|
What this covers
|
A business commissions a new website. The design is better in every visible way. Six weeks after launch, inquiries have halved, and nobody can explain it because nothing on the new site looks broken.
This outcome is common enough to be predictable, and the cause is almost never the design. It is the addresses.
Addresses and Pages Are Two Different Assets
A website is made of pages, but search engines and other websites do not refer to pages. They refer to addresses. The distinction sounds academic until a rebuild, at which point it decides whether the project gains or loses.
When another site links to a business, that link attaches to a specific address. When a search engine ranks a page, the ranking is recorded against that address. Years of accumulated signals sit on the address, not on the words displayed at it.
Ranking signals accumulate against an address over time. That is the sentence a redesign has to be planned around.
What Actually Changes in a Rebuild
Most redesigns change the structure as well as the look, because the old structure grew by accretion and the new one is planned. That is usually an improvement for visitors and a hazard for search.
|
What changed |
What search engines see |
Effect if unmapped |
|
Page redesigned, address unchanged |
Same address, new content |
Minimal. Rankings usually hold |
|
Address changed, redirect in place |
Old address points to new one |
Most signals carry across |
|
Address changed, no redirect |
Old address returns an error |
Signals stranded, rankings lost |
|
Page removed entirely |
Address returns an error |
Any links pointing at it are wasted |
|
Address changed, redirect to the homepage |
A redirect to something unrelated |
Treated as a soft error, little carries |
The fourth and fifth rows are where most of the damage happens, and the fifth is the one people are surprised by. Redirecting every retired address to the homepage looks tidy and accomplishes very little, because the destination does not answer what the old address answered.
Why Nobody Notices for Six Weeks
Search engines do not re-crawl a site the moment it changes. They arrive on their own schedule, find the new addresses, find errors at the old ones, and adjust over the following weeks. The decline is gradual, which is exactly what makes it hard to attribute.
By the time inquiries are visibly down, the launch is a month and a half back. The design has been praised by everyone who has seen it. The connection is not obvious, and often the redesign gets credit for being fine while the business quietly assumes demand has softened.
The Mapping Step
Preventing it is not technically difficult. It is a list, made before the build rather than after.
- Export every existing address on the current site, including old pages nobody remembers
- Note which of those addresses currently receive search traffic or have links pointing at them
- For each one, decide the closest equivalent address on the new site
- Where a close equivalent genuinely does not exist, decide deliberately whether the address retires or redirects to the nearest parent
- Test every redirect after launch rather than assuming the rule worked
The fourth point is the one that gets skipped. A redirect pointing somewhere unrelated is not better than a clean retirement, and pretending otherwise produces a site full of misleading destinations.
The Page That Earns the Most Gets Touched Least
There is a temptation during a rebuild to standardize everything, including the address of the page that currently performs best. The new structure is more logical. The address would match the pattern. It is the correct decision in every respect except the one that matters.
A page that already ranks has accumulated the signals that make it rank. Moving it introduces risk for a tidiness benefit no visitor will ever perceive. The reasonable rule is to leave the best-performing addresses exactly where they are, even when they break the naming convention, and accept the inconsistency as the price of not gambling.
Titles and Headings Carry Information Too
Addresses get the attention because their failure is dramatic. A quieter version of the same problem happens in the text.
When a rebuild rewrites every page title and heading in a new house style, pages that ranked for specific phrasing can lose it. The new wording may be better prose. Search engines were matching the old wording to what people typed.
The workable approach is additive: keep the phrasing that is earning, and add the new material around it rather than replacing it. Where a page is not earning anything, the writer has a free hand.
Checking the Work After Launch
|
Check |
What it tells you |
When |
|
Crawl the old address list |
Which redirects fire and which error |
Launch day |
|
Search Console coverage report |
Which addresses search engines now consider valid |
Week one and week four |
|
Compare traffic by page, not in total |
Which specific pages lost, rather than whether the site did |
Weeks two to six |
|
Confirm the new addresses are indexed |
Whether replacements are actually eligible to rank |
Week two |
Comparing totals is the trap in that table. A site-wide figure can look stable while the three pages that generated inquiries have collapsed and a blog post has picked up incidental traffic that replaces the number without replacing the value.
What This Means for a Metro Market
In a city with real competition, the recovery window matters more than it does elsewhere. Positions vacated during a rebuild get filled, and the businesses that filled them do not give them back quickly.
That is the argument for treating the mapping step as part of the build rather than a technical afterthought. An agency handling web design in Kansas City should be able to produce the address list before the first mockup, because it is also the document that decides the new site’s structure. The same reasoning is set out in a guide to improving local rankings, and the relevant business profile shows the market it is written from.
Staging, and Why the Old Site Should Stay Reachable
A rebuild that replaces the live site in one action removes the ability to check anything against the original. Once the old pages are gone, questions about what a page used to say, which heading it used or whether an address existed at all become unanswerable.
Building on a staging address and switching over solves this, provided the staging copy is kept for a few weeks afterward rather than deleted on launch day. It costs nothing to retain and it is the only reference available when a page that used to perform stops performing.
One detail that gets missed: a staging site left open to search engines can itself get indexed, which creates a duplicate of the whole site at a different address. The convention is to block it while it is in progress and remove the block at launch, and forgetting the second half of that is a common way for a new site to stay invisible for weeks.
What to Measure in the First Month
|
Week |
What to look at |
What a problem looks like |
|
Launch day |
Every old address, crawled |
Any that error instead of redirecting |
|
Week one |
Which new addresses are indexed |
Important pages still missing |
|
Week two |
Search impressions by page |
A page that had impressions and now has none |
|
Week four |
Inquiries by source |
Search inquiries down while other sources hold |
The last row separates a search problem from a general downturn. If direct and referral inquiries are steady and only search has fallen, the rebuild is the cause and the address list is where to look.
Where the Time Actually Goes
Owners are often surprised that the mapping work takes longer than expected and the design work takes less. On a site with a few hundred addresses accumulated over years, listing them, deciding a destination for each and testing the result is genuinely a day or two of careful work.
It is also the part that can be done before anything is designed, which makes it useful scheduling. The list does not depend on what the new site looks like. It depends only on what the old one contained, which is knowable immediately.
What Is Reasonable to Expect
A well-mapped rebuild usually holds its rankings and then improves on them, because the new site is faster and better structured. A poorly mapped one loses ground for a quarter and recovers over two.
Neither outcome is visible in the design. Both are decided before anyone opens a design tool, which is why the address list is worth asking about in the first conversation rather than the last.
The question that surfaces it is simple enough to ask any supplier: what happens to the addresses of the pages that currently rank. An answer describing a mapped list and a testing step is a supplier who has done this before. An answer about the new structure being cleaner is a supplier about to cost a business a quarter of inquiries it will struggle to attribute to anything.


