The German site was built by head office in 2019, translated from the corporate template, and has not been touched since. It reports a few hundred impressions a month. Somebody in another country is now deciding, on the basis of that number, whether the Stuttgart subsidiary deserves a marketing budget.
This is an uncomfortably common situation among engineering firms with a German arm. The subsidiary sells into the automotive and machinery supply chain around Stuttgart. The website reflects a corporate identity written elsewhere, for a market that behaves differently, measured against territories that behave differently again.
None of the numbers are wrong. The comparison being made with them is.
What an inherited site cannot do
A corporate template performs the job it was designed for: presenting a group consistently across markets. What it cannot do is answer a question posed by a German technical buyer, because that question was never part of its brief.
The concrete symptom is always the same. The German pages carry the group's vocabulary — capability statements, values, a heritage paragraph — and none of the four specifics a purchasing engineer looks for: process, tolerance class, batch range, material. Those four exist inside the subsidiary. They have simply never been written down in German on a public page.
Who searches, and who signs
In the supply chain around Stuttgart these are rarely the same person, and the distinction changes what the numbers mean. Search happens early, at the point where a shortlist forms. Signing happens months later, often after a plant visit and an audit.
Four people, four different queries, one subject
They appear side by side in the keyword view with nothing to distinguish them.
- The design engineer. Searches during specification, months before any tender, usually by material or process. Highest volume, longest distance from a purchase order, and consistently over-weighted when priorities are set by volume.
- The technical buyer. Searches by standard, certification and capability. Reads a page for thirty seconds and leaves if the technical figures are absent. Moderate volume, high commercial relevance.
- The quality engineer. Searches by standard designation before an audit. Very low volume; almost always ignored; the single query type where a dedicated page still wins easily.
- The project manager with a supply failure. Searches urgently for a replacement supplier, often from a phone on the shop floor. Lowest volume, by far the highest conversion, and the reason mobile performance still matters in a desktop-dominated segment.
Because these four appear undifferentiated in the reporting, priorities set by search volume will consistently favour the design engineer and neglect the project manager. That is the wrong way round commercially, and it is invisible until somebody separates the roles by hand.
Why the German property looks worse than it is
Three structural reasons make an industrial German property read as weak when it is merely different. Each of them is worth stating explicitly to a head office before any budget conversation.
| What head office sees | What is actually happening | What to report instead |
|---|---|---|
| Low absolute traffic | The relevant terms have twenty to thirty monthly searches nationally | Enquiries with recorded origin, year on year |
| Volatile week-on-week figures | At thirty clicks a month, one visitor moves the percentage | Quarterly or annual comparison only |
| Poor conversion rate | Half of all enquiries arrive by telephone and are never attributed | A recorded origin field in the enquiry log |
| Weak performance versus other markets | Those markets sell different products through different channels | Contribution per enquiry, not sessions |
The third row is the one that changes minds. In industrial supply, a substantial share of enquiries arrive as a phone call after somebody read a page. No analytics package will attribute that call. The only mechanism that works is a question asked during the first conversation and written into the enquiry record — unglamorous, easily forgotten, and the only bridge between what the reporting shows and what the business decides.
Matching the view to the question
The Search Console reporting inside the Semalt workspace is split into eight views rather than one configurable report. For a subsidiary reporting upward, that split is convenient, because each view corresponds to a question somebody actually asks.
- Overview across every verified property. Clicks, impressions, click-through rate, trends and health flags for all sites at once — the only view a group head of marketing normally needs.
- Keywords with position history. Position over time, click trend, and breakdowns by device and country for each individual term. The view that answers what actually changed.
- Pages. The same metrics cut by landing page. In engineering this almost always reveals that a reference project or a data sheet carries the result, not the capability overview.
- Countries. For a subsidiary, the view that separates genuine German demand from group-wide noise arriving on a German URL.
- Keyword dynamics. Entries and exits from the top 3, top 10 and top 30. The one view with a direction rather than a state, and the only one worth a weekly habit.
Frequency matters as much as content here. The overview is monthly. Keywords and pages are fortnightly. Countries and devices are quarterly. Keyword dynamics is weekly and takes fifteen minutes. Subsidiaries that try to review everything weekly stop reviewing anything by the third month.
One further note on frequency. Where a subsidiary shares a property with the group — a single domain with a language folder — the country breakdown stops being optional and becomes the first filter applied to every view. Without it, traffic from six markets is aggregated into one figure, and the German line is invisible inside it. All eight views sit in the same Semalt workspace as the campaign itself, so applying the filter once carries through rather than being reapplied per export.
The competitor you have never met
Search Console describes your own property and nothing outside it. It cannot tell you which suppliers appear alongside you, how strong their domains are, or whether the field is tightening. The position-tracking side covers that across six views, and in this segment it usually produces one specific finding.
The competitor from the tender list
A comparable firm in the same region, known to sales, discussed at every review meeting.
- Rarely shares many search terms
- Competes at the signature, not at the shortlist
The manufacturing portal
A national marketplace or directory holding the process terms, often with a decade of domain history.
- High shared-keyword count
- Frequently a channel rather than a rival
The second card deserves a strategic decision rather than a reflex. Competing head-on against an established portal for a generic process term is a multi-year project with an uncertain outcome. Being listed on that portal produces enquiries within weeks. For a subsidiary with a modest budget and a head office asking for results, the second route is usually the defensible one — and it does not preclude building specific pages for the narrower terms where the authority gap is small.
The corporate capability page
Translated from the group template. Describes values, heritage and a philosophy of quality.
- No figure a buyer can compare
- Identical to nineteen sister sites
The process page from the plant
Names the machines, the tolerance class held in series, the batch range and the materials run regularly.
