The network site, three regional sites inherited when offices merged, a customs brokerage on its own domain because it is a separate legal entity, and a warehousing brand nobody has updated since 2019. Five properties. Two of them compete for the same terms, and no report anybody currently receives would ever show that.
Logistics groups collect domains the way they collect legal entities — because operations, licensing and acquisition history all push in that direction. Each addition made sense at the time. What emerges after fifteen years is a set of properties that has never been looked at as a set.
What goes unasked is not how each performs. It is whether any two are taking positions from one another, and that is structurally invisible to reporting organised one property at a time.
What five separate views cost every month
While each property lives in its own account, an overhead accrues that appears on no invoice: putting the picture together. Signing in five times, exporting five times, then a spreadsheet where one person manually matches up column names — ending in a figure that only its author could explain.
Where marketing occupies part of one person's week, half a day a month on reconciliation is a meaningful share of the capacity — and it produces no outcome, only figures in a comparable shape. The resulting total is also hand-built, which means it cannot be audited and therefore should not be put in front of a principal who may ask how it was derived.
The second effect does more damage and is entirely invisible in separate reports. When two properties you own go after identical terms, whatever attention exists gets split between them. Read on its own, each report looks unremarkable. Set them next to each other and it becomes obvious that one is suppressing the other — and in a group of this shape the pair is frequently the network site and a regional site inherited from a merger, both describing the same corridor.
Seeing them together without merging anything
What the Semalt workspace does here is keep several domains, several Google logins and several people in one place with filtering that spans the lot — while leaving each property exactly as separate as the licensing arrangements require it to be.
Three mechanisms, three distinct purposes
None of them consolidates a single thing. All three make a scattered group legible.
- Tying the Google logins together. Connect the accounts and the search properties come out of their separate compartments. In most groups that is the moment somebody first sees all five listed on a single screen.
- Granting one property to an outside login. The agency handling the warehousing brand receives that property and nothing further. Beyond the obvious, it means access ends precisely when a contract does rather than when somebody remembers to change a password.
- Labels that hold across every screen. Tag something as network, regional, brokerage or dormant and the tag applies across every screen simultaneously. The reports split accordingly; the data underneath is never fragmented.
In practice it is the labels that do most of the daily work, and the campaign side inherits them, so placement activity can be reviewed per label as well. One dataset can be presented several ways without anybody maintaining several reports: everything for the principal's summary, one label for the regional review, another for whoever still owns the dormant brand. Same figures, three framings, no manual joining in between.
Being explicit about the limits is worth doing. No domains get combined, no campaigns get merged, and no software decides which property should carry which corridor — that stays a commercial judgement, and in a group with separate licensed entities it is frequently a legal one as well. What changes is that the judgement can finally be made against figures that are genuinely comparable.
Five patterns from the first joint review
Network site against regional site
Both describe the same corridor. Search engines pick one, and it is frequently the regional site because it is older.
- Visible at once in the ranking table
- Decide which one leads, which supports
The brand from an office that no longer exists
Nobody has edited it since the merger, yet links still point at it and it still occupies positions for terms that now belong to another property.
- Count the references before deciding
- Redirecting beats leaving it live
One property carrying the group
A single site produces most of the result while four others return residual figures nobody has quantified.
- Spend tracks the entity chart, not the outcome
- The label view puts the ratio on one screen
Five properties, five builds
Five suppliers, five content systems, five ways of forming a URL, accumulated over a decade and a half.
- Level them up first, compare afterwards
- Skip that and three months go on measurement noise
A fifth pattern belongs specifically to groups with a licensed entity in them and causes the most confusion when it appears: the brokerage site outranking the network site for the group's own name. That is arguably correct behaviour — the brokerage page is often more specific and more linked — and it means a shipper searching for the group lands on the entity least equipped to sell the wider service. Nothing about this is fixed by configuration. It is fixed by deciding what the brokerage page should say about the rest of the group, which is an editorial decision nobody has been asked to make.
