You exhibit in Cologne once a year. Your buyers live here all year. That asymmetry is the whole problem: you are being judged in a market you cannot observe, using a language you may not read, on a schedule set by somebody else's calendar.
An exporter with a stand at interzum, a software vendor at gamescom, a food producer at Anuga — all three share the same situation. Their website is in English. Their prospective buyers search in German. They have a Search Console property, and it reports numbers that look plausible and answer none of the questions that matter to them.
The good news is that the necessary answers are already inside the data. They are simply not in the default view. Getting to them requires three specific slices and a willingness to ignore the headline figure entirely.
Measuring a market you visit once a year
The instinct of most head offices is to compare Germany against other territories using a single number: sessions, or clicks, or a share of total traffic. That comparison is close to meaningless when one territory has a physical event driving a concentrated burst of interest and the others do not.
A ninety-day window that contains a Cologne trade fair and a ninety-day window that does not are two different measurements of two different situations. Presented side by side with a percentage change between them, they produce a conclusion that is arithmetically correct and commercially wrong.
The country and device split is not an appendix
For a business whose buyers are geographically concentrated and whose website is not, the country breakdown is the primary view rather than a supplementary one. It answers the question the headline number cannot: of everything that happened, how much of it happened in the territory that matters.
Reading total impressions
Global impressions rise, the team celebrates, and the German figure is flat or falling underneath.
- Filter to Germany before drawing conclusions
- Then filter again to the terms that carry intent
Device split before the fair
Pre-fair research happens on desktop; on-site lookups happen on a phone in a hall with poor reception.
- Mobile weight rises sharply in fair week
- Page weight matters more than usual there
The device observation is not a general best-practice remark. In the halls at Koelnmesse, connectivity is uneven and visitors are standing up, holding a coffee, comparing two suppliers in ninety seconds. A page that takes eight seconds to render is not slow in that context — it is absent.
Impressions without clicks usually means a language gap
Here is the pattern that appears most often in the accounts of foreign exhibitors: impressions from Germany are substantial, clicks are not. The reflex explanation is ranking. The actual explanation is usually that the result was shown, read in German search context, and understood to be a foreign supplier who may not deliver, may not invoice in euros, and may not answer in German.
What the impression-to-click ratio is telling you
The same shape in a chart, three unrelated causes, three different pieces of work.
- High impressions, low clicks, German queries. A trust and clarity problem, not a ranking one. The listing does not state delivery territory, invoicing currency, or language of support — and a buyer will not click to find out.
- High impressions, low clicks, English queries from Germany. Usually procurement staff at international firms. This is your most valuable segment and the one most often dismissed as noise.
- Low impressions, high click rate. The page is well matched but barely visible. This is the easiest situation to improve, because the message already works.
- Impressions rising, click rate falling, position stable. New queries have attached themselves to the page. Check what they are before treating the falling rate as a decline — frequently the page has started appearing for terms it should never have matched.
The three facts worth stating explicitly in title and description are delivery territory, invoicing arrangement, and language of support. None of them are keywords. All three change click behaviour measurably, because they remove the reason a cautious buyer had for not clicking.
Which view answers which question
The Search Console analytics inside the Semalt workspace is arranged as eight separate views rather than one configurable report. For a team working at a distance from the market, that arrangement is helpful, because each view corresponds to a question somebody in the business actually asks.
| Question asked by | The question | The view that answers it |
|---|---|---|
| Head office | Is Germany worth the stand fee? | Property dashboard with clicks, impressions, trend and health flags |
| Sales | Which products are people looking for here? | Keywords with position history and click trend |
| Marketing | Which page is doing the work? | Pages, with the same metrics cut by landing page |
| Anyone, in fair week | Are we visible on phones in the hall? | Devices breakdown |
| Export management | Is the demand actually German? | Countries breakdown |
| Whoever owns the budget | Are we gaining or losing ground? | Keyword dynamics: entries and exits from the top 3, 10 and 30 |
The last row is the one to build a routine around. Everything above it describes a state; only that view describes a direction. A report that says the average position is 14.3 produces a nod. A report that says seven terms left the top ten last week produces a decision.
