when the thing is a fact with one value per customer that changes what you do: the plan, the referral source, the vehicle. Two to five of them. Everything else is a note, and a note is fine. See custom field examples by industry for the field, tag and note distinction drawn out.

In Frontiva

The CRM and lead management cover the three things above: one contact timeline carrying every channel, leads with an owner and a visible score, follow-up tasks with an overdue view, and tags and custom fields with the short list kept short by somebody.

What matters more than the feature list at this size is that the inbox and the record are the same thing rather than two systems that sync. A conversation that arrives by text, web chat or phone lands on the contact, so the Thursday question has an answer without anybody having copied anything.