CRM tags and segments for messenger sales

Tags in a CRM are only useful when they answer «who will I message later». A tag is a fact you attach by hand, and a segment is an automatic selection built from tags and fields.

Eight to fifteen tags is enough for a shop of any size. Tagging systems rarely fail from having too few. They fail from having too many: once the list is longer than the screen, people pick whatever’s nearest, and the data stops meaning anything.

What principle should you invent tags by?

Don’t describe the customer — describe the future action. Before creating a tag, ask one question: «will I ever send a message or a selection to everyone carrying this?»

1. 1

   Works

   «Bought footwear», «asked about wholesale», «wanted instalments» — each implies a future message.

2. 2

   Doesn't work

   «Nice», «difficult», «regular» — a judgement, not a fact. Next month someone applies it differently.

3. 3

   Actively harmful

   «VIP» with no criterion. Everyone interprets it their own way and the segment comes out random.

Which tag categories does a shop need?

Three groups cover most needs:

What they bought. Product category, not the specific item. «Footwear», «accessories», «care». This is the basis for upsells and seasonal selections.

How they buy. Wholesale, instalments, collection, prepayment only. These save time in every subsequent conversation.

Why they didn’t buy. Too expensive, size unavailable, still thinking, waiting for restock. The most underrated group: it’s what gives you a reason to return with concrete news instead of «how are you doing?».

How do tags become segments?

A segment is the intersection of two or three conditions matching one specific message. A single tag isn’t a segment.

Working examples:

• «Bought footwear» + last contact over 3 months ago → new-season message.
• «Size unavailable» + category tag → message on the day it lands.
• «Asked about wholesale» + didn’t buy → a separate offer with terms, not a general broadcast.

Each of those has a specific message behind it that, without tags, you’d have to send to everyone — losing the relevance that made it work.

How do you stop the system collapsing?

Two rules. First: tags get applied during the conversation, not «later when there’s time». Second: once a quarter, open the list and delete everything that was never used.

The second point is unpleasant and the most important. Tag lists grow on their own; nobody ever deletes anything voluntarily — and it’s the sprawl, not any single wrong tag, that kills the system.

A twelve-tag starter set

What they bought: category 1, category 2, category 3 (substitute your own)

How they buy: wholesale, instalments, collection, prepayment only

Why they didn’t buy: too expensive, size unavailable, waiting for restock, thinking

Special: delivery went wrong (so next time you approach differently)

Twelve tags, each applicable in a second, each with a future message behind it.