Many small firms look at their website traffic once a month, see a few hundred page views and not a single form enquiry – and draw the obvious conclusion: the site is doing nothing. Meanwhile the phone keeps ringing. The gap is in the measurement, not in the website. When people have a problem with a product or service they bought, 89 percent reach for the phone and 48 percent use a contact form or customer portal (Bitkom 2016). The most important route to a job therefore shows up in no report at all. This article shows how to attribute calls to the website without collecting customer data: the tap on the phone number as a countable event, the difference between a tap and a conversation that actually happened, one question at the phone that takes five seconds to answer, the reading by hour and weekday – and the point at which call tracking becomes awkward under data protection law.
Why a website looks useless even though it produces calls
The usual report for a small website counts page views, time on site and – on a good day – submitted forms.
What it does not count is the route that most often leads to a job for a tradesperson, a workshop, a practice or a law office: the phone call. When people have a problem with something they bought, 89 percent reach for the phone, 72 percent visit a branch, 58 percent write an email and 48 percent use a form on the website (Bitkom 2016); several routes could be named. Looking only at the form count therefore measures the smaller share and treats the larger one as non-existent. On top of that, online research and an offline purchase have long belonged together: a substantial share of in-store buying decisions is prepared by prior online research. The website is working – its effect simply ends in the handset rather than in a statistic. How a firm turns that into enquiries systematically is covered in the article on the tradesperson website that generates enquiries.
This gap has consequences beyond a wrong gut feeling. Anyone who believes the website does nothing cuts in the wrong place: maintenance gets postponed, content is deleted, or the very service page that triggers most calls is left to go stale. Conversely, time flows into pages almost nobody visits. Small and medium-sized firms make up roughly 99 percent of all companies in Germany (Federal Statistical Office) – businesses, in other words, where nobody spends half a day analysing numbers. All the more reason for the few numbers you do look at to reflect the real route to a job. Which figures actually trigger a decision is set out in the article on website metrics for small businesses.
The first step out of the gap is surprisingly small: a tap on a phone number is an event like any other and can be counted. On a phone this route is the normal case – 70 percent of mobile searchers have called a business straight from the search results (Google/Ipsos 2013), and 98 percent at least rarely feel the need to call a business while searching on a smartphone (Google/Ipsos 2013). For local services, arranging an appointment is the most frequently named reason to call: 71 to 76 percent name it (Google/Ipsos 2013). Counting these taps shows for the first time which page triggers calls – the home page, a particular service page, the contact page or the article about a common repair.
The phone number belongs on every page as a link
A tap can only be counted if there is something to tap. A phone number that sits on the page as plain text is not a calling route on a mobile device but a transcription job: select, copy, switch to the phone app, paste.
Every one of those steps loses people. Set up as a link – technically a tel link – a single tap opens the dialler right away, and that tap is exactly the event that can be counted. On a desktop computer either nothing happens or a phone application starts, depending on the setup; no harm is done, because the number remains readable there. The effort is a one-off and takes minutes, while the insight lasts. How much weight mobile visits now carry is covered in the article on mobile usability for a business website.
Check the number in four places
A tap is not yet a conversation
As soon as the first numbers arrive, the second insight follows: a counted tap is not the same as a conversation that took place. Several stages sit between the two, and something can go wrong at each of them.
Someone taps by accident, cancels the dialling, ends up in a queue, hears a busy tone or hangs up because nobody answers. In the other direction there are calls that produced no tap at all: the number was read off the screen and typed into another device, dialled from the address book or copied from a note. The tap is therefore an approximation, not a cash book – and as an approximation it is still worth considerably more than no number at all.
Tap on the number
The countable event on the website. It says: someone started a call here, and on this particular page.
Connection established
Only here does the phone ring in the business. Drop-offs before the ring stay invisible in every report.
Somebody picks up
The point where most enquiries are lost: outside opening hours, during an appointment, out on site.
The matter is named
Only the conversation reveals whether this is a job, a follow-up question, an appointment or a complaint.
Origin clarified
One short question assigns the call to the website, the search, a recommendation or the sign on the van.
Job or appointment
The result that counts. It ends up in the calendar and the order book, not in the traffic report.
