A better report starts with a better question.
An advertising report is a view of recorded activity. It can help explain what was delivered, what was observed after an interaction and where a campaign needs investigation.
Explore the platform reference
It cannot make incompatible definitions compatible simply by placing them on the same chart. Before choosing a visualisation, decide which operational question the report is meant to answer.
Embed Measurement Notes is an independent guide to advertising metrics and reporting decisions. It uses small invented datasets to explain denominators, cost calculations and the limits of attribution. No chart on this page is a live campaign feed. The aim is to make a report reproducible enough that a colleague can understand both the number and the decision attached to it.
Name the event before naming the metric
Start with an event that a system records. An advertisement may create an impression record, an interaction may create a click record and a completed form may create a submission record. The names sound familiar, but their exact definitions and collection conditions matter. Two systems can use similar labels while observing different parts of the journey.
Write a small metric dictionary before assembling the report. Include the event, the system that records it, the reporting period and any important exclusions. If an enquiry must be reviewed before it is considered qualified, keep the raw submission and the qualified outcome separate. Otherwise a later change in qualification can appear to be a change in advertising performance.
The dictionary does not need to become a lengthy manual. A clear sentence beside each critical measure is enough to prevent a reader from confusing a page view with a visit or an enquiry with a customer. The practical test is whether another person can reproduce the calculation without asking what the team meant by the label.
Keep the denominator visible
A rate expresses a relationship between quantities. Clickthrough rate relates recorded clicks to impressions. A conversion rate depends on the particular conversion and interaction definitions being used. A percentage without its denominator leaves the reader to guess which population the number describes.
Suppose an invented placement records 20,000 impressions and 160 clicks. The arithmetic clickthrough rate is 0.8 percent. If another placement records twenty clicks from 1,000 impressions, its rate is two percent, but its sample is much smaller. The higher percentage alone does not establish that it will remain higher under comparable delivery or produce more useful outcomes.
Show both the rate and the underlying counts. This helps the reader distinguish a stable-looking large record from a ratio that changes sharply when one event is added. It also discourages a common reporting mistake: averaging percentages from unequal groups instead of calculating a combined rate from the combined numerator and denominator.
Put the denominator to work.
These results describe the entered amounts. They are not a forecast, a suggested bid or evidence of traffic quality. With a zero denominator, the corresponding rate is undefined.

Work through the cost calculation
The calculator on this page uses spend, impressions and clicks to compute arithmetic CPM and CPC. CPM is spend divided by impressions, multiplied by one thousand. CPC is spend divided by clicks. These are descriptive calculations for the values entered, not forecasts or suggested bids.
For example, a hypothetical expenditure of 150 currency units across 50,000 impressions and 300 clicks produces a CPM of three units and a CPC of 0.50 units. The calculator does not know whether those clicks were relevant, whether the inventory was suitable or whether a later event became a customer. Keep the financial arithmetic separate from the quality assessment.
Zero denominators need an explicit result rather than a misleading zero cost. If no clicks are recorded, cost per click is undefined for that record. Negative spend and negative counts are not valid inputs for this simple tool. The interface explains these cases so that a missing or invalid observation cannot quietly become a favourable-looking performance figure.
| Invented record | Calculation | Read it as |
|---|---|---|
| 160 clicks / 20,000 impressions | 0.8% | Recorded clickthrough rate |
| 20 clicks / 1,000 impressions | 2.0% | A higher rate from a smaller group |
| 150 spend / 300 clicks | 0.50 units | Arithmetic cost per recorded click |
Add dimensions only when they change the decision
A dimension groups a record: campaign, source, device or another agreed category. Breaking a report down can reveal an issue that a total hides, but every additional grouping also creates smaller samples and more opportunities to overinterpret noise. The useful question is what action a particular breakdown supports.
A device view might reveal a destination that is difficult to use on a phone. A source view might identify a concentration of delivery that needs inspection. A creative view might show that two messages are attracting different kinds of visitors. These are starting points for investigation, not automatic explanations of causation.
Keep the total available and show how the groups relate to it. Unknown or missing categories should remain visible when they represent a meaningful part of the record. Hiding them can make the remaining chart look cleaner while making its interpretation less honest. If the grouping is unstable because identifiers changed during delivery, repair the mapping before using it to compare campaigns.
Use campaign labels that survive the handoff
Campaign parameters can help an analytics system associate a visit with a named source, medium and campaign. They are labels passed through a route, not evidence that a particular advertisement caused an outcome. A consistent naming convention makes those labels useful; an improvised convention can split one campaign into several apparently unrelated rows.
Choose a stable campaign identifier and decide how creative variants will be distinguished. Keep the spelling and case consistent with the reporting process. Avoid putting personal or sensitive information in URL parameters, which can be copied, logged or shared as part of the address. The campaign record should explain the identifier without requiring a reader to decode a private conversation.
Test the complete destination route. A redirect can drop a parameter, a link can point to an old page and a copied address can preserve an unintended value. Confirm that the intended labels reach the landing page and are recorded as expected. This is a practical implementation check, separate from deciding which attribution method the business will use.

