Nineteen country sites, one group domain, three acquired brands still on their original addresses, and a Search Console account per market because that is how each one was set up. Nobody in the group can answer which of those properties produced an enquiry in Germany last quarter.
Estates like this are assembled rather than designed. A market opens, an agency sets up a property, an acquisition arrives with its own domain, a campaign gets a microsite that nobody retires. Each decision was reasonable at the time. The result is a portfolio where the most important question has never been asked.
That question is not how each property performs. It is whether any two of them are working against each other — and no arrangement of separate reports can answer it.
The cost of nineteen separate views
While each market lives in its own account, a specific overhead accrues that appears on no invoice: assembling. Several logins, several exports, then a worksheet in which one person reconciles column headings manually, ending in a total whose derivation cannot be checked by anyone who did not build it.
At group scale this stops being an irritation and becomes a structural problem. A monthly reconciliation across nineteen markets is not half a day; it is a role. And because the output is manual, it cannot be audited, which means the one number the board sees is the number nobody can trace back to a source.
The second effect is more damaging and entirely invisible in separate reports. Where two of your own properties compete for the same terms, the attention divides between them. Each individual market report looks acceptable. Only side by side does it emerge that one property is suppressing another — and in an international estate the two are frequently in different countries, reporting to different managers, neither of whom has any reason to look.
What a shared workspace changes structurally
The arrangement in the Semalt workspace is not one tool per site but one workspace holding several domains, several Google accounts and several people, with filters that reach across all of them.
Three mechanisms behind the consolidated view
Individually unremarkable. Together they solve the scattered-access problem that group estates always have.
- Google accounts joined into one group. Several logins are linked, so Search Console properties stop sitting in separate compartments and appear in a single listing — which is usually the first time anybody sees all nineteen at once.
- Individual sites issued to outside addresses. One domain can be handed to a different email address — a local agency, the manager of an acquired brand — without exposing the remainder of the estate.
- Site tags applying across every view. One label — a region, an acquisition wave, a retired brand family — applies to every screen at the same moment, separating the reporting without fragmenting the underlying data.
The third mechanism does most of the daily work. Tags let the same underlying data be cut several ways without maintaining several reports: the board view filters to everything, the regional view to one tag, the acquired-brands view to another. Same figures, different frames, no reconciliation.
It is worth being explicit about the limits. No domains are combined, no campaigns are merged, and no software decides which property should carry which role — that remains a commercial judgement made by people. What changes is that the judgement can finally be made against figures that are actually comparable. A group hoping the workspace will resolve its portfolio question will be let down; a group hoping it will finally make the question answerable gets precisely that.
Five patterns that surface at the first joint review
The country site and the group domain
Both hold middling positions for the same German process term. Consolidated, one would stand considerably higher.
- Visible immediately in the ranking view
- Decision: one leads, one supports
The brand still on its old domain
Untouched since the deal closed, consuming attention and holding inbound references nobody has counted.
- Compare clicks and references first
- Redirect beats leaving it running
One market carrying the estate
A single property produces most of the result while eighteen others return residual figures nobody has quantified.
- Spend tends to track staff numbers rather than outcome
- The tag view puts the true ratio on one screen
No two properties built alike
Trailing slashes here and not there, a dedicated sitemap in one market and a shared group file in the next.
- Align before comparing anything
- Otherwise a year is spent on artefacts
A fifth pattern appears only in international estates and is the most expensive to find late: two properties publishing the same content in two languages, both aimed at buyers who search in English. In Benelux and the Nordics a large share of commercial searching happens in English regardless of the local language, so both properties contest the same queries, each holds a middling position, and each market review reports adequate performance. Only the shared-keyword column shows they are the same page twice. The remedy is rarely deletion; it is deciding which property owns the English terms and giving the other a genuinely local reason to exist. Which property owns which terms is settled during keyword research, and the editorial separation in our content strategy.
