Customer Conversations in One Place

No affiliate links on this site, and nothing here is ranked.

What the task actually is

Anybody in the firm being able to see what was said to a customer and what was agreed.

Not pipeline management. Not forecasting. Not campaign automation. Those are the features the category is built around, and they belong to organisations with a sales function.

The task most small firms have is that a customer rings, somebody who was not on the previous call answers, and the previous conversation is in another person's head.

The four sizes

One person, few customers. Email plus a note per customer. Where it stops: anybody else needing to see it, and anything you need to search across.

Two or three people, shared mailbox. Everybody sees the correspondence. Where it stops: anything said on the telephone, and anything that is a decision rather than a message.

A shared record per customer — a document, a sheet, a simple database. Where it stops: volume, and linking conversations to work.

And a real system, where the record is the working surface rather than a place to write things down afterwards.

The fourth is where the category earns its price, and the honest test is whether anybody currently updates the third.

What makes the record fail

Partial adoption, which is the constraint on everything.

A customer record holding two thirds of the conversations misleads whoever reads it, which is worse than an acknowledged absence — because the reader believes they are informed.

Entry burden falling on the person who gets nothing back. The person taking the call types the note; the person reading it is somebody else. That arrangement is not stable, and it is the reason most small-firm systems decay within a year.

And recording the wrong thing. A log of every email is noise. What is worth recording is decisions, promises and anything unusual — three or four lines per customer per month, not a transcript.

What to check before buying

Whether it captures what actually happens. Most conversations in a small firm are telephone calls and messages, not emails, and a tool built around email threads records the wrong half.

Whether it works on a phone, badly or well, since that is where a call gets noted or does not.

What comes out. Customer records are the longest-lived data a small firm holds, and a list of contacts without the correspondence is not the same thing.

And who else sees it. Customer information carries obligations that do not depend on the size of your business.

The version that works before buying anything

One document per customer, in the shared drive, named consistently.

Date, what was discussed, what was agreed. Three lines.

It is searchable, everybody can use it, it exports trivially and it costs nothing — and it fails at exactly the point where volume makes finding things slow, which is a threshold you will notice.

Where it works, that is the answer. Where it does not, you now know precisely what failed, which is a specification rather than a category.

Why a CRM gets bought

Rarely because the record failed.

Usually because somebody left, and the firm discovered that what they knew about their customers left with them. That is a real and expensive discovery, and the response is correct even if the tool chosen is oversized.

Or because a customer was embarrassed — asked something they had already explained, chased for something already paid.

Or because a bank, an investor or a buyer asked what the customer base looks like and nobody could answer.

The first three are record problems and a shared document solves them. The fourth is a reporting problem and it is the one that genuinely wants a system.

Knowing which prompted the search changes the size of the answer, and it is worth asking before looking at anything.

The pipeline question

Most small firms are sold pipeline management with the record they wanted.

A pipeline is a forecast — how much work is likely to close and when. It is useful where the answer changes a decision: hiring, ordering, committing to something.

Where nothing would change, it is a set of fields somebody has to keep current for no return, and it is the first thing to stop being updated.

Ask what you would do differently if you knew. If the answer is nothing, the pipeline is decoration and the record is the product.

Migrating from nothing

The good case, and it is worth doing deliberately.

Do not import everything. Years of old email into a new system produces a system full of noise on day one.

Start with active customers, three lines each, written from memory. An afternoon.

And let history stay where it is. The old email is still searchable and it does not need to be inside the new tool to be findable.

The obligation nobody mentions

A record of customer conversations is personal data, and that carries requirements which do not scale with the size of the business.

Where it is held, who can see it, how long it is kept, and what happens if somebody asks for a copy or asks for it to be deleted.

A shared document has the same obligations as a purchased system — the tool does not create them and does not discharge them.

Which is an argument for deciding deliberately where it lives, rather than letting it accumulate in whichever mailbox happened to receive it. Additional context is available from Trello.

The short version

• The task is anybody being able to see what was said and agreed, which is not pipeline management
• Four sizes: notes plus email, a shared mailbox, a shared record per customer, and a real system where the record is the working surface
• Records fail through partial adoption, through entry burden on somebody who gets nothing back, and through logging everything instead of decisions
• Three or four lines per customer per month is the useful volume; a transcript is noise
• Check whether it captures telephone calls, whether it works on a phone, what exports, and who else sees the data
• Start with one document per customer in a shared drive; where it fails, the failure is a specification