From click to enquiry: where people drop off along the way

From click to enquiry

When you have traffic but no enquiries, the question is not what to change but where to look. Most writing on this starts with ready answers: the form, the button, the price. This one starts with measurement, because without it every change is a guess. Then it goes through the places where people genuinely drop off, and which of them you can test with data.

In short

• Before anything else, check what exactly counts as a conversion: a click on a phone number is not a completed call, and a submitted form is not a qualified enquiry.
• Where the loss sits is measured, not assumed. Four funnel metrics: how many get past the first screen, how many reach the form, how many start it and how many submit it.
• Phone enquiries in service businesses often go unrecorded, and while they are missing from the sums, those sums lead to wrong decisions.

Table of contents

• First: what counts as a conversion
• The maths of cost per conversion
• How to measure where the loss is
• The places where people drop off
• The first screen
• The price indication
• The form: what to remove and what to keep
• The calls that never reach the reports
• Response time
• The quality of the enquiries
• Store versus service business
• The order of work
• Frequently asked questions

First: what counts as a conversion

This is the question that decides whether every figure after it means anything. In most accounts we open, "conversion" covers several different things at once, which is why the averages are unreadable.

What is not the same thing

What to do

Separate them into distinct actions with different values rather than adding them into one number. Then add one more level that almost nobody tracks: how many of the enquiries are qualified. That is the only figure that genuinely connects the advertising to the business.

Which types of action can be recorded, including calls from ads and offline actions, is set out in the Google Ads help centre.

The maths of cost per conversion

Cost per recorded conversion is cost per click divided by the conversion rate. Both are in the equation, but they do not move with equal ease.

A worked example

From our own account for the Bulgarian market. The figures are for recorded conversions - what exactly sits behind them determines what the percentage means, so check your own account's action list before comparing:

What the formula shows and what it does not

At €0.60 per click, the same rate gives a recorded conversion at about €31. At 3% and an unchanged cost per click, about €24. These become cost per enquiry only if the actions being counted really are enquiries: with mixed actions the sum means something else.

What the formula does not tell you is how much work each of the two changes takes. Lifting the rate may be a matter of one paragraph or of three months' work on the product and the pricing. So do not assume that work on the page is always the cheaper route: it has its own cost and its own uncertainty. The only certainty is arithmetic - at a low rate, every recorded conversion comes out more expensive at the same cost per click. The click itself does not get dearer.

Where the ceiling is

The rate does not rise indefinitely. For high-value services with a long cycle, the range we see in practice is between 2 and 5 per cent. That is a reference point from specific projects, not an industry standard. Planning a budget around these figures is covered in the article on the Google Ads budget.

How to measure where the loss is

This is the substantive part. Without these four figures, everything else in this article is a list of possibilities rather than a diagnosis.

What to record

1. The share getting past the first screen. Shows whether the message holds attention.
2. The share reaching the form. Shows whether there is enough reason between the screen and the form to continue.
3. The share starting the form. The gap against the previous figure shows whether the form puts people off on sight.
4. The share submitting it. The gap shows whether it puts them off while filling it in.

Add clicks on the phone number and the email address, and for a store: add to cart, begin checkout, completed order.

These four funnel metrics localise the drop-off, they do not prove its cause. Few people getting past the first screen does not necessarily mean a weak headline: someone may have found the phone number straight away and called. Few started forms does not mean the form puts people off - the answer may simply have been found earlier on the page. The numbers tell you where to look; the hypothesis gets tested by making a change.

And one technical caveat, because this is where people often go wrong: the standard scroll event in GA4 fires at 90 per cent of the page's vertical depth, in other words practically at the bottom. It measures neither getting past the first screen nor reaching the form - those two metrics need separately configured events.

How to read it

Why per landing page, not site-wide

The site-wide rate mixes brand traffic with cold traffic. Brand traffic usually converts considerably better and lifts the average. The figure that matters is the rate on the page where people arrive from advertising and from organic search.