The consolidated ranking view
Across every domain being followed, the position-tracking side keeps one ordered table: a single score per property alongside its mean position, how many terms it holds and a movement line covering the last twenty-eight days. On paper that is one more set of numbers; in practice it is the screen where priorities get decided.
| View | Question answered | Typical consequence |
|---|---|---|
| Consolidated score per property | Which site stands best overall? | Budget follows the leverage rather than the headcount |
| Keyword count per property | How broadly is each positioned? | Breadth weighed against depth per market |
| Strongest page per property | What is the result actually resting on? | Protect and extend that page before anything else |
| Who else holds these terms | Which domains contest our field? | Your own properties are listed among them |
| Twenty-eight-day trend | Where is the estate heading now? | Short-term response instead of annual review |
The fourth row carries the uncomfortable finding. Every domain contesting the same terms is listed, including yours. In an international estate this is frequently the first evidence that a well-meant local initiative in one market has begun taking positions from another — something no market-level report will ever surface, because each one is written from inside a single property.
Access in an estate with agencies in it
Group estates almost always include external parties: a local agency in one market, a freelancer in another, a platform team in a third. The usual arrangement is a shared account whose credentials circulate, and it holds until somebody's contract ends.
- Local agency. Receives exactly the property it works on, under its own address. When the contract ends, that single grant is withdrawn — no password change, no effect on eighteen other markets.
- Country manager. Sees the reporting for their tag. Understanding the German figures does not require access to the Nordic ones, and giving it anyway generates questions nobody wants.
- Group marketing. Needs the consolidated view and nothing else. A PDF of up to 250 rows, in the group's own logo and colours, covers it entirely.
- Finance or board. Wants periods that can be set against one another; being current matters far less. A scheduled export at each quarter end serves better than an account nobody signs into.
What is gained here is less a security improvement than a reduction in noise. When each role sees precisely its own section, the questions that arise from somebody encountering a figure meant for another audience simply stop occurring — and in a group, those questions consume more time than the reporting itself.
Reports that reach the people they are for
Report layouts are configurable and can carry the group's own mark and palette. Rather more consequential than appearance are the row limits attached to each of the two output paths, since those dictate which material can go where in the first place.
The division follows on its own. Anything to be read goes in the PDF and stays under 250 rows. Anything to be processed goes in the export, where ten thousand rows are available. Both paths offer the same building blocks — movement over time, headline figures, tables that sort, compact trend lines, breakdowns by country and by device. What differs is solely whether the result gets read or gets processed.
One account shared by four people
Credentials passed on at handover. No way to establish who changed a setting, or when.
- Every departure forces a change for everyone
- External parties hold the same keys as staff
One address per participant
Each person works under their own identity and sees only the properties assigned to them.
- Grants withdrawn individually
- Actions remain attributable afterwards
One detail that saves recurring trouble in international teams: authorising Google happens once and covers the mailbox, the search reporting and the analytics account in the same step. Three separate setup steps become one, and the most frequent cause of a market quietly losing one of its data feeds — an authorisation that lapsed months ago while the other two kept working — disappears with it.
The record that outlives the people
Each project carries a dated sequence: assistant answers, generated reports, new placements with the referring domain's rating and traffic, items open and closed, and campaign notices. Four filters cover everything, links, files and tasks, and each task holds one of three states — active, deferred or cancelled.
With one site you remember; with nineteen nobody does
In a group, staff turnover guarantees that undocumented reasoning disappears within two years.
- Replies built on this project's own figures. A routing stage picks, per question, which blocks are relevant — Google reporting, position data, campaign state or figures you supplied — loading between none and three of them.
- Output as it is written, twenty messages retained. Enough to follow a line of enquiry across several questions; short enough not to become an archive.
- Placements listed with the numbers that matter. Each one arrives with the rating of the site carrying it and that site's traffic, rather than as a bare address.
- Searching the entire record as text. The feature nobody values in month one and everybody uses in month eighteen, when somebody needs to know what was settled about an acquired domain two springs ago and why.