One practical note on the eight views: they do not need to be looked at with the same frequency. The property dashboard is a monthly view. Keywords and pages are fortnightly. Devices and countries are worth a look before and after the fair and rarely in between. Keyword dynamics is the only one that rewards a weekly habit, and twenty minutes a week is enough for it. Teams that attempt to review everything weekly stop reviewing anything after two months — the schedule matters as much as the data. All eight sit inside the same Semalt workspace as the campaign itself, which removes the step of exporting one system to compare it with another.
What Search Console cannot see
Search Console describes your own property and nothing else. It does not know which German suppliers appear alongside you, how authoritative their domains are, or whether a new entrant has moved into your term space. The position-tracking side of the workspace covers exactly that gap across six views.
- Ranking score per site. A consolidated figure across tracked terms, alongside average position and keyword count — useful when a head office wants one number per territory.
- Top keywords by volume and current position. The list from which priorities are set, ordered by what people actually search rather than by what the catalogue calls things.
- Best-performing pages. Rarely the pages the marketing team expects. Often a specification sheet or a delivery-terms page nobody considered strategic.
- Competitors sharing your keyword field. Domains appearing for the same terms, with Domain Authority and the count of shared keywords — the fastest way to identify a German competitor you had not heard of.
- Global ranking across the portfolio. For exporters running one domain per market, a single ranked list instead of six separate reports.
The competitor view has a particular value for exhibitors. A German buyer preparing for a fair rarely searches for a company; they search for a capability. The domains that appear for that capability are your actual competitive set for that week, and they are frequently not the companies whose stands are near yours.
There is a second use for the same view that exhibitors tend to discover late. Because it lists domains by shared keywords rather than by category, it surfaces distributors, trade portals and industry directories alongside manufacturers. For a foreign supplier, those intermediaries are often the more realistic route into German search results in the first year — a listing on a portal that already holds the term will produce enquiries sooner than a page competing against it. The campaign side of My SEO treats those as placement opportunities rather than competitors, which is usually the correct reading.
The competitor you can see
The firm two aisles over with a bigger stand. Easy to watch, and rarely the one taking your enquiries.
- Shares your audience for three days
- Shares your search terms rarely
The competitor you cannot
A domestic supplier with no stand, holding the capability terms your buyers search before they travel.
- Appears in the shared-keyword column
- Domain Authority tells you how hard the term is
Comparing the two columns is uncomfortable and useful. Stand budgets are set against the visible competitor and enquiries are frequently lost to the invisible one, because the shortlist is assembled before anybody arrives in Cologne. Where the shared-keyword count is high and the authority gap is small, the term is worth writing for. Where the gap is large, the honest answer is that a page will not close it within a single fair cycle, and the budget belongs somewhere else.
Working backwards from the fair date
Because the date is fixed and the reporting is delayed, everything has to be planned in reverse. The following schedule assumes the standard expectation that a change takes four to eight weeks to become visible, with a margin for the fact that visibility and traffic are not the same thing.
| Weeks before | Work | What you should be able to see |
|---|---|---|
| 12 | Country and device baseline recorded; competitor set identified | a written starting point, not a feeling |
| 10 | Two pages written for the capabilities you can realistically win | pages live and internally linked |
| 8 | Pages submitted for indexing, internal links from the home page | bot visits logged, status recorded |
| 6 | Titles and descriptions state territory, invoicing and support language | click rate on German queries begins to move |
| 4 | Mobile weight of the fair-relevant pages reduced | device split stabilising |
| 2 | No further changes; observation only | a clean measurement window |
The final row is the one most often ignored and the most important. Changing pages during the two weeks before a fair destroys the ability to interpret what happened during it. Whatever you learn will be attributed to the wrong cause, and the same mistake will be repeated the following year.