What period you need

At a few enquiries a month, one month proves nothing. Work in quarters and in absolute numbers. A difference between 1.8 and 2.1 per cent on forty clicks is noise.

The practical consequence: do not make two changes in one period. If you change the landing page headline and the form fields in the same week, the next report will not show which of the two worked, and at low volumes you will not get a second chance to separate them. One change per period is slower but gives answers.

The places where people drop off

The list below follows the order of the journey, not order of weight. Which one matters most in your case is shown by the figures from the previous section, not by a general rule: a broken form may be the whole problem, or it may be marginal.

The first screen

The first screen answers three questions before the person has decided whether to stay. That is a content task, not a design one.

The three questions

• Is this for me? "SEO for online stores with over 500 products" says more than "Professional SEO services". The second applies to everyone and therefore to nobody.
• What exactly do I get? Not benefits, but the content of the service: "monthly retainer: technical audit, content, internal links, reporting".
• What happens now? A call, a quote or a meeting, and how long it takes. People hesitate more over the unknown than over the price.

What gets in the way

A large image with no information eating the screen on mobile. A strapline that does not say what you do. A rotating gallery where the message changes before it has been read. Two calls to action competing with each other. A consent banner covering the content.

The price indication

A page with no price indication at all loses two groups: those for whom it is expensive, and those for whom it looks suspiciously cheap. This is an observation from practice, not a measured quantity: separating the effect of the price from everything else on the page would need a dedicated test with and without an indication.

How to give a price when it depends on many things

1. A starting figure: "from" is indication enough.
2. What changes it: volume, timeframe, number of pages, competition, language.
3. What is and is not included in that figure.
4. How an exact price is reached and how long that takes.
5. Whether there is a minimum commitment period.

What you gain

The indication filters by budget before time has been spent on either side. The person who sees it either leaves straight away or writes better prepared. Put it where the hesitation starts: after the description of the service and before the form, not at the end.

The form: what to remove and what to keep

The rule is not "the fewer fields the better". The rule is: every field either helps you answer or qualifies - otherwise it goes.

What always stays

• Name.
• One way to get in touch, chosen by the person.
• A free field for describing the need.

What earns its place in some cases

A field for quantity, for a desired timeframe, for a budget range, or the option to attach a file. For a technical or project-based service these fields improve the outcome: they save a whole round of clarifying questions and filter out enquiries that will not become projects. The decision depends on whether filtering is worth more to you than volume.

What gets in the way almost everywhere

• Mandatory company registration number, registered address and three phone numbers at the enquiry stage.
• Two email fields for confirmation.
• Long mandatory consent checkboxes with no short explanation.
• A captcha that fails on mobile.

After submitting

A useful confirmation says three things: we have it, we will reply within a stated time, here is a number if it is urgent. And test the chain: send a real enquiry from a phone once a month. This is the check that most often surfaces a hidden defect - mail going to spam, a notification not being sent, an error only in one browser.

The calls that never reach the reports

In service businesses a significant share of enquiries arrive by phone and go unrecorded. That distorts every sum downstream.

What happens without recording

You see one rate while the real one is higher. You decide the channel is not working, reduce the budget and stop something that brings customers. We have seen campaigns shut down over figures that were incomplete.

The minimum

1. The phone number on the site is a link that dials with one tap.
2. Clicks on it are recorded as a separate action.
3. The source of every call is logged in a CRM or a simple spreadsheet.
4. A monthly comparison: form enquiries plus calls against spend.

The question "how did you find us" is not precise and people misremember, but three months of such records change decisions more than it sounds.

Response time

Response time is part of conversion, not only of service.

What happens with a slow response

The person who wrote to you has usually written to one or two others as well. An enquiry you answer two days later is not necessarily lost, particularly for a large or unusual service, but a delayed response can reduce the chance of an order.

The practical minimum

• A reply within the working day, even if only an acknowledgement and one clarifying question.
• One named person responsible for incoming enquiries.
• A prepared first reply asking exactly what you need in order to quote.
• A rule for what happens to an enquiry arriving on Friday afternoon.