- Comparable against a rival
- Answers the audit question directly
The distance between those two cards is the whole of the opportunity available to most German subsidiaries, and it requires no budget approval from abroad. The information on the second card already exists inside the plant; it sits in quotations, in audit documentation and in the heads of two or three people. Publishing it in German is an afternoon of work and does not touch the corporate identity at all, because it adds pages rather than altering existing ones — which is also why it tends to get approved.
A review rhythm that fits small numbers
Because absolute figures are small, this market needs a slower cadence than a consumer business. A quarterly cycle with one short weekly habit has proven workable.
| Cadence | View | Decision it supports |
|---|---|---|
| Weekly, 15 minutes | Keyword dynamics, entries and exits | Check losses technically, link winners internally |
| Monthly | Pages and countries | Assess how much demand is genuinely German |
| Quarterly | Competitors and global ranking | Decide which capability pages to write |
| Annually | Year-on-year enquiries with recorded origin | The only defensible performance statement |
Where the weekly habit is genuinely impossible to sustain, the fallback is to let the campaign automation carry the term selection and to keep only the monthly review. That is a worse outcome than the weekly slot and a considerably better one than an annual panic before the budget round.
The weekly quarter-hour has exactly two jobs: to catch a lost position while the cause can still be reconstructed, and to notice when a term has risen without any effort — the moment when an internal link from a strong page produces the most benefit. Clicks, impressions and average position deliberately stay out of that weekly slot, because at this volume they fluctuate randomly and generate discussion without substance.
One habit is worth adopting alongside the weekly slot, and it costs nothing. Whenever a term rises into the leading positions, note the date in the same file where enquiries are logged. Six months later, when somebody asks whether the activity produced anything, that file contains two columns that can be laid against each other — the dates on which visibility changed and the dates on which enquiries arrived. In a business with a handful of orders a year, temporal proximity is the only evidence available, and it exists only if somebody wrote both columns down while they were happening.
This matters more in a subsidiary than in an owner-managed firm, because the audience for the evidence is elsewhere and arrives with a spreadsheet rather than with context. Two dated columns and a contribution figure will carry a budget conversation. A chart of impressions will not, however good it looks. The campaign side of My SEO keeps the visibility half of that record automatically; the enquiry half remains a manual discipline and always will.
Making the German number legible abroad
Reporting that leaves the country serves a different purpose from reporting used at the desk where the work happens. It will be read by someone who has never looked at a German results page and who is lining this territory up against five others on the same slide, which puts a hard limit on how much explanation it can carry.
Five items are usually enough: the enquiry count with recorded origin against the same period last year, whichever three pages carried that result, which terms entered and left the leading positions, how much of the demand was genuinely German, and a single explanatory sentence. Full keyword lists, competitor tables and country breakdowns belong in the export for the working level. Reports carry your own logo and colours, which decides more often than it should whether a document gets forwarded to a board or filed unopened.
The problem that appears in year two
Subsidiaries change marketing staff more often than head offices assume, and each change resets the institutional memory. The person who chose the terms leaves, the person who inherits the account finds a set of decisions with no reasoning attached, and the work begins again.
A written record instead of institutional memory
Every question asked, every report generated, every link created and every open item, in one dated sequence per project.
- Replies built on this project's figures. Before answering, a routing stage picks which blocks of data are actually relevant to the question — Google reporting, position data, campaign state, or figures you supplied yourself — and loads between none and three of them.
- Replies appear as they are written, with twenty messages retained. Sufficient to follow a train of thought across several questions, and deliberately short of becoming a filing cabinet.
- Every placement reported with its own numbers. A link that appears is listed together with the rating of the site it sits on and that site's traffic, rather than as a bare address to look up separately.
- Full-text search across every message. The function that matters at handover: what was decided about this term last spring, and on what grounds?
For a subsidiary this is worth more than any single feature of the analytics. A record that survives a departure converts three years of scattered effort into something a successor can continue. Where the underlying campaign work sits — term selection, link building, page recommendations — is described in our content strategy, and the technical prerequisites under technical SEO. Which capability terms are worth pursuing at all is settled first, during keyword research.
Open the German property in the dashboard
Questions from German subsidiaries
Our German pages are translations of the corporate site. Is that a problem?
For the corporate and heritage pages, no. For the pages meant to produce enquiries, yes. A translated capability page competes against pages written by people who know how a German purchasing engineer describes the requirement, and it loses on phrasing rather than on substance. The pragmatic route is to translate the corporate material and write the four or five commercial pages from scratch, using the vocabulary that appears in incoming enquiries.
Head office wants one number for Germany. What should it be?
Enquiries with a recorded origin of "website", compared against the same period last year, stated alongside the average contribution per enquiry. Producing it costs more work than reading off a session total, and it is the only formulation that holds up when somebody asks a second question. Where a single visibility figure is unavoidable, the consolidated score across tracked terms will do the job — on the condition that everyone present understands it measures presence in results and not money.
Our terms have almost no search volume. Is any of this worth doing?
That depends entirely on order value. At twenty monthly searches and a five-figure average order, one additional win a year justifies considerable effort. The useful question is not how much volume exists but how many additional orders make the activity pay — and that figure lives in the business, not in any tool.
Should we build a page for every district or county we serve?
Only where there is something specific to say. Forty location pages differing by a place name get judged as interchangeable and are partly not indexed at all. In technical B2B the capability page is far more effective, because someone searching for a manufacturing process filters distance themselves. Five good capability pages comfortably outperform forty location variants.
What happens when our marketing manager leaves?
Whatever is written down survives; everything else does not. In practice that means three things worth maintaining deliberately: a record of which terms were chosen and why, an enquiry log with recorded origin, and access issued to individual addresses rather than a shared password. All three cost very little while somebody is present and are impossible to reconstruct once they are not.