The consolidated ranking table
Everything being tracked appears in a single ordered list. Each row gives a property, its overall score, the mean position it occupies, how many terms it holds, and a short line covering the previous month. Written out like that it reads as yet more reporting. In use it is the page people open when they cannot agree on what to do next.
| View | Question it settles | Consequence |
|---|---|---|
| The overall score, property by property | Of the five, which is in the strongest position? | Work goes where results are, not where the org chart points |
| How many terms each one holds | Is this property spread wide or focused? | A judgement about coverage versus concentration |
| The single strongest page on each | What is holding up the numbers? | Protect that page before ordering anything new |
| Everyone else appearing on these terms | Which domains occupy our vocabulary? | Several of them will be your own |
| The last four weeks of movement | Which direction is the group travelling? | Correction inside a month rather than at year end |
Row four is the one that produces awkward conversations. It lists every domain competing on those terms without exception — which includes the ones you pay for. In a logistics group this is usually the first hard evidence that a regional site has quietly become the primary property for a corridor the network site was meant to own — which might be the right arrangement or entirely the wrong one — but either way it is now something to decide on purpose instead of something that keeps happening by default.
Who holds which key
Groups of this shape almost always involve outside parties: an agency on one regional site, a contractor on the brokerage, a former colleague still maintaining the dormant brand as a favour. The default is a shared account whose credentials circulate, and it holds until somebody's arrangement ends.
- An outside agency. Works on exactly one property, signed in as itself. When the arrangement ends, that single permission goes and nothing else is touched — no credential shared with four other properties needs changing.
- Entity or regional management. Sees the reporting for its label. Understanding the brokerage figures does not require access to the network site, and in some groups it is not permitted either.
- Principal or group management. Needs the consolidated view and nothing beneath it. A printable document of up to 250 rows, in the group's own mark, covers that entirely.
- Whoever handles the next audit or acquisition. Needs comparable periods and figures whose derivation can be traced. A scheduled quarterly export serves far better than a live account nobody signs into.
The last item deserves planning before it becomes urgent. In a sale, a merger or a compliance review, the question arrives sooner or later: where did this traffic come from and how were these links obtained. Five exports stitched together in the week the question lands look exactly like five exports stitched together. One drawn from a single workspace, backed by an entry-by-entry history, reads differently — and the people best placed to spot the distinction are precisely the ones examining you.
A single login circulating among six parties
Whoever takes over the role inherits the password. Tracing a settings change back to a person is not possible.
- One person leaving means resetting for everybody
- Outside contractors carry identical access to employees
Separate identities throughout
Everybody signs in as themselves, and what is visible to them is exactly what they were given.
- Access removed individually and immediately
- Every change traces back to somebody
One detail that removes a recurring irritation in groups spread across time zones: Google needs authorising only once, and that single step covers the mailbox, the search reporting and the analytics account together. Three separate setup exercises become one, and with them disappears the classic failure of this kind, where one permission lapsed some months back, the other two kept functioning, and nobody noticed until a report came up empty.
Reports that reach their audience
You can arrange a report however you like and put the group's mark and colours on it. Far more decisive is a mundane constraint: each of the two output formats accepts only so many rows, and that ceiling determines what is capable of travelling by which route in the first place.
Everything follows from those two numbers. Material intended for reading goes into the printed version and has to fit inside 250 rows. Material intended for further calculation goes into the export, with ten thousand rows available. Neither route is short of building blocks — trends across time, headline figures, sortable tables, compact movement lines, breakdowns by country and by device are all present in both. The only real difference is who receives the output and what they intend to do with it.
What survives the next restructuring
One property you remember; five you do not
Offices combine, entities change hands, and anything not written down evaporates within a couple of years.
- Answers drawn from this project alone. Ahead of replying, a routing stage identifies which material is actually relevant — the search reporting, the ranking data, the campaign's state, or figures you supplied — and pulls in anything from none of it to three of them.
- Text appears as it is produced, with twenty exchanges kept. That is enough history to follow one thread over several questions and deliberately too little to accumulate into an archive.
- Each placement arrives with its metrics attached. Every entry carries the strength rating of the site hosting it and that site's visitor volume, so nothing has to be looked up separately.
