Guide

Gmail labels are not a CRM

September 15, 2026 · 9 min

Labels answer “does this belong to Acme?” A CRM answers “what is in play, what it is worth, and what happens next.” Those are different jobs.

The most common Gmail CRM is a nested label tree: Clients/Acme, then Sales/Negotiation, then a star because you will “deal with it on Thursday.” It looks organised. It is a filing cabinet. Filing cabinets do not forecast.

A CRM is a picture of deals. Each deal has a stage, a number, a date you can miss, and a next action in one sentence. Gmail labels cannot hold any of that. They were never designed to. Pretending they can is why people bounce to a spreadsheet in March and abandon the spreadsheet in April.

This is not an argument against labels. Use them. Search is faster when Acme has a colour. The argument is that labels are taxonomy, not pipeline — and sales needs a pipeline.

What a label can actually tell you

A label is a boolean. The thread has it or it does not. Nested labels add a folder illusion: you can pretend Sales/Hot is a stage. Gmail still treats it as a tag. Nothing stops you from leaving a closed deal in Hot for six months. Nothing totals the money. Nothing asks what you promised to send.

Multiple labels on one thread make it worse, not better. The same email can be Client, Q4, Waiting, and Urgent. That is how people file. It is not how people sell. When you open the inbox on Tuesday, you still have to reconstruct the deal from the last three messages.

Stars and Important are even thinner. They mean “I felt something when this arrived.” Mood is not a stage.

Five things labels cannot store

If you sell by email, you keep asking the same five questions. Labels have no field for any of them.

Company name, a private note, and a link back to the thread belong on a card for the same reason. They are attributes of the deal, not of the message.

The nested-label trick, and why it fails

Power users try this: Pipeline/Lead, Pipeline/Contacted, Pipeline/Negotiation, plus a per-client label. Filters auto-apply the client tag. You promise to move the pipeline label by hand.

Three weeks later the board in your head and the labels on the threads have diverged. Moving a label is two clicks and a search. Dragging a card is one gesture you can see. The visual is the product: if you cannot see every live deal in one glance, you do not have a CRM. You have a search query.

Filters also cannot decide what a deal is. Auto-labelling every message from a domain dumps newsletters, invoices, and “quick question” threads into Sales. Then you stop trusting the folder. Trust is the whole system.

Two deals, one client: the real test

Imagine Marta at Vidal. You are closing a €950 workshop for October. In the same week she asks about a separate Q1 retainer. In Gmail that is often one thread, or two threads with the same labels.

A CRM needs two cards: different values, different close dates, different next steps. Labels collapse them. Spreadsheets can split them, but then you copy the sender twice and you are no longer in Gmail. The honest design is one card per conversation you refuse to lose, with fields Gmail will not invent for you.

If you only ever have one live thing per company, you can fake this for a while. The first time a client buys twice, the filing system lies.

A setup that keeps labels and adds a pipeline

You do not have to delete your label tree. Run two systems with two jobs.

  1. Keep client labels for search. Apply them how you already do.
  2. Stop using labels as stages. If Sales/Negotiation exists, archive the habit, not necessarily the label.
  3. Put every live deal on a board with five columns: Lead, Contacted, Negotiation, Won, Lost. Rename later if you must.
  4. On each card: company, value (empty is honest), expected close, next step in one line, a note if there is a trap (“invoice the parent company”).
  5. Link the card to the thread. The email stays the record. Do not paste the thread into notes.
  6. Start the morning on the board, not in Unread. Move cards. Then answer mail.

Only threads you pick go on the board. Do not ingest the mailbox. Newsletters are not deals. A robot that “creates contacts from every sender” is how CRMs fill up with junk and people stop opening them. You already know which threads hurt to lose.

What “good” looks like on a Tuesday

You open Gmail. Beside the inbox — or one click from the thread — you see columns. Negotiation has three cards and a total. One close date is this week. The next step on that card is “send revised SOW.” You do that before you clear Unread.

That picture is the job. Labels cannot draw it. A full Salesforce instance can, and should, if you have a team that lives there. For one to ten people who close in Gmail, the board should sit where the mail already is.

KanbanMail is a free Chrome extension (and a Gmail add-on) that injects that board into Gmail. You send a thread to a column. We store the contact, the thread link, and the fields you type. We do not scan the mailbox and we do not pull Gmail through our server. If you want the wider method, start with how to use Gmail as a CRM. If the columns themselves are the open question, read how to build a sales pipeline inside Gmail.

FAQ

Can Gmail labels work as a CRM?

No. A label tells you a thread belongs to a client. It does not store deal value, an expected close date, a next step, or more than one live deal with the same company.

What should I use instead of nested sales labels?

A board with one card per deal, a stage you can drag, and the fields Gmail will not show in a list: value, close date, next step and notes. Keep the thread as the record.

Should I still label sales mail?

Yes, if labels help you search. Filing and forecasting are different jobs. Keep labels for search; put the pipeline on cards.

See the pipeline labels cannot show

Free Chrome extension. One click from the thread to a card with value and a next step.

Install on Chrome — free