In short and in plain language
Is your website bringing enquiries? This article shows small firms how to measure it in a simple way.
A few numbers a month are enough. How often are you seen in search? How often is your main page viewed? How many enquiries come in?
Compare each month with the month before. A single number says little.
You can usually count a lot without consent. This includes calls, emails and forms.
Decide first what you will change for each result. We set up the counting and read the numbers with you.
A website is ordered quickly, and then something odd happens: it simply runs. Month after month, with nobody looking at it. At some point somebody in the business asks the question that unsettles every investment - „is this actually bringing us anything?“. That question can only be answered with numbers. 79.02 percent (Eurostat) of enterprises in the EU with ten or more employees now have their own website, and in Bavaria the figure reached 92 percent (Bayerisches Landesamt für Statistik) of enterprises with at least ten persons employed in 2025. Evaluating data in the business is rare, though: only 33.02 percent (Eurostat) of enterprises with ten or more employees analyse data with their own staff at all, and among small enterprises with 10 to 49 employees just 11.45 percent (Eurostat) use business intelligence software. This article shows which four monthly numbers are enough for a small business, which of them you may collect without consent, and how you recognise that a change to the site has paid off.
Key takeaways
- Four numbers are enough: visibility in search, views of your most important service page, contact routes opened, and enquiries that arrived - read off on a fixed day monthly.
- A single monthly figure says little. It becomes readable next to the previous month and, once available, next to the same month a year earlier.
- More is possible without consent than people think: what the server logs anyway, and what you count yourself - calls, emails, forms -, does not touch the consent requirement of Section 25(1) (TDDDG); the data protection rules for further processing still apply.
- A number without a consequence is wasted time. Decide in advance what you change at which result, otherwise the monthly sheet turns into filing.
Why hardly anyone looks
The gap in data analysis is documented, and it is not a German peculiarity. In 2025, only 33.02 percent (Eurostat) of EU enterprises with ten or more employees analysed data with their own staff. Adding external providers brings the figure to 39.85 percent (Eurostat) - still not one in two. And where analysis happens, it rarely concerns customers: 17.62 percent (Eurostat) of enterprises analyse data about their customers. The gap is starkest in software. Among small enterprises with 10 to 49 employees, 11.45 percent (Eurostat) use business intelligence software, compared with 69.24 percent (Eurostat) among large enterprises. The size difference is no accident: large organisations have somebody whose job it is, while in a small business evaluation is one task out of forty and keeps sliding to the end of the week.
No fixed appointment
Without a date in the calendar, the review does not happen. It is not urgent, it does not announce itself, and in a full week it loses every contest for attention.
Worry about consent
Many skip measurement entirely because they want to avoid a banner. Yet part of the numbers can be collected without anything being stored on or read from the visitor's device.
Too many metrics at once
Anyone opening an analytics tool for the first time sees dozens of charts and knows less afterwards than before. Four numbers that belong together carry further than forty that merely sit side by side.
The German picture confirms this. In Bavaria, 92 percent (Bayerisches Landesamt für Statistik) of enterprises with at least ten persons employed had their own website in 2025 - the site itself is long since the normal case. A countable conclusion rarely sits on it: only 27 percent (Bayerisches Landesamt für Statistik) of enterprises with their own website provide an online ordering, reservation or booking system. And 14 percent (Bayerisches Landesamt für Statistik) of enterprises there use business intelligence software, the kind of tool that evaluates company data. For a business without a booking system, success does not show up in a till but in calls, emails and completed forms. No tool counts those events by itself: they end outside the website - in the inbox, on the phone, in the calendar. Anyone using online appointment booking has it easier here: the conclusion happens on the site itself.
Four numbers are enough: the funnel for small businesses
Between the first search result and the enquiry in your inbox there are four stations. At each you lose part of the interested audience, and that is normal. What matters is not the individual number but the point where the drop is noticeably steeper than at the others. That is where work on the site pays off, and where it evaporates if you decide without measuring. Write down all four values and a pattern emerges after two or three months - no additional tool needed, just a sheet of paper.
- Visibility in search: how often did the page appear in the results, how often was it clicked? The search service produces this number and shows whether you make the shortlist at all. If it rises without the views following, title and description are the problem.
- Views of your most important service page: the home page is not the metric, the page describing the service is. That is where an enquiry is decided, so it deserves its own line.