Of the three states, "deferred" is the one that matters most across an estate — and it is the reason the timeline in My SEO stays usable rather than becoming a backlog. Much of the work across an estate is not wrong, merely not now — restructuring an acquired brand while the main market runs a product launch, for instance. Deferred items stay findable without cluttering the active list. The two alternatives both fail: keep everything open and within a year the list goes unread; cancel instead and the justification disappears with the entry. Terms and addresses can also be handed to the assistant in batches, which spares somebody the intermediate spreadsheet.
For an international estate the record has a second function that outweighs the first. Group reorganisations move properties between owners every few years, and each move erases whatever was only ever held in somebody's head. A searchable timeline converts three years of scattered local effort into something a successor in another country can pick up — which is, in most groups, the difference between a market that compounds its work and one that restarts it every second year. The assistant itself sits in My SEO Stream and pulls only the data blocks a given question needs.
Bringing an assembled estate under control
| Step | Action | Result |
|---|---|---|
| 1 | Write down every domain the group owns, unclaimed ones included | a verified list instead of an assumed one |
| 2 | Join the separate Google logins into a single group | all properties appearing on one screen |
| 3 | Assign tags along commercial lines, not technical ones | a filter that reaches every view |
| 4 | Work through the ranking table and the shared-term column | the places where two of your sites collide |
| 5 | Fix one standing report per audience | no more monthly circular to the whole group |
| 6 | Record decisions in the project timeline | reasoning that survives a reorganisation |
Take time over step three. A tagging scheme that has to be redone later consumes more effort than it ever saved. Keep the number low and let the categories mirror how the business is actually run rather than how the sites were built — a region label earns its place, a platform name does not. And step one is less trivial than it looks — in most groups of this size, at least two live domains turn up that no current employee can account for. The technical alignment those domains then need is covered under technical SEO.
All of this — the listing of properties, the consolidated ranking, the tagging and the report definitions — can be applied to the group's actual domains inside the Semalt dashboard instead of being rebuilt in a spreadsheet each month. Synchronisation runs continuously in the background, so nobody has to trigger a refresh before a review.
Open the estate overview in the dashboard
Questions from group and country teams
Is this worth doing with only two properties?
Yes, though the argument differs from the one at nineteen properties. Administrative overhead is negligible with two. What matters is whether they collide, and that only shows in a side-by-side view. If anything the stakes are higher at two, because a group domain competing with its own country site for the same terms is the most frequent arrangement in this sector — and the one least likely to be noticed from inside either report.
What should happen to a domain that came with an acquisition?
Measure before deciding. Check what traffic it still receives and whether external sites still reference it. If either figure is meaningful, a carefully mapped redirect is better than switching it off. If both are near zero, it costs a renewal fee and some attention. Either way the decision belongs in writing — twelve months later nobody remembers why a redirect exists, and it gets removed by accident.
Can a local agency see only its own market?
Yes, and specifically: a single property is granted to that agency's own address, and the rest of the estate stays invisible to it. This removes the circulating shared password and has the useful property that access ends precisely when a contract does, without anybody having to change credentials that eighteen other markets depend on.
How many tags should an estate have?
As few as will do the job. For most groups three to five will cover it, drawn along whatever line the business is actually managed by: geography, division, or the distinction between core properties, acquisitions and things nobody has retired yet. Push past a dozen and the scheme reproduces the very muddle it was meant to clear up. That usually happens when each property is tagged by whoever happened to set it up, rather than once, centrally, by someone looking at the whole list.
Does every property need its own campaign?
Campaigns are purchased per domain while the reporting stays shared. In practice that means one market can be worked actively while others are only observed, without losing the consolidated view. In estates where a single property carries most of the result, that split is usually the more economical arrangement — and it is invisible until the estate is viewed as a whole.
What is the most common mistake when consolidating?
Retiring a property before anyone has looked at what points to it. Traffic can be nil while inbound references are not, and those disappear along with the domain. Four steps in this order avoid it: inventory, measure, decide, redirect — with retirement last and often unnecessary. Running a close second is failing to write the reasoning down, because by the next reorganisation the redirect looks like an accident and somebody removes it believing they are tidying up.