- Full-text retrieval across the entire log. Of no interest whatsoever in the first months, and the thing that saves a day of work a year and a half later, when a successor needs to know what was settled about the dormant brand and on what grounds.
Inside My SEO Stream each project keeps a dated log holding the assistant's replies, generated reports, new placements, items open and closed, and campaign notices. Every item sits in one of three states — running, postponed or dropped — and "postponed" is what keeps the log readable. A great deal of work across a group is not wrong, only not now: restructuring the dormant brand while the network site handles a peak season, for instance. Leave everything open and within a year nobody opens the list; drop things instead and the reasoning leaves with them.
Bringing an accumulated group under control
| Step | Action | Result |
|---|---|---|
| 1 | Write down every domain the group owns, renewals nobody recognises included | a verified list instead of an assumed one |
| 2 | Join the separate Google logins into a single group | every property on one screen |
| 3 | Apply labels that mirror how the group is run, not how the sites were built | one filter working everywhere |
| 4 | Go through the ranked list together with the shared-term column | a map of where your own properties get in each other's way |
| 5 | Fix one standing report per audience | no more circular to the whole group |
| 6 | Record the decisions in the project log | reasoning that outlasts a restructuring |
Step one is less trivial than it looks: in most groups of this age at least one live domain turns up that nobody can account for, usually a defensive registration somebody later built a page on. Step four is where the actual work sits, because deciding that the network site owns a corridor and the regional site supports it requires agreement between two managers who have never had to agree about it. The question of which property lays claim to which term gets answered in keyword research; how the writing is divided between them is covered by our content strategy; and the levelling-up that must happen before any comparison means anything falls under technical SEO.
Everything described here — the inventory, the ranked comparison, the labels and the standing reports — runs against the group's real domains from inside the Semalt dashboard, rather than being reassembled in a spreadsheet at the start of every month. Data refresh happens on its own, continuously.
Open the group listing in the dashboard
Questions from group and entity management
Our entities are legally separate. Does joint reporting create a problem?
Reporting jointly and managing jointly are different things, and conflating them is what stops most groups from doing either. Looking at figures side by side to detect self-competition touches no responsibility and changes no property. Deciding what a page says remains with whichever entity owns it. Writing those two sentences into a procedure note usually settles the question permanently.
What should happen to the brand from a merged office?
Measure before deciding. Establish what traffic it still receives and whether external sites still link to it. Where either figure is meaningful, mapping each old page to its replacement retains more value than a shutdown ever will. Where both are negligible, all the domain consumes is a renewal invoice and occasional thought. Whichever way it goes, put the reasoning in writing: twelve months on, the redirect looks unexplained and gets deleted by whoever is tidying that quarter.
Can an agency see only its own property?
Yes. The relevant property is attached to a login the agency already owns; nothing else the group holds is visible from there. What is easy to miss is the second benefit: no shared password remains in circulation, and closing an arrangement means withdrawing a single assignment instead of changing a credential that four other properties still rely on.
How many labels does a group this size need?
Somewhere between three and five, following whichever division the group genuinely operates along. Network, regional, brokerage and dormant will cover most cases in this sector. Push past a dozen and the scheme reproduces the confusion it was meant to remove, which tends to happen when each property is labelled by whoever set it up rather than once, centrally, by somebody reviewing the whole inventory.
Does every property need its own campaign?
Campaigns are bought per domain; the reporting covers all of them regardless. So one site can be worked on while the others simply run and are observed, and nothing drops out of the overall view. Since it is normal rather than exceptional for one property to account for most of the result, that uneven allocation is usually also the cheaper one — a fact that stays hidden until somebody looks at the whole set at once.
What is the most common mistake when consolidating?
Switching something off without first checking what points at it. Visitor numbers can sit at zero while inbound links remain, and those go the moment the domain does. Follow an order and the problem disappears: list what exists, measure it, take a decision, set up the redirect — with retirement last, and usually unnecessary once the redirect is in place. Second on the list is doing the technical part while quietly skipping the conversation about which property owns what, which leaves you with a tidy inventory of the identical collision.