The quality of your database sets the ceiling on your results. The best marketing message fails against a random list of numbers, and a simple message succeeds against a well-segmented audience. This article covers the part most people rush: organising contacts.
Importing from Excel
Mersal imports Excel files with two important steps that happen automatically:
- Number cleaning: normalising format (spaces, dashes, leading zero, country code).
- Upsert: an existing number is updated rather than duplicated — so the list never doubles up.
These two solve the biggest problems with imported lists: duplicates, and badly formatted numbers that fail at send time.
Preparing the Excel file properly
An import succeeds or fails on the file itself. Before uploading:
- One header row at the top: not two, and no merged cells.
- Phone column formatted as text, not number: Excel strips the leading zero when a column is formatted as a number.
- Country code: best written explicitly (
+20). If it's missing, set the default country on the import screen. - One column per fact: "Ahmed - Cairo - wholesale" in a single cell cannot be split.
- Remove total rows: a summary row at the bottom of the file becomes a very strange contact.
Always trial a small file (20 rows) first to confirm the column mapping is right before uploading 5,000.
Custom attributes
Name and number aren't enough. Mersal supports 12 custom attribute types you can add to contacts: text, number, date, select list and more. Practical examples:
- City — to send a branch-specific offer only to that area.
- Last purchase date — to target people who haven't bought in 3 months.
- Customer type (wholesale / retail) — so price lists differ.
- Birthday — for an automatic greeting with an offer.
These attributes aren't just for display: you use them as personalisation variables inside messages, and as conditions in automation.
Choose attributes carefully
The common mistake is defining 20 attributes on day one and finding 17 of them empty two months later. The only test: will I use this to target, or inside message text? If not, don't define it.
An attribute that's missing on half your contacts is worse than not having it, because a campaign filtering on it silently excludes half your audience.
Groups
Groups are the core targeting unit. There's a default "My customers" group, and you can create groups and move contacts between them. The practical rule: one group per marketing intent, not per data source.
So "wholesale customers" and "no purchase in 3 months" are sensible groups, while "October 2025 import" is not a group — that's a fact belonging in an attribute.
Segments that actually work
- By recency: bought in the last 30 days / 90 days / over a year ago. The single most valuable split there is.
- By value: your top 20% of spenders deserve entirely different messages.
- By geography: essential if you have branches or differing shipping.
- By preferred channel: who replies on WhatsApp versus who opens email.
Bulk actions and archiving
Search, export, bulk actions and archiving. Archiving matters: instead of deleting an inactive contact and losing its history, you archive it — it leaves campaigns while the data stays.
The Leads Hub
Leads you extract land in a separate hub with review, normalisation and validation before they become real contacts. That barrier is important: it stops raw, unreviewed data from polluting your clean database.
A 30-minute cleanup plan
- Export your current list and count duplicates and numbers missing a country code.
- Define just 3 custom attributes — the ones that matter most, not everything.
- Create 3 groups by marketing intent and move contacts in.
- Archive any contact with no engagement in a year.
Frequently asked questions
What if I import the same file twice? No duplication — existing numbers are updated rather than added again.
Can a contact be in more than one group? Yes, and it's normal — a wholesale customer in Cairo belongs to both.
Are invalid numbers removed automatically? They're flagged at import time, and you can review them before committing.
