Analytics that helps you make a decision
We help connect marketing to real business events. We define what counts as an enquiry, a suitable request, a booking or an order, and check available sources. A report should explain what to do next, not just show more charts.
What the business gains
Indicators with a clear meaning.
Every figure is linked to an event definition and source.
Distinct stages.
A view, enquiry, booking, actual service and payment do not become one total.
Data quality checks.
Duplicates, incomplete events and consent limitations are visible in the report.
A basis for decisions.
The team can discuss what to improve on the website and in its work.
A click, an enquiry and a sale are different events
A visit may not result in an enquiry. An enquiry submitted through a form may be a duplicate or may not match the offer. A confirmed booking can be cancelled, an order can remain unpaid and a payment can be refunded. Calling everything one “conversion” gives budget decisions an incomplete picture.
We begin with definitions and who confirms each event. A figure in a report needs a clear source, period and meaning.
Follow real events before deciding what to improve
Visit
Enquiry
Confirmation
Fulfilment
Payment
From the enquiry: duplicate
Link to the original event before counting.
From confirmation or payment: cancellation or refund
A separate outcome, not completed revenue.
From the visit: unknown source or missing link
State the limit before crediting a channel.
Illustrative sequence, without invented figures: visit → enquiry → confirmation → fulfilment → payment; cancellation is a separate outcome. A missing source or duplicate limits the conclusion and needs checking before a channel receives credit.
What pNet does
Agrees useful indicators
We choose the goal and events along the customer journey. We define counting rules, exceptions and the person responsible for data quality. Different business models need different measures: a visit, programme participation and an accepted B2B order cannot be measured identically.
Checks sources and connections
We examine data from the website, advertising platform, CRM and selected financial system. We define stable identifiers and a permitted matching method. A matching time or email does not always prove that events belong to the same person or order.
Sets up necessary measurement
We prepare selected events and data transfers within the agreed scope. We account for access, permitted processing and actual consent settings. We do not send unnecessary personal details just because an advertising interface has a field for them.
Checks quality and limits
We look for duplicates, gaps, delays, different time zones and inconsistent definitions. We distinguish an event that did not occur from missing data about it. We test the selected journey before treating an attractive dashboard as proof of correctness.
Prepares a clear view for a decision
We show results alongside limitations and changes. We distinguish supported conclusions, working hypotheses and places where data is insufficient. A recommendation should follow an observation, not a desire to increase the budget at all costs.
Examples of useful distinctions
Salon: an enquiry, confirmed booking, completed visit and payment are counted separately. A cancellation or refund should not remain an unexplained successful sale.
Academy: interest in a programme differs from an assigned place, payment and actual participation. A corporate order for several places is not the same as several independent enquiries.
Supplier: a qualified enquiry, accepted proposal version, order and delivery have different dates. A long sales cycle is accounted for explicitly, not hidden in one monthly figure.
What one report cannot establish
Platforms link sales to advertising differently: they count different periods after an impression or click and allocate credit to channels in different ways. Their figures therefore cannot simply be added together and called unique sales. Revenue is not profit, and growth observed after a change does not prove that the change alone caused it.
When a sound experiment or comparable check is possible, we discuss it separately. With limited data, an honest statement of uncertainty is more useful than a number that only looks precise.
The result of pNet's work
The team receives indicator definitions, a source map, selected measurement and a report suited to a specific decision. It is clear what has been checked and what still needs improvement. After launch, we help monitor data quality and update reporting as processes change.
Illustrative report: an observation must support a decision
Example structure without invented metrics. Use only available, permitted sources; these rows do not claim an installed measurement system.
One decision from a checked observation
| Event | Source to verify | Limit or exception | Team decision |
|---|---|---|---|
| EventEnquiry | Source to verifyWebsite receipt and CRM owner | Limit or exceptionDuplicate, failed delivery or unknown source | Team decisionRepair the enquiry path before increasing spend |
Full illustrative report: sources, limits and decisions
| Event | Source to verify | Limit or exception | Team decision |
|---|---|---|---|
| EventEnquiry | Source to verifyWebsite receipt and CRM owner | Limit or exceptionDuplicate, failed delivery or unknown source | Team decisionRepair the enquiry path before increasing spend |
| EventConfirmed booking or order | Source to verifyConfirmation in the responsible business system | Limit or exceptionCancellation, changed terms or no confirmation | Team decisionSeparate accepted work from unconfirmed interest |
| EventPayment and actual fulfilment | Source to verifyFinancial record and delivery or visit confirmation | Limit or exceptionRefund, missing link or different reporting period | Team decisionState the data limit before crediting revenue to a channel |
After the website launches, analytics helps decide which page to improve, which advertising test to stop and what prevents the team from handling enquiries. Measurement takes account of applicable consent settings for data processing and the availability of sources. No report turns a growth hypothesis into a proven effect without a valid comparison.
Important: A funnel of advertising events without a connection to orders and payment does not prove revenue or ROI. Cancellations and refunds are counted separately.
Frequently asked questions
An interface can be prepared, but definitions and verified data are needed first. Otherwise it will display wrong conclusions quickly and clearly.
It is one source with its own rules. For a business decision, it is useful to compare it with confirmed CRM and financial events within the permitted scope.
Only if agreed costs, results and a suitable way to connect them are available. With incomplete data, we use a limited indicator and state clearly what it means.
No. We start with necessary events and the minimum data. Processing, access and transfers are agreed separately.
Define the decision that needs data
Tell us what you want to decide: choose a channel, improve a page, assess enrolment or understand enquiry quality. We will identify the minimum useful measurement.