And the practical consequence: if nobody picks up enquiries within the day, raising the budget increases the lost opportunities, not the customers.

The quality of the enquiries

Twenty enquiries at €10 each can be worse than eight at €25. The count on its own is not a measure.

What makes an enquiry good

• The person is within the scope of the service by size and budget.
• They roughly understand what they are buying.
• They have a decision or at least a clear deadline for one.
• They can be served with current capacity.

How the page filters

The price indication filters by budget. The "who it is for" description filters by size. The description of the process filters by readiness. A page without those three lets everything through, and the time then goes into conversations that will not become projects.

This is also the argument against cutting the form mechanically: some of the filtering is cheaper done before the conversation than inside it. Which suits you better depends on capacity - with a full schedule, filtering is worth more than volume; with spare capacity, the reverse.

What to track

Record what share of enquiries become customers. That figure tells you more about the page than the conversion rate does. If enquiries rise while the share falls, that is a reason to look: possible causes are changed traffic, seasonality, different handling of enquiries, or simply a small sample. The page is one of the hypotheses, not the answer.

Track one more thing that almost nobody logs: why enquiries fall through. An enquiry from outside the service area, below the minimum budget, or for something you do not offer may point to missing information on the page, but it may equally mean the traffic is arriving on unsuitable queries. The two differ in whether the queries people arrive on match the service. Three months of such notes is enough to see which it is.

Store versus service business

In a store the funnel is visible

It can be followed step by step: product page, cart, delivery details, payment. The largest loss is usually in one of the later steps, but not always: it may be in the product page or in unsuitable traffic. The figures show which. That case is covered in the article on how an online store loses sales before the advertising, and what the page itself has to contain in the article on ecommerce product pages that sell and rank.

In a service business the funnel is shallower but still measurable

It is sometimes claimed that service businesses have no funnel. They do: visit, getting past the first screen, reaching the form, started form, submitted form, qualified enquiry, quote, contract. There are more steps and they are softer, and the reasons for dropping out are harder to read, but the same thing gets measured.

The order of work

First: test the chain

Send a real enquiry from a phone. Click the phone number. Check that the notification arrives and that the action is recorded. Surprises here are frequent and expensive, and the fix is quick.

Second: set up the measurement

Separate the actions, add the four funnel metrics from the measurement section and wait for enough data. Without this, the next steps are gambling.

Third: wherever the figures point

If it drops before the first screen, the headline and the match to the query. If it drops before the form, price, process and proof. If it drops in the form, the fields. Not in that order, but in the order the data shows.

What is left for later

Recording calls, a rule for response time, and logging the share of enquiries that become customers. If the page sits on a site that is not ranking organically either, the order of checks is different and is covered in the article on why a website is not ranking in Google. If you would like us to go through yours together, get in touch.

Frequently asked questions

What conversion rate is normal for a services website?

From practice, the range for cold traffic is between 2 and 5 per cent, closer to the lower bound for more expensive services with a long cycle. That is a reference point, not a standard: it depends on the niche and above all on what you count as a conversion. In our own account for the Bulgarian market, recorded conversions run at 1.93 per cent of clicks. Look at the rate per landing page, because brand traffic lifts the average.

Should I publish prices if every project is different?

At least an indication is worth it: a starting figure, the factors that change it, and how an exact quote is reached. The reason is practical rather than proven with numbers: the indication filters by budget before time has been spent on either side, and it saves a round of clarifications. If there is resistance internally, start with a "from" figure and the factors.

How do I tell whether the problem is the traffic or the page?

First check that the recording works: missing measurement looks exactly like missing enquiries. Then look at whether the queries people arrive on match what you sell. Suitable queries with no enquiries point towards the page, but do not prove it on their own: a small sample, a long decision cycle and unrecorded phone enquiries are all possible. The quickest check is the search terms report.