A related point about the two-week freeze: it applies to the pages, not to the observation. Recording what you see during the fair — which queries spike, which device split shifts, which page receives the traffic — costs nothing and produces the baseline for next year. Most exhibitors return from a fair with a stack of business cards and no record at all of what happened in search during those five days, and then rebuild the same guesswork twelve months later. Fifteen minutes of screenshots on the Friday is worth more than a week of analysis the following spring, if only because the two-day reporting lag means the full picture will not exist until after you have gone home. The Semalt dashboard keeps those figures without a manual refresh, since background workers synchronise the data continuously.
Reporting to people who are not in this market
A report written for a head office in another country has a different job from a report written for the person doing the work. It has to survive being read by somebody who has never seen a German search result and who is comparing this territory against five others.
Keep the summary to figures with an attached action. German clicks year on year, the three pages carrying the result, entries and exits for the period, and one sentence of explanation is usually the whole of it. Everything else — full keyword lists, country splits, competitor tables — belongs in the export, where the working level can use it. Reports can carry your own logo and colours, which matters more than it should: a document that looks like a company document gets forwarded, and a tool export does not.
Reducing the distance between question and answer
The recurring difficulty for a team outside the market is not access to data but the time it takes to interrogate it. Somebody in Milan or Manchester has a question about German queries at four in the afternoon and gets an answer three days later, by which point the question has changed.
A project timeline instead of a mailbox
Answers, automated reports, new links and open tasks in one chronological record per project.
- Answers grounded in project data. A routing step decides, per question, which data blocks to load — Search Console, SERP, campaign, user-supplied — anywhere from none to three depending on relevance.
- Streamed responses, twenty messages of history. Enough context to continue a line of questioning without restating it, short enough not to become an archive.
- New links with donor metrics. Each newly created link arrives with the referring domain's rating and traffic rather than as a bare URL.
- Full-text search across every message. The function that earns its keep in year two, when somebody asks why a decision was taken before the last fair.
None of this replaces judgement about the German market. It shortens the loop between a question and a defensible answer, which for a team operating at a distance is the difference between deciding and guessing. The underlying campaign work — term selection, link building, page recommendations — sits in the same place, and our keyword research approach describes how the initial term list is built for a market you do not live in. The technical prerequisites are covered separately under technical SEO, and the practical checks for a local presence under local SEO.
If you would rather see the country split, the keyword dynamics and the competitor set against your own property than read a description of them, the workspace shows all three side by side once the property is connected.
Open your Search Console data in the dashboard
Frequently asked questions
Our site is in English. Do we need German pages to be found by German buyers?
For terms with commercial intent, in most categories, yes. English pages can rank for English queries made from Germany, and that segment is real — it is largely procurement staff at international firms. But the majority of category searches happen in German, and an English page competing against German pages for a German query starts at a disadvantage that no amount of technical work removes. Start by measuring how much of your German impression volume comes from each language before deciding how much to invest.
How far in advance of a fair should we start?
Twelve weeks is a workable minimum, and the reason is arithmetic rather than caution. A new page needs to be discovered, crawled and indexed, which alone can take days to weeks. After that, four to eight weeks typically pass before movement becomes distinguishable from noise. Starting eight weeks out is possible; starting three weeks out means you are paying for a stand and hoping, which is not a plan.
Why does our click rate fall while our position improves?
Almost always because new terms have started matching the page. They generate impressions and no clicks, which pulls the average down while the terms you care about are unaffected. Sort the keyword view by impressions and look at the recent entries with zero clicks. If they are off-topic, the falling rate is a side effect rather than a problem — although it is worth asking why the page is matching them at all.
Head office wants one number for Germany. What should it be?
Clicks from Germany on commercially relevant terms, compared against the same period last year, with the fair date marked. It is more work to produce than a raw session count and it is the only version that survives a follow-up question. If a single consolidated figure is genuinely required, the ranking score across tracked terms serves that purpose — provided everyone understands it describes visibility rather than revenue.
Is it worth tracking competitors we never meet at the fair?
Yes, and they are frequently the more important set. Buyers preparing for a fair search by capability rather than by company name, and the domains holding those terms shape the shortlist before anyone walks into a hall. A domestic supplier who does not exhibit at all can still take the enquiry, simply by being the result that appears when someone searches the capability on the train to Cologne.
