Open the customer list of most Facebook and WhatsApp shops and you find one of two problems. Either there are no tags at all, or there are 140 tags like "eid offer 2", "saree new", "Saree-New" and "vip??", added by four people over two years. Both end the same way: when Eid comes, the owner sends one offer to everyone.
Everyone means the woman who only buys kids' clothes, the wholesaler from Narayanganj, and the customer who refused two COD parcels. Some of them block the number. The offer that should have made money damages the WhatsApp number instead.
This guide shows a tag system a small team can actually keep clean, how tags differ from segments, and which groups are worth messaging separately.
What is customer segmentation in a WhatsApp CRM?
Customer segmentation in a WhatsApp CRM is splitting your contacts into groups so each group gets messages that suit it. A CRM (customer relationship management system) built around WhatsApp keeps one record per phone number, with the chats, orders and labels attached to it.
Segmentation matters more on WhatsApp than on Facebook. A post on your Page is easy to scroll past. A WhatsApp message lands on the customer's phone next to messages from family. If it is not relevant, the customer can block or report you in two taps, and those signals feed your number's quality rating.
What is the difference between tags and segments?
A tag is a label on a contact; a segment is a saved audience. You use tags to record facts about people, and segments to pick who receives a campaign.
| Tag | Segment | |
|---|---|---|
| What it is | A label such as "Wholesale" or "int-saree" | A saved group of contacts |
| Who sets it | Your team, an import, store sync or an automation | You, once, when you build it |
| Changes when | Someone adds or removes the tag | Static: you edit the list. Dynamic: the rules re-check every time |
| Best for | Interests, buyer type, warnings | Campaign audiences like "opted in, seen in the last 60 days" |
A simple way to remember it: tags describe a person, segments choose a crowd.
Do not confuse contact tags with conversation labels in your inbox. A label like "COD pending" belongs to one chat and is done when the chat is done. A tag like "Repeat Buyer" stays with the customer for months.
Which tags should an online shop create?
Create tags in four groups: where the customer came from, what they want, what stage they are at, and anything your team must be warned about. Most shops need no more than 15 to 30 tags in total.
Source tags
Where the contact first came from. Useful when you want to know whether your Instagram followers or your Facebook ad crowd buy more.
src-fb-ad,src-instagram,src-website,src-shop-counter,src-referral
Interest tags
What they asked about or bought. These drive most of your campaigns.
int-saree,int-three-piece,int-kids,int-skincare,int-gift-box
Stage tags
Where they are in their life as a customer.
First Order,Repeat Buyer,Wholesale,Lapsed
Flag tags
Internal warnings. Never use these in a customer-facing message.
flag-parcel-refused,flag-wrong-address,flag-complaint-open
A naming rule that keeps the list clean
Pick one pattern and write it in your team's notes: lowercase, a short prefix, a dash, no spaces. "int-saree" and "Saree New" are the same idea, but a campaign aimed at one misses the other. If three people tag, one person owns the tag list and is the only one who creates new tags.
How do tags get onto contacts without manual work?
Tags stay accurate when most of them are added by rules, not by memory. Here are the common ways, from most to least reliable:
| How a tag is added | Example |
|---|---|
| Store sync | A Shopify or WooCommerce customer with several orders gets "Repeat Buyer" |
| Automation on a keyword | Customer writes "wholesale price?" and an automation adds "Wholesale" and assigns the chat to sales |
| Automation on an order | Order placed, add "First Order" |
| Import | A CSV of last year's Eid buyers gets "eid-2025-buyer" on every row |
| Agent in the inbox | After a chat about kids' clothes, the agent adds "int-kids" |
An automation for wholesale enquiries might look like this:
Trigger: incoming message contains "wholesale" or "paikari"
Step 1: add tag "Wholesale"
Step 2: assign conversation to Rafiq (sales)
Step 3: reply: "Paikari order er jonno amader sales team apnake ekhoni message korbe."
The agent-in-the-inbox route is the weakest, because a busy agent on Eid night will not tag anyone. Use it only for interests the automation cannot catch.
How do I build segments that are worth messaging?
Build a segment by combining who is allowed to receive a message (opt-in) with who is likely to care (activity, location, language). Then add the interest tag at send time.
Example rule-based segments for a fashion shop:
| Segment | Rules | Use it for |
|---|---|---|
| Active WhatsApp customers | WhatsApp opt-in is yes AND last seen in the last 60 days | Regular offers and new arrivals |
| Sleeping customers | WhatsApp opt-in is yes AND last seen more than 120 days ago | One win-back offer, then stop |
| Bangla speakers in Bangladesh | Country is BD AND language is Bangla | Bangla templates |
| New this month | Created in the last 30 days | A welcome message with your delivery rules |
Two habits make segments safer:
- Always include opt-in. A segment without the WhatsApp opt-in rule is a list of people who might report you. See opt-in for what counts.
