I initially considered combining three related sourcing workflows into one product.
On the surface, they looked similar:
Find a product.
Organize the information.
Prepare it for sale.
But the more closely I mapped the workflows, the more different they became.
One workflow focuses on preparing products from 1688 for Coupang and Naver Smart Store.
Another deals with multiple Chinese sourcing platforms, where access methods, product data, and failure patterns are different.
The third explores safety checks before a seller moves toward an actual listing.
Each workflow has different:
• platform constraints
• data requirements
• error conditions
• policy risks
• approval boundaries
Combining them would make the product look larger.
But it would also make responsibility less clear.
In ecommerce operations, automating the wrong step does not simply produce an inconvenient result.
It can create incorrect listings, duplicated products, policy violations, or account risk.
So I decided to separate the product boundaries while keeping the underlying logic connected.
Two workflows are now functional MVPs under validation.
The third remains a concept and validation scope until its safety boundaries are clear enough.
This reminded me that product architecture is not only about grouping similar features.
Sometimes, good architecture means knowing what should not be combined.
When do you decide that related workflows belong in separate products rather than one larger platform?
#ProductArchitecture #EcommerceOperations #SystemsThinking
Great post! I'm currently setting up an intercity transport platform in the US market and looking to connect with mobility tech founders.).
Thanks, Branko. Your intercity transport platform sounds interesting.
My work is broader than mobility tech, but I focus on turning complex workflows and decision points into structured digital products. I’d be interested to hear what part of my post connected with what you’re building, and what stage your platform is currently at.
The workflow differences justify separate boundaries, but the stronger signal is user behavior. Have customers actually preferred one workflow without needing the others bundled?
Not yet—and that is exactly the next assumption I need to test.
The current separation came from operational differences and risk boundaries, not from confirmed customer preference.
I now want to observe whether users:
• complete only one workflow repeatedly
• naturally move between two or more workflows
• ask for a single bundled experience
• value separation because it makes each step clearer and safer
If most users need all three in sequence, the products may remain separate internally while being presented through one unified interface.
If user behavior consistently clusters around individual workflows, that would support keeping them as separate products.
So for now, I see the product boundaries as a working hypothesis, not a validated conclusion.
The workflow-clustering test is the key one now. If you’re open to it, what’s the best email to reach you on?
I’m open to discussing it. You can reach me at connect.geniusbrain@gmail.com.
I’d be especially interested in your thoughts on how you would structure the workflow-clustering test.
Thanks! I’ve just sent it over.
Looking forward to hearing your thoughts whenever you have a chance.
Thanks. I haven’t received the email yet, including in my spam folder.
Could you please confirm that you sent it to connect.geniusbrain@gmail.com?
I've just resend it to same address. please check now
I checked your email and sent you a reply.