- Contact routes opened: how often was the form opened, how often was the phone number tapped, how often were the directions viewed? This step shows intent before an enquiry exists.
- Enquiries received: calls, emails, forms, bookings - added up, without judgement. Quality comes later; first the quantity counts, because only it can be compared across months.
In practice, the jump from step two to step three is the most expensive, and it has to do with content. The most common function of a company website in the EU is describing goods and services with price lists: 66.36 percent (Eurostat) of enterprises offer it. Anyone providing no orientation expects a trust that readers need not extend elsewhere - details in the article on whether prices belong on your website. The second brake is the form: every additional mandatory field costs conclusions, and the article on the contact form with fewer fields explains which fields can go.
Work through one month once
What you may count without consent
The legal position is more precise than everyday practice suggests. Under Section 25(1) (TDDDG), storing information in the end user's terminal equipment and accessing information already stored there are permitted only if the end user has consented on the basis of clear and comprehensive information. The trigger is therefore the visitor's device, not evaluation as such. Cross that line and, under the fine provision of the same act, you risk a fine of up to 300,000 euros (TDDDG). What the server logs in the course of delivery does not access the terminal equipment - the data protection requirements for the subsequent processing, for instance retention periods and the truncation of address data, continue to apply independently. Classifying an individual case belongs in professional hands; the essentials are in the article on the cookie banner for small websites and in the GDPR basics.
Three numbers nobody needs a banner for
How the web as a whole measures can be quantified, and the figures explain why the consent question feels so present. On 64 percent (Web Almanac) of the desktop pages examined, a reach measurement service is running, with a single provider accounting for 53 percent (Web Almanac). At least one third-party service that sees visitor data is embedded by 75 percent (Web Almanac) of all pages examined, and the measurement cookie of the most widespread service sits on 46 percent (Web Almanac) of pages. That is the normal state of the web, not a requirement for your business. If you only want to know whether more enquiries arrive than last year, you do not need this apparatus - and save part of the banner logic with it.
| What you want to know | Where the number comes from | Consent under Section 25 TDDDG |
|---|---|---|
| How often does the page appear in search? | The search service's report on your own domain | not affected, the number arises at the search service |
| Which page was delivered how often? | Hosting server log, evaluated in aggregate | not affected as long as nothing is stored on the device |
| How many enquiries arrived? | Counted at form submission, in the inbox, in the calendar | not affected, no information on the terminal equipment |
| Which contact route was opened? | A separate sub-page per contact route or a server-side redirect | not affected when no identifier is used |
| How does a visitor move through the site? | Measurement script in the browser with its own identifier | required |
| Where did a visitor come from across several visits? | Recognition via a cookie or a stored identifier | required |
The monthly sheet: one page, four fields
The form matters more than the tool. A monthly sheet is one row per month in a table that belongs to you and that you fill in within ten minutes. It lives in the business, not in somebody else's account, and it survives every change of tool. Open a tool each time instead, click through views and note nothing, and after a year you hold not a single comparison. The list below describes what belongs on the sheet.
- A fixed date: the third working day of the following month, as a recurring calendar appointment
- Four numbers: search impressions, views of the most important service page, contact routes opened, enquiries received
- The previous month in the column beside it, plus the same month a year earlier once it exists
- One line of notes: what was different this month? Public holidays, school breaks, an advert, a site rebuild, a storm
- One sentence on the consequence: what will you change before the next date, and how will you tell whether it worked?
- Filed as your own table, not only as a view in a provider's tool
That last line is a precaution, because analytics tools do not keep their data indefinitely. The default view of the search performance report shows click and impression data for the last three months (Search Console-Hilfe); to look further back you change the date range and export. Behavioural analysis is similar: data on age, gender and interests is retained for 2 months (Google-Hilfe zur Datenaufbewahrung) regardless of the settings, and data from cross-device recognition for at most 26 months (Google-Hilfe zur Datenaufbewahrung). A three-year comparison therefore comes from your own exports, not from the tool. Anyone filing them should also know who owns the associated accounts and files - see the article who owns your website.
