Four websites, one company, and nobody still employed who was in the room when the second and third were registered. The task is not to pick a favourite but to establish, from evidence rather than seniority, which of the four has an audience of its own.
Estates like this begin with one sensible-sounding sentence. A competitor with a single office outranks you for a phrase containing the word Göteborg, and somebody concludes that the way to rank in Göteborg is to own a website about Göteborg. The domain is registered that afternoon. Eighteen months later the reasoning is applied to Malmö, and a year after that to the north, by which point it is no longer an argument but a habit.
What remains is four properties describing one business in nearly identical words, three of them under twenty pages, none in anybody's job description. Unwinding that is not a technical exercise but an argument about which site the sales manager is fond of, and it can only be settled with numbers collected before anyone knew they would be used to close a website. Hence putting the estate into one workspace first and deciding second.
Nobody ever decided to run four websites
Take a firm installing and servicing ventilation in commercial buildings, based in the middle of the country. Its original domain dates from 2011 and carries four hundred pages. The Göteborg domain, registered in 2019, has fourteen. The Malmö one has eleven, nine of them the Göteborg pages with the place names swapped. The northern one has a homepage and a news item from 2022 announcing that it exists.
None of the three is lazy or dishonest. Each was built by somebody with a real observation — a regional competitor outranking a national supplier — who reached for the only lever they knew. Its cost is not the registration fee: four properties accumulate links, mentions and trust separately, so six years of effort was divided four ways instead of compounded once.
An estate of this shape is also invisible from inside any single Search Console account, since each site looks modest alone. The pattern — four properties chasing the same phrases, three thin, one quietly carrying the business — appears only when they are read side by side, which is what nobody has the tooling to do.
What a shared workspace is, before it becomes a decision
A workspace here is not a folder of shortcuts. It is an account holding several websites at once, keeping a separate working record for each, and letting colleagues move between them on one login.
One account, one project per website
For the company whose estate grew by accident and must now be read as one picture.
- A dashboard spanning every verified property. Clicks, impressions, click-through rate, trends and health flags for all four sites at once, where the pattern becomes visible.
- A working record per site. Each project keeps its own stream: questions and answers, reports as they generate, new placements with the referring page's rating and traffic, and open tasks.
- Site tags that filter the whole panel. A tag is not a folder. Selecting one narrows every view you open afterwards, so a grouping is defined once and holds.
- Colleagues invited freely. Charges follow the website rather than the number of people signed in, so nobody need share credentials.
- Data that refreshes itself. Background processes keep some thirty-five views current across eleven integrations, so nothing is reloaded by hand.
Tagging does the work a folder tree cannot. Tag the three thin properties as candidates and the original as the survivor, and every view afterwards reads both ways round: the estate with the candidates counted in, and without them. That comparison is most of the analysis.
Rank tracking adds the figure the finance director will ask for: one portfolio-wide ranking score with a twenty-eight-day trend curve, sitting above the individual sites. It answers the only real question about a consolidation — whether the company ended up better off.
Which of the four actually earned an audience
The instinct is to keep whichever site has the most traffic. That settles nothing, since the original is larger for reasons unconnected to whether the regional sites deserve to exist. A regional site earns independence by having something the parent would not inherit if it absorbed the pages tomorrow.
Five tests separate the cases, and none needs an opinion. Run each with one property selected at a time, over ninety days rather than twenty-eight, since regional volumes here are thin.
| Test | Where you read it | Points towards keeping | Points towards folding |
|---|---|---|---|
| Searches for its own name | Keywords, brand terms | People look for it unprompted | Only the parent name appears |
| Links earned independently | Competitor and portfolio views | Referring domains the parent lacks | Every link came from your own sites |
| Queries won here and nowhere else | Keywords, properties compared | A distinct set with its own intent | The same phrases, worse positions |
| Pages unique to this site | Pages view | Local references, staff, real cases | Templates with a place name swapped |
| Where enquiries originate | Your own enquiry records | Enquiries citing this site by name | Everything arrives on the main number |
Applied to the ventilation firm, the tests surprise those who commissioned them. The northern site fails all five and closes without discussion. The Malmö site fails four, defended only because a former colleague built it. The Göteborg site passes three: branded searches for its own name, two referring domains from a trade body that never linked to the parent, and harbour ventilation queries the parent has never ranked for, written by somebody who did that work.