A simple division of labour follows from this chain: the website counts how often a call was started. What became of it is known only to the person who held the handset. Together the two numbers form a picture; on their own each of them misleads. A firm with forty taps and six recorded jobs per month knows more about its website than one with four hundred page views and no idea what happened next. If both numbers stay low although visitors arrive, it is worth looking at the usual causes in the article on visitors but no enquiries.
Scroll table sideways
| Question | Tap on the number | Note after the call |
|---|---|---|
| Which page triggers the call? | Shows the page where the tap happened | Rarely remembered during the call |
| How many calls arrived? | Counts attempts, not connections | Counts what actually rang |
| What was it about? | Stays open | Job, appointment, question or complaint |
| Phone or computer? | Readable from the device type | Irrelevant during the conversation |
| Did it turn into a job? | Stays open | Ends up in the calendar |
| How did the caller know the firm? | Only if a tap really happened | One question, one answer |
The one-sentence question at the phone
The most reliable attribution costs nothing and takes five seconds: “May I briefly ask how you came across us?” Ask it at the end of the call, once the matter is settled – not at the start, where it feels like an obstacle.
Most callers answer willingly and with surprising precision: through the search, through the website, through a recommendation, through the van, through the sign on the building. What matters is that the question stays open and does not suggest an answer. Ask “You found us online, I assume?” and you get a confirmation instead of information.
For the record a tally sheet with five fields next to the phone is enough: search, website directly, recommendation, existing customer, other. One stroke per call, one sheet per month. Anyone already recording digitally sets up the same five fields as a selection in the job form. What counts is not the accuracy of any single assignment but the ratio across a hundred calls. Deliberately record no names and no phone numbers in this list – it is a count, not a customer file, and stays outside data protection law from the outset. How this view of the numbers condenses into a monthly routine is described in the article on metrics for small businesses.
After four weeks you have a number nobody had before: the share of calls that came via the website or the search. Put it next to the counted taps and the arithmetic either fits or stands out. If thirty taps add up in a month and twenty-two callers name search or website as their route, the picture is consistent. If forty callers name the website although only twelve taps were counted, the number is evidently being read off rather than tapped – a sign that it sits in a hard-to-reach place, is not set up as a link, or is mostly read on a desktop computer.
Reading calls by hour and weekday
The second reading that pays off immediately is the one over time. Taps on a phone number carry a timestamp, and in small firms calls rarely spread evenly.
Two waves are typical: early morning, before the caller's own working day begins, and late afternoon after work. In many trades a Monday peak comes on top, because the weekend has piled up whatever was left over on Friday. Anyone who knows this distribution can lay it against their own staffing – and sees straight away whether the hours with the most calls are also the hours when somebody is at the phone. Firms whose order book is full anyway use the same reading to steer enquiries rather than turn them away; the article on managing demand through the website shows how.
The question is not whether you want to be reachable around the clock – for a small firm that is neither feasible nor necessary. The question is whether the availability you promise matches the actual calling times, and whether the website states those times correctly. A stated availability from 8am to 5pm on the contact page helps little if the strongest hour is 5pm. How opening hours and callback times stay current in the long run is shown in the article on opening hours and availability.
- Read the taps by hour and mark the two strongest time windows.
- Lay your own staffing underneath as a second row: who actually answers, and when?
- Name the gaps – lunchtime, days out on site, holiday cover, Friday afternoon.
- State callback times on the website instead of promising constant availability.
- Match the voicemail message to the times that really apply.
- Offer a callback form for the fringe hours that works without a conversation.
- Measure again after four weeks: has the distribution shifted?
- Think about the season – heating failures in winter and garden enquiries in spring move the peaks.
Two rows of numbers, one decision
Missed calls: the most expensive invisible number
No value is recorded so rarely and costs so much as the missed call. It leaves no trace in the website report, no message in the inbox and frequently no second attempt either.
How often it happens is clear from the customer side: 44 percent of those who had bad experiences with a customer service reached nobody or received no answer, and for 75 percent making contact was time-consuming and cumbersome (Bitkom 2016). For a small firm the arithmetic is uncomfortably simple: a missed call equals an enquiry that does not make it into a quote – and because 89 percent reach for the phone when a purchase causes a problem (Bitkom 2016), this is precisely the route where most is left lying.