Export while the data is still there
The phone first, the rest after
Anyone reading their numbers should look at the mobile view first: that is where most of the traffic arises. Almost 9 out of 10 (Eurostat) internet users in the EU went online with a mobile device in 2025. The consequence is uncomfortable: a page can look impeccable at a desk and still lose enquiries, because the phone number hides behind a menu on a handset or the form jumps while you type. It may also be too slow: on mobile devices, 48 percent (Web Almanac) of the websites examined reached good values on the Core Web Vitals in 2025 - a page that appears too late loses interested people before anything is counted. What helps is in the articles on mobile optimisation and loading speed.
Read it split by device
Record the four numbers monthly, separately for mobile devices and desktop. Only the split shows whether a drop affects everybody or only the half operating with a thumb.
Count the call
On a handset the call is the shortest route to an enquiry. Not counting it underrates your own site. A tally sheet by the phone for four weeks gives a reliable order of magnitude.
Put the local listing next to it
Part of the enquiries do not arise on the site but in the map service. The business profile listing has its own numbers, and they belong beside the monthly sheet.
A word on the Web Almanac figures in this article: they do not come from a survey. For the Web Almanac, metadata from over 16 million (Web Almanac) websites is collected monthly and made publicly available - those percentages are therefore measured shares of pages examined, not self-reporting by respondents. The figures from Eurostat and from the Bayerisches Landesamt für Statistik, by contrast, are survey data: they rest on questionnaires put to enterprises and households and reflect what the respondents report about their own business. Both kinds are dependable for the question of what is common in general, and unsuitable for the question of what is common in your trade in your town. For the second question there is only one source: your own four numbers across twelve months.
The hardest metric is not the one nobody can measure, but the one somebody writes down every month and nobody reads afterwards.
Five mistakes that make the numbers worthless
Evaluation rarely fails on technology. It fails because the numbers are not comparable, are read too late, or must carry a statement they cannot carry. The five patterns below cost the most in practice - and all five can be avoided when the measurement is set up.
- The comparison is missing. A number on its own is not a statement. 486 views are many or few depending on what stood there the month before. Without a reference value, every number becomes a story matching the mood of the day.
- The period wanders. Look at four weeks once and a calendar month the next time, and you are comparing 28 days with 31. That is a good ten percent of difference before anything has changed.
- Your own visits are counted. Anyone opening the site daily produces part of their own views. Your own addresses belong excluded, otherwise the metric grows with your busyness.
- The change is not noted. If the home page was rebuilt, an advert placed and a service added in one month, no effect can be attributed afterwards. One line of notes per month prevents that.
- No decision follows. Numbers from which nothing follows are a hobby. Every fixed date needs a sentence on what changes before the next - even if that is the deliberate decision to change nothing.
A sixth point concerns not the measurement but its use. Anyone turning good numbers into an advertising claim must be able to substantiate it - for environmental claims, separate rules apply from September, described in the article on the new rules for environmental claims. And anyone using the numbers for a campaign should separate the destination page instead of sending everything to the home page; why that makes evaluation possible at all is in the article on the landing page for adverts.
How we set the measurement up
We build websites at a fixed price and set the counting up at the start, because doing it afterwards is more work. That includes a clearly named service page as the destination, a separate address per contact route so the path shows in the server log, and a form whose submissions are counted on sending. Reading that off and entering it into the monthly sheet is part of our website maintenance - which also describes what else needs doing continuously, and why a site without care slowly loses visibility, as the article on website maintenance shows. Which building blocks are included in which plan is set out in our packages.
For existing sites we start with an inventory: which numbers already exist, which can be collected without consent, and which step of the funnel has been blind? Out of that comes a table with four columns and a date in the calendar - nothing more in the first six months. Anyone wanting to know more can measure deeper later; the order matters, not the scope. What we take on is listed under services, the costs stand openly on the pricing page, and the ongoing evaluation runs through the same website maintenance that covers updates and backups.
Sources and studies
Related Articles
Title and Description: Your Own Entry in Search Results
A Google entry has three parts: address line, title link and snippet. What you can specify, what Google assembles itself and how to check the outcome.
How Long Does It Take Google to Find Your Website?
From launch to an index entry Google states one or two days to a few weeks. What affects the pace, what does nothing, and what to check on launch day.
Get Found by ChatGPT and AI Search with a Small Website
Why small businesses, clubs and freelancers without a clean website are missing from AI answers, and how a starter website with FAQ and location data helps.