- Preview before you send. If "Sleeping customers" suddenly has 6,000 people when you expected 800, a rule is wrong.
Which messages should go to which group?
Send each group the one thing it is most likely to want, and send less than you think. A rough plan for a shop selling clothing and cosmetics:
| Group | Message | How often |
|---|---|---|
| int-saree AND opted in | New saree collection with three photos | When stock lands |
| Repeat Buyer | Early access to the Eid sale, a day before everyone else | Twice a year |
| Wholesale | Updated paikari price list as a PDF | Monthly |
| Sleeping customers | One "we miss you" offer | Once |
| flag-parcel-refused | Nothing promotional | Never |
Here is what a targeted marketing message template might look like for the saree group:
Assalamu alaikum {{1}}, notun Jamdani collection eshe geche. Apni age saree dekhechilen, tai apnake prothome janalam. Dekhte "Show" likhe reply din. Offer na chaile "STOP" likhun.
The opt-out line is not decoration. A customer who can say STOP does not need the block button.
What mistakes ruin a tag system?
The usual mistakes are too many tags, no owner, and using tags as a dumping ground for campaigns.
- One tag per campaign. "eid offer", "eid offer 2", "eid final" tell you nothing a month later. Track campaigns in your campaign reports, not in tags.
- Tags as the only opt-in record. Opt-in should live in the opt-in flag, not in a tag called "ok to message".
- Duplicate contacts. If the same buyer exists three times under different spellings, tags split across the copies. A CRM that matches on phone number fixes this at the source.
- Never deleting. Review the tag list once a quarter. Merge near-duplicates and delete tags with almost no contacts.
How does SaleMag handle tags and segments?
SaleMag's CRM keeps one contact per phone number in each workspace, with coloured tags, custom fields and separate WhatsApp, SMS and email opt-in flags. Tags come from your team, CSV or Excel imports (you can tag every imported row), store sync from Shopify, WooCommerce or BigCommerce, and the add-tag and remove-tag steps in the automation flow builder.
Segments are static lists or dynamic rules combined with AND or OR, on fields such as country, language, source, opt-ins, created date and last seen date. Segment rules do not filter by tag; instead, you pick a tag directly as a broadcast audience. Broadcasts skip contacts without WhatsApp opt-in automatically. SaleMag's AI chatbot answers customers but does not tag them by itself, so set up automations for the tags you care about.
Next steps
Write your four tag groups on one page this week, clean up the old tags, and set one automation for wholesale enquiries. Then read how to send a WhatsApp broadcast without getting blocked and how to recover abandoned carts on WhatsApp. When you are ready to try it with your own list, start free.
Step by step
- 1
Write down your tag groups
Decide on four groups (source, interest, stage, flag) and a naming pattern such as 'int-saree' or 'flag-parcel-refused'.
- 2
Clean the existing list
Import your customer sheet with tags applied to every row, merge duplicates by phone, and delete tags nobody uses.
- 3
Automate the obvious tags
Add automation rules that tag a contact when they send a keyword like 'wholesale' or place an order.
- 4
Build segments for campaigns
Save rule-based segments such as 'opted in and active in the last 60 days' and preview the audience before each send.
Frequently asked questions
What is the difference between a tag and a segment in a WhatsApp CRM?
A tag is a label on a contact, such as 'Wholesale' or 'Kids wear'. A segment is a saved audience, either a fixed list or a set of rules like 'country is Bangladesh and WhatsApp opt-in is yes', that you use for campaigns and automations.
How many tags should a small online shop use?
Most small shops do well with 15 to 30 tags. If a tag has fewer than a handful of contacts after a month, merge it or delete it.
Can I send a WhatsApp broadcast to one tag only?
Yes. In SaleMag you can pick a tag, such as 'Repeat Buyer', as the audience of a WhatsApp broadcast, and only contacts with that tag who have opted in to WhatsApp receive it.
Can a SaleMag segment rule filter by tag?
No. SaleMag segment rules work on contact fields such as country, language, source, opt-ins, created date and last seen date. To target a tag, choose the tag directly as the campaign audience.
Does the AI chatbot tag customers automatically?
No. In SaleMag the AI chatbot answers customers but does not tag them. Tags come from your team, imports, store sync or an automation's add-tag step.
Should I tag customers who refused a COD parcel?
Yes. A private flag tag such as 'flag-parcel-refused' helps your team ask for confirmation or an advance before the next order. Keep it internal and never use it in a customer-facing message.
About the author
WhatsApp & F-commerce automation, Dhaka
The Dhaka team that builds SaleMag, the AI sales inbox for WhatsApp, Messenger and Instagram used by online sellers in Bangladesh.