Counting missed calls is easier than it sounds. The mobile phone keeps a list, every phone system keeps a list, and even an answering machine shows how often it kicked in. It is enough to note two figures once a week: calls answered and calls missed. A share of one in five is not rare in practice and goes unnoticed as long as nobody looks. What matters is calling back the same day, because in the meantime the customer dials the next number. How strongly response time decides the job is covered in the article on response time for enquiries.
A missed call is an enquiry that happened but appears in no report – and therefore does not get fixed.
Data protection: where the data-minimal line runs
In attributing calls there is a point at which the effort stops being technical and becomes legal.
Methods that assign every visitor an individual phone number in order to link a later call to a session work with personal data: a caller's number is, under Article 4 number 1 (GDPR), information relating to an identifiable person, and combining it with browsing history creates a profile. If identifiers are stored on or read from the terminal device for that purpose, Section 25 (TDDDG) applies as well – then consent is required, and with it returns the consent banner that many firms have only just got rid of. What triggers a banner and what does not is set out in the article on the website without a cookie banner.
The data-minimal line runs where you count instead of attribute. A counter that records twenty-seven taps on the phone number on a service page processes no information about individual people and therefore needs neither consent nor a banner. Article 5 paragraph 1 letter c (GDPR) requires anyway that no more data be collected than the purpose needs – and the purpose here is to see which page triggers calls. Totals are enough for that. Anyone who still wants to work with rotating phone numbers needs a legal basis under Article 6 (GDPR), a note in the privacy policy under Article 13 (GDPR) and a data processing agreement. For a firm with twenty calls a day, that effort is typically out of proportion to the insight gained.
A second point concerns the consistency of the number to the outside world. Changing numbers in search results, directories and the business profile make local discovery harder, because name, address and phone number are read as one connected signal; the guidance for the business profile therefore requires the main number to remain listed (Google documentation). A fixed number that is the same everywhere is the calmer solution for small firms – and with counted taps nothing is lost, because the link to the page is created on the website and not in the phone system. How to keep that profile in good shape is shown in the article on the business profile at the search engine.
The callback form as a safety net
Outside your reachable hours you need a second route that does not ring.
A callback form is the leanest version: name, phone number, a time window to choose from, plus one line about the matter – more fields lower the completion rate without improving the answer. What matters is the promise beside it: “We call back on working days until 5pm, usually within a few hours” is more concrete and more honest than “We will be in touch as soon as possible”. And because a submitted form is a countable event too, the measurement gap closes by itself at this point: tap, callback request and form enquiry then stand side by side. What matters when building one is described in the article on the contact form; how to sort enquiries in advance is shown in the article on the price estimator on your website.
A callback form also takes pressure off when many calls are about the same thing. Existing customers often need no conversation at all, just a form for the recurring case – a maintenance appointment, a repeat order, a fault report. How to map that cleanly is covered in the article on service forms for existing customers. Every call that becomes unnecessary this way frees the line for the conversations that really do turn into jobs. That, too, is a form of measurability: whatever arrives as a form is counted before anyone works on it.
How XICflow brings together calls, forms and pages
In XICflow the phone number is not text but a building block: it sits in the header, on the contact page and in the footer, is set up as a calling link everywhere and is maintained in a single place.
Every tap on it is counted – as are submitted forms and route requests – and the figures sit right next to the page views. So you see not only how many people visited the site, but which page makes the phone ring. The measurement works without cookies and without recognising individual visitors, so it needs no consent banner. Which building blocks are included is shown in the overview of XICflow services; how a site comes about is described in how XICflow works.
What XICflow does not take over is the conversation itself. The question about origin is still asked by the person at the phone, and whether an enquiry becomes a job is decided in the business. The technology makes sure the way there is short: a number that can be tapped, a callback form for the fringe hours, opening and callback times maintained in one place, and a fast, mobile-friendly page. How finished sites look can be seen in the example websites, and what building one costs is shown in the price overview. At the end there is a figure that was missing before: not just how many visitors came, but how often that turned into a call.
Sources and Studies