Key takeaways
- Repair when the shop works but is slow, awkward or dated
- Rebuild when the catalogue model cannot express what you now sell
- Write down the three changes you want next year, then ask if the platform blocks them
- Blocked means rebuild; merely tedious means repair
- A working set of indexed URLs with history is an asset a rebuild puts at risk
- A new front end on the existing back end is often the right middle path
Rebuilding feels decisive. It also resets the clock on every problem you have not diagnosed, and brings a migration risk you did not have before.
Sometimes it is still right. Here is how to tell.
Repair is usually right when
- The shop works, but it is slow, awkward or visually dated
- The catalogue model fits how you sell, even if the interface does not show it well
- The problems are identifiable and bounded, such as a slow category page or a broken filter
- You depend on search traffic and the current URLs are performing
That last point carries more weight than most people give it. A working set of indexed URLs with history is an asset, and a rebuild puts it at risk.
Rebuild is usually right when
- The catalogue model cannot express what you now sell
- The platform blocks a commercial change you have already decided to make
- The codebase is unmaintained to the point that changes are unsafe
- Security or platform support has genuinely lapsed
No amount of repair fixes a product model that cannot describe your products.
The question that settles most cases
Write down the three changes you most want to make in the next twelve months. Then ask whether the current platform blocks them, or merely makes them tedious.
Blocked means rebuild. Tedious means repair, and usually for a fraction of the cost.
The hidden cost of rebuilding
A rebuild is not just the build. It is:
- The migration of products, customers, orders and reviews
- The URL map and redirect testing
- Staff relearning an admin they already knew
- A period of elevated risk around launch
Those are real costs and they are routinely left out of the comparison.
A middle path people forget
You can rebuild the storefront without replacing the commerce engine. If the admin, the catalogue and the order flow work, but the customer-facing site is the problem, a new front end on the existing back end gives most of the benefit with a fraction of the migration risk.
This is frequently the right answer and it is rarely the one proposed, because it is less work to quote for a full rebuild.
Frequently asked
Questions readers ask
Should I rebuild my ecommerce site or fix the existing one?
Write down the three changes you most want to make in the next twelve months, then ask whether the current platform blocks them or merely makes them tedious. Blocked means rebuild. Tedious means repair, usually for a fraction of the cost.
What are the hidden costs of rebuilding?
Migrating products, customers, orders and reviews; producing and testing the URL map; staff relearning an admin they already knew; and a period of elevated risk around launch. These are routinely left out of comparisons.
Can I rebuild the front end without replacing the platform?
Often, and it is frequently the right answer. If the admin, catalogue and order flow work but the customer-facing site is the problem, a new front end on the existing back end gives most of the benefit with far less migration risk.





