01 The problem with reported numbers
Every system counts by its own logic. An ad platform credits itself with an enquiry if someone clicked an ad beforehand. An analytics tool counts a form submission. The form itself also accepts spam and test entries.
- Spam and automated entries are counted as enquiries.
- The same person enquires twice and is counted twice.
- Several channels claim the same enquiry for themselves.
- Faulty tracking counts too much or nothing at all.
None of these errors is unusual. Together they mean that the sum of reported results often differs considerably from what actually reaches the business.
02 Four checks before a result counts
- 1
Does measurement work?
Forms, calls and purchases are checked with test runs. If an event is missing or fires twice, every further number is worthless.
- 2
Is the result genuine?
Spam, test entries and duplicate enquiries are filtered out – by clear, traceable rules.
- 3
Which action does it belong to?
The origin is assigned to an action. Where that is not clearly possible, it is shown as such.
- 4
Did it reach the business?
Matching with the customer database shows whether the enquiry became a contact, a quote or an order.
03 What this means for decisions
The key figure is not the cost per reported conversion but the cost per verified result. Only then can channels be compared fairly and budgets be shifted sensibly.
The gap itself is also a signal. If the difference between reported and verified results suddenly grows, a spam wave or a tracking error is often behind it – long before it shows in revenue.
04 The limits of attribution
Not every effect can be assigned to an action. Recommendations, conversations at a trade fair or a company vehicle on the road leave no click trail. It is better to show such results honestly as “not assigned” than to credit them to whichever channel happened to be involved last.