- Judge distinctness, not size. Ask what is lost by moving the smaller site's content. For a template with a place name swapped, nothing is.
- Count referring domains, not links. Forty links from your own four sites are one relationship. Two from an organisation that chose to link are two, and only that kind survives intact.
- Write the verdict down with its evidence. Record it in that site's project stream with the figures behind it, because somebody who was not consulted will reopen the question in eight months.
The indexing budget belongs to the account, not the website
Here is the constraint that changes the arithmetic of a multi-site estate, and it is easy to read past. The URL tracker carries a daily allowance of 1,000 addresses per account. Not per website. A workspace holding four properties has the same ceiling as one holding a single site, and the four share it.
Batch size and daily allowance do different jobs. Ten thousand addresses is what you hand over at once; a thousand a day is how fast the queue drains. Hand over ten thousand and the workspace is committed for ten days — unfortunate if most belong to a site you are about to retire.
Sitemaps have their own shape — nested indexes followed three levels down, 1,000 sitemaps to a job, two jobs running and twenty waiting — and an estate mid-consolidation is when those limits get noticed, so stagger it. The per-address log of bot visits is where you watch a redirected address being picked up.
Folding three sites into one without losing the record
Done properly, a consolidation is a transfer, not a deletion. Content with value moves into a regional section of the surviving site, old addresses point permanently at their counterparts, and retired domains stay registered rather than lapsing.
Anything specific to the place
Real references, named staff, an address with opening hours, a local case — the material the parent never had.
- Becomes a proper regional page
- Linked from the national service page
Templates with a place name in them
A page differing from its sibling by two words adds nothing on arrival and dilutes what is there.
- Redirect to the service page
- Do not recreate it to be safe
The domain, still pointed
Renewal is trivial, and the redirects must keep working long after anyone remembers why.
- Renew it indefinitely
- One hop, never a chain
The verified property
Keep the retired site verified in the workspace and watch its addresses being crawled and released.
- Six months of observation
- Tagged as retired, not removed
For about two months the numbers then look worse, because three properties stop reporting before their contribution surfaces on the fourth, and anyone reading a single dashboard concludes the project failed. The portfolio score protects the decision: it counts the estate as one thing, so a transfer reads as a transfer. Keyword dynamics helps too, since it records crossings rather than totals — a phrase leaving the old site and appearing on the new one is the transfer working.
When the dealer in Umeå really is a different company
All of that assumes one owner and one balance sheet. Change the assumption and the same shape — one brand, several websites, overlapping words — becomes a structure that must not be consolidated.
Consider an importer of forestry and groundcare machinery with fourteen dealers across the country. The dealer in Umeå is a separate legal entity with his own owner, staff, workshop and customers, selling your machines under a shared brand. His website is his. Head office cannot fold it into anything, because he is not a region of the importer but a business with a franchise agreement.
A workspace holding sites that belong to other people
For importers and dealer networks, where the estate is legitimate and the question is access, not consolidation.
- Google accounts linked as a group. Gmail, Search Console and Analytics authorised in one consent flow, with standalone accounts tied into groups, so verification stays with each domain's owner.
- Sharing one site at a time. A single property goes to a named email address. The recipient sees that site and nothing else — not the network, not your tags, not who else has access.
- Access that can be taken back. What you shared stays listed in your overview and can be withdrawn without renegotiating anything or changing a shared password.
- Reporting that carries the right name. The report builder takes a logo and colours, so a document reaching a dealer looks as though the importer sent it, not a tool.
A structure problem
One balance sheet behind every site. The remedy is fewer properties and better regional sections on the survivor.
An access problem
Independent companies under a shared brand. Nothing merges, and the whole question is who may read whose figures.
