Hi! I run an Ethiopian coffee business selling packaged roasted coffee beans and ground coffee. Online orders come through my Shopify store, while other customers use a form on my website or send me WhatsApp messages.
Right now I copy everyone into a spreadsheet by hand, and it's a mess. Same customer shows up two or three times, WhatsApp orders never make it in, and I can't tell my best buyers from one-time visitors.
I'm looking for someone to put all of this into one clean, deduplicated contact list automatically, ideally sorted by how valuable each customer is. Thanks!
First, I’ll make a safe copy of the customer information you already have and clean it separately. You and your team can keep taking orders as usual while I work.
Next, I’ll bring together the Shopify, website, and WhatsApp records, remove duplicates, and check that each customer’s details and order history are correct. Any new orders placed during the setup will be included before the new system goes live.
n8n is the best solution for this type of task. It already handles this kind of connection and routing well, and the workflow is visual, so you can follow what happens at each step. That also makes it easier for someone else to take over, maintain, or change later. The goal is a dependable system you can own, not clever code that only its original developer understands.
The final result is one reliable customer list: each person appears once, their information is up to date, and your best customers are easy to identify. I’ll also keep the original data as a backup.
Even a well-built workflow will run into messy data and the occasional outage. Here’s how I’d plan for the situations most likely to cause trouble.
Say Shopify does not hear back quickly enough, so it sends the same order again. If the workflow treats it as new, you end up with two copies of the order and possibly two customer records.
I’d check Shopify’s order ID before saving anything. If that order is already there, the workflow updates it instead of creating another copy.
Someone orders through Shopify with one email, then messages on WhatsApp from a phone shared with their business. The records might belong together, but there is not enough evidence to be sure.
I would leave both records untouched and send them to a review queue. You can compare them side by side and merge them only when the match is clear.
A regular customer writes, “Send me the usual,” without mentioning a product, quantity, email, or delivery address.
The workflow would keep the original message, point out what is missing, and put it in the review queue so someone can follow up instead of letting the system guess.
Airtable hits a rate limit or briefly stops responding just as a new order comes in.
I’d have the workflow wait and try again a few times. If it still cannot connect, it keeps the order safe, records what failed, and sends an alert so nothing disappears quietly.
The contact reaches the CRM, but the customer score or Slack notification fails. Running the whole workflow again could create a duplicate.
I’d record where the workflow got to, then restart it from the failed step. The completed work stays in place and the contact is not added twice.
A loyal wholesale buyer orders less often but spends much more. A simple order-count rule could rank them below someone making frequent small purchases.
I’d test the scoring rules against customers you already know, keep the calculation easy to inspect, and let you adjust a score when the numbers miss the bigger picture.
Paste the JSON into a blank n8n workflow, then connect your own Shopify, WhatsApp, Airtable, OpenAI, and Slack accounts. The routing, deduplication, review, retry, and alert structure comes with it.
Explore next
Explore essays about software, AI, attention, and the systems shaping everyday life.