Treat attribution as a set of rules
Attribution describes how a system assigns credit for an observed outcome across eligible interactions. Different rules and windows can produce different allocations from the same journey. A report needs to state the method and observation conditions rather than presenting the assigned credit as an unqualified account of what caused the outcome.
Imagine a visitor who sees one campaign, returns through another channel and completes a purchase later. A report focused on the final interaction may tell a different story from one that allocates credit across several recorded interactions. Neither can recover interactions the system did not observe merely by changing the chart style.
When two reports disagree, compare their event definitions, eligibility windows and observation coverage before choosing the more attractive result. A platform report and a website analytics report may answer different questions. Understanding that difference is often more useful than forcing the totals to match through an unexplained adjustment.
Keep time and revision in the report
A reporting period needs a time zone and a clear extraction time. Some outcomes arrive after the original interaction, and some records are adjusted later. A report captured immediately after a campaign ends may therefore differ from a later view of the same period. Preserve that timing instead of treating the later number as proof that the earlier export was careless.
Distinguish a genuine correction from a new observation. If duplicate submissions are removed, explain the correction and retain the definition used. If an outcome arrives later, explain the observation delay. Both change a total, but they imply different lessons for the next campaign and for the reporting process.
Use a simple version note when a report is updated. State what changed, which figures are affected and whether the operational conclusion changed with them. Replacing a file without a note can leave different colleagues making decisions from incompatible versions while believing they are discussing the same record.
Choose a chart that exposes the comparison
A chart should make the intended comparison easier to inspect. A trend needs a consistent time scale. A comparison between categories needs readable labels and an appropriate baseline. A cost table may be more useful than a decorative chart when the reader needs exact amounts and calculation definitions.
Do not use an attractive visual treatment to conceal missing data. A gap in the record should not automatically become a smooth connecting line. A category with no observed outcomes should not disappear simply because it complicates the layout. The design should preserve the distinction between zero, missing and not applicable.
The visual examples in this journal use deliberately small scenarios so the reader can reconstruct the arithmetic. In a real report, pair the chart with the definitions and counts needed to assess it. Keep colour meanings consistent and provide text labels so the comparison remains understandable without relying on colour alone.
Bring the buying platform into the measurement plan
Before a campaign runs, compare the identifiers, reporting fields and destination requirements supported by the buying setup with the measurement plan. A report cannot reliably break down a creative identifier that was never recorded or reconcile an event that the team defined differently after delivery began.
Reacheffect is the sponsored commercial reference for this publication. Review the provider's current platform information alongside the events and dimensions your own process can capture. This page does not reproduce a live account report or certify that a particular integration is available for every campaign configuration.
The practical handoff is a short agreement: these identifiers will be used, this event will be counted and this period will be reviewed. Keep the agreement with the creative and destination versions. It gives later analysis something firmer than a collection of screenshots whose definitions may have changed between exports.

Finish with an action the evidence supports
The final section of a report should explain what happens next. A low observed visit count after many recorded clicks might justify checking the destination route. A poor enquiry qualification rate might justify reviewing the message and the offer. Neither observation automatically identifies the single cause of the problem.
Separate the diagnosis from the proposed test. State what is known, what remains uncertain and which change would help distinguish the competing explanations. If several campaign settings change together, preserve that fact and keep the conclusion correspondingly broad. Precision in the language matters as much as precision in the arithmetic.
A useful report leaves a reader able to reproduce the important number, understand its limits and explain the next decision. It does not need an impressive number of panels. It needs a coherent connection between the question, the record and the action that follows from it.