Who is allowed to read whose enquiries
This is where dealer networks stall, and the sticking point is never technical. Head office wants to see how each dealer's site performs, fair enough when the brand's reputation is partly in their hands. The dealer in Umeå notes that head office is asking to read his enquiry flow while offering nothing back, and declines — also fair enough, since some of that flow concerns machines that are not yours. An arrangement running only upwards is a reporting obligation, not data sharing, and it is complied with badly. The version that holds is reciprocal, written down before anyone connects anything.
| Party | Sees | Does not see | What they get back |
|---|---|---|---|
| Importer, own sites | Everything on brand and product sites | — | A picture the network cannot assemble |
| Importer, dealer sites | Only what each dealer shared | Dealers who declined; other brands | An honest view, not a mandated one |
| Dealer, own site | Everything on it, without asking | Other dealers' properties | Analysis he would not have bought |
| Dealer, brand sites | Product and support pages shared with the network | Commercial and recruitment areas | Evidence of what the brand sends him |
| Agency or consultant | The one property they were engaged for | The rest of the estate | Access ending with the engagement |
Notice what the fourth row buys. A dealer who can watch demand arriving on the brand's own pages in his territory stops experiencing the arrangement as surveillance, which is what makes the third row acceptable to him.
- Name the grant in the contract, not in an email. Dealer agreements are renegotiated on a cycle, and an arrangement living in a mail thread is re-argued whenever a territory changes hands.
- Review the access list each quarter. The consultant hired for a migration, the agency you replaced, the dealer who gave up his franchise last spring — all keep access until somebody removes it.
- Be open about the campaign level. Whether a site runs the automated tier at 149 USD a month or the reviewed tier at 500 explains most of the differences a network dashboard will show.
Questions that come up during a consolidation
Can we put all four sites in one workspace and decide later?
That is the right order. Verification, tagging and the shared dashboard cost nothing, and a quarter of comparable data is what the keep-or-fold argument needs. Decide first and measure afterwards, and the decision belongs to whoever is most senior in the room.
How long should we watch a regional site before closing it?
Ninety days at minimum, preferably a window covering that site's busy season. Volumes are low enough here that a month is mostly noise, and a seasonal business shut down in its quiet quarter looks dead when it is out of season.
Does the daily indexing allowance really cover the whole estate?
Yes. A thousand a day is an account ceiling, so four properties share it rather than each receiving their own. Submit for the surviving site and anything that passed the tests, and leave the retiring ones out.
Our dealers use different agencies. Does that break the shared view?
No. Each agency works inside the property it was given, and the dealer decides separately whether to share it upwards. Access is granted per site to a named address, so several parties can hold different slices of one network without seeing each other.
What happens to the English pages when a regional site is folded in?
They usually improve, because the foreign reader was never served by the split. A company establishing operations here reads a national page as reassurance and a county page as a limitation, so one national site with real regional sections answers both audiences.
Decide what the estate is before you decide what it costs
An estate assembled by accident will not be unwound by argument. Each of those four domains has somebody who remembers why it was registered, and none has a defender who has looked at what it does now. A workspace decides nothing for you; it produces what the discussion has lacked, which is four properties described in the same terms over the same window. A panel holding campaigns, Google data and indexing together is useful here mainly as a way of ending an argument with evidence.
The two estates here look alike from outside and are opposites underneath. One is a company that divided itself four ways and should be three-quarters smaller. The other is fourteen separate businesses under one brand, where nothing merges and the question is who may read whose numbers. Ownership tells them apart, and the wrong remedy is how a franchisor loses a dealer's cooperation for a decade.
Three things stay outside any workspace: which regional audiences you intend to serve, a judgement about vans and delivery days; how you explain a consolidation to whoever built the retired sites; and what a dealer is owed for his enquiry data. What the panel removes is the fortnight of assembling comparable numbers by hand, which is where these projects usually die. How the generative layer describes multi-site companies is covered in the market research section, and the structural work that follows sits under our services.
Start with the estate you have rather than the one you mean to end up with. Verify all four properties, tag three as candidates and one as the survivor, and give it a quarter before anybody proposes a redirect: open a workspace and connect the sites you already own. If a regional site turns out to have an audience that would not follow its content home, you have found the exception. If none does, you have avoided an argument you were going to lose on volume rather than evidence.
