Getting the Structure Right Took Longer Than Building the Site

Some products only become credible after the early versions are discarded. The first pass can prove effort, but it often fails to prove structure. When the shape is wrong, shipping faster mostly locks the wrong assumptions in place.

UK Shortlists was built this way: started, paused, rebuilt, and simplified more than once. Different routes and page models were tried, then removed. The hard part was not getting pages live. The hard part was deciding what the site should be strict about and what it should ignore.

The useful outcome was a category-first architecture with shortlist logic that makes comparison clearer. That did not appear in a single attempt. It came from abandoning weaker structures, reducing overlap, and making tighter calls on what deserved to stay.

That is why repeated restarts are not automatically a warning sign. If each restart removes ambiguity and improves the decision model, the product is getting more believable, not less focused.

A better structure is often earned through replacement, not defended through momentum.