Define a new customer before examining the results
A new customer should correspond to the first purchase recognised under the business's rule within the available history. Decide whether the unit is a person, a company or another commercial relationship. Keep the definition unchanged between periods so that a classification change does not look like growth in customer acquisition.
In a business-to-business service, two contacts at the same organisation may represent a single customer. In a consumer shop, the same person may buy using different email addresses. If you automatically count every record as someone new, the reported base grows without the business acquiring that many customers.
You also need to decide how to handle cancelled purchases and contracts that never started. One possible rule is to recognise the first commercially confirmed purchase and keep cancellations identified. Explain the choice rather than switching between initiated orders, signed contracts and received payments according to whichever outcome looks more convenient.
A former customer who returns still has a previous relationship with the company. You can create a reactivation group when the distinction is useful, but do not turn that person into a new customer by erasing the history. The report should help explain that recovery of demand.
To prepare your next step, explore AI data analysis for businesses.
Use commercial history as the basis for classification
Classification should begin with sales records or the commercial system that stores the customer relationship. The advertising platform adds attribution information but does not replace that history. Connect orders and customers using consistent identifiers, check duplicates and retain a category for cases that have not yet been resolved with confidence.
Start with a simple set of fields: sale identifier, customer identifier, first known purchase date, current sale date and relevant value. Add the known source in a separate field. This allows you to correct a classification without rewriting the commercial history supporting the report.
If you migrated systems and have only recent history, state that limitation. The absence of purchases in the imported file does not prove that the person never bought before. Marking the case as unknown is more rigorous than presenting the entire migrated customer base as recent acquisition.
Review the highest-impact cases first, such as customers with multiple orders or high-value sales. Avoid merging records merely because the names look similar. An incorrect merge can conceal two real customers, while an unresolved duplicate can invent an additional one in the report.
Show the composition before summarising the return
The report should present customers and first-purchase revenue separately from purchases made by existing customers. Keep the commercial total and unknown cases visible too, so that the components can be reconciled. Only then add advertising spending and the relationships you want to examine between costs and the outcomes observed.
Consider a hypothetical example with ten thousand euros of revenue during the period. Six thousand come from first purchases, three thousand from existing customers and one thousand remain uncertain. Presenting all ten thousand as new-customer revenue would conceal a material difference, even if the overall sales total were correct.
If advertising spending was two thousand euros, dividing total revenue by advertising gives five. Dividing classified first-purchase revenue by the same spending gives three. These are different descriptive ratios; neither demonstrates, simply by existing, that every euro in its numerator was caused by advertising.
The format can be short as long as it preserves these definitions. A campaign report that supports operational follow-up should make it possible to understand what changed in revenue composition and which decision that change supports, without hiding the share that still requires investigation.
Distinguish acquisition from attribution and incrementality
Being a new customer describes the previous relationship with the company; attribution describes the rule connecting a sale to marketing interactions. Incrementality asks what additional outcome happened because of the activity. These are three different questions, and they need to remain distinct when you interpret the figures in a report.
A new customer may arrive through a recommendation, an unprompted search or several interactions that include advertising. An existing customer may buy again after seeing a relevant advert. Do not automatically give advertising exclusive credit for every sale in the first group or assume every sale in the second would have happened anyway.
When campaigns specifically target acquisition, monitor first-purchase revenue alongside available source information. State the limits of the association, especially when several channels participated. Assessing causality more rigorously requires an appropriate comparison or an experiment designed to answer that question, rather than simply relabelling the sales total.
This also prevents premature decisions about campaigns with modest direct returns. Before switching them off, seek to understand which sales they are helping, keeping observed evidence, interpretation and relationships that remain hypotheses separate from one another.
Follow customer development after the first purchase
The first purchase reveals only part of acquisition value, particularly when the customer relationship continues over time. Follow groups by entry date and observe subsequent purchases, contribution and retention. This requires comparable observation periods, so that recently acquired customers are not compared with others who have had longer to buy.
Imagine two groups acquired in different months. The older group has already had an opportunity to renew a service, while the recent group is still finishing its first delivery. Higher accumulated revenue in the first group is insufficient to demonstrate better acquisition; observation time is part of the comparison.
Nor should you finance expensive acquisition using repeat purchases that you merely hope will happen. Use history compatible with the offer and keep realised revenue separate from future scenarios. If continuity has not yet been demonstrated, the report should support a more limited decision about further spending.
Connect subsequent revenue with delivery and account management costs. A second purchase may add economic value or consume nearly the entire amount in additional work. The objective is to understand the quality of the acquired relationship, rather than adding all purchases together and calling the result profit.
Turn the report into a decision about growth
Separating new and existing customers should help determine where to invest and what to improve. Observe whether growth comes from acquiring customers, increasing repeat purchases or reactivating older relationships. Compare that composition with margin and capacity, and record which change you intend to test before increasing the available budget.
If revenue grows mainly within the existing base, that can indicate a healthy commercial relationship and an offer relevant to people who already know the company. At the same time, it may reveal that new-customer acquisition needs attention. Both conclusions can coexist without diminishing the retention team's contribution.
If first purchases rise while continuity weakens, review the expectations created and the delivery experience. Acquisition reporting becomes more useful when connected to the indicators of business progress. The team can then avoid optimising for an initial transaction that creates a relationship difficult to maintain.
End each review with a verifiable action and a date for the next assessment. That might mean resolving unknown customers, monitoring a recent cohort or testing a message for a specific segment. Classification improves management when it makes the next investment decision more precise and better informed.