Advertising Incrementality: How to Measure Whether Campaigns Generate Additional Sales
Advertising reports can show impressive conversion figures, growing revenue and a strong return on advertising spend. However, these results do not necessarily mean that every recorded purchase was caused by advertising. Some customers were already planning to buy, while others might have reached the same business through organic search, direct visits or recommendations. Advertising incrementality helps marketers identify the additional sales that would not have occurred without a campaign. This approach provides a more realistic understanding of marketing effectiveness and helps businesses decide which advertising investments contribute to actual growth. Understanding how incrementality works is particularly important in 2026, when customers frequently interact with several marketing channels before completing a purchase.
What Is Advertising Incrementality and Why Is It Important?
Advertising incrementality measures the additional business results directly caused by advertising activity. Instead of counting every purchase connected to an advertisement, it estimates how many sales would have happened if the advertising had not been shown. This difference separates genuine advertising impact from existing customer demand. For example, a regular customer might search for a familiar retailer, click a sponsored result and complete a purchase. The advertising account records a conversion, but the customer may have purchased the same product without seeing that advertisement. Incrementality measurement helps businesses understand whether advertising influenced purchasing behaviour or simply appeared during a decision that had already been made.
The main principle is to compare actual campaign performance with a credible estimate of what would have happened without the advertising. Imagine an online retailer recording 1,500 purchases during a promotional period. A properly designed experiment indicates that approximately 1,200 purchases would have occurred without the campaign. The estimated incremental contribution is therefore 300 additional orders. These figures provide more useful information than the total number of recorded transactions because they show the potential increase in demand caused by advertising. The business can then compare the value of those additional orders with campaign expenditure to understand whether the investment delivered a worthwhile financial return.
Incrementality becomes particularly valuable when businesses use several advertising channels, operate remarketing campaigns or spend significant amounts on branded search. Customers often encounter multiple advertisements before completing a purchase, and different reporting systems may assign credit to the same transaction. A shopper might see a social media advertisement, receive a promotional email and subsequently click a paid search result. Although each interaction can appear important in marketing reports, the customer might have purchased the product regardless. Incrementality testing helps marketers avoid overestimating campaign performance and provides stronger evidence for decisions about advertising budgets, customer acquisition and long-term marketing strategy.
How Incrementality Differs from Traditional Marketing Attribution
Marketing attribution and incrementality measurement answer different questions about advertising effectiveness. Attribution identifies which marketing interactions receive credit for a completed conversion. Depending on the selected model, that credit may be assigned to the final click or distributed across several interactions. This information helps businesses understand customer journeys and identify the channels commonly involved before a purchase. However, attribution cannot establish whether those interactions actually caused the transaction. Incrementality focuses on the additional outcome created by advertising. Its central question is whether the customer would still have purchased if the advertisement had never appeared.
Consider a retailer that spends £6,000 on remarketing advertisements and receives £30,000 in attributed sales revenue. According to the advertising account, the campaign generates five pounds in revenue for every pound spent. However, an incrementality experiment might show that most customers were already likely to purchase. If the experiment estimates only £9,000 in additional revenue, the incremental return on advertising spend is 1.5 rather than 5. This difference changes the financial interpretation of the campaign. Although advertising generated additional revenue, the amount may not be sufficient to cover advertising costs once product margins, fulfilment expenses and other operating costs are considered.
Both approaches remain useful when applied correctly. Attribution supports everyday campaign monitoring, helps identify customer interactions and provides information for advertising optimisation. Incrementality testing offers stronger evidence when deciding whether a campaign deserves additional investment. A business can therefore use attribution reports to understand where purchases are recorded while relying on controlled experiments to estimate how many purchases advertising actually creates. In 2026, this combination is especially relevant because customer journeys increasingly involve several devices, advertising channels and interactions that cannot always be connected to a single identifiable individual.
How to Measure Incremental Sales from Advertising Campaigns
One of the most reliable methods for measuring advertising incrementality is a controlled experiment involving comparable customer groups. The first group is eligible to receive the advertising campaign, while the second group is deliberately excluded from it. The excluded customers form a control group, providing an estimate of purchasing behaviour without the tested advertising activity. Ideally, participants are assigned randomly so that differences in purchasing intentions, customer characteristics and previous interactions are distributed between the groups. If the advertising group produces more purchases than the control group, the difference can indicate an incremental advertising effect, provided the experiment has been conducted correctly.
Before launching an experiment, businesses need to establish precisely what they want to measure. Additional completed purchases, revenue, new customers and repeat orders can all be relevant outcomes, but each answers a different commercial question. The marketing team should also decide which audience will participate, how long the experiment will run and how results will be recorded. Both groups should experience similar pricing, product availability and general business conditions. Major promotional changes during the experiment can make the results harder to interpret. The testing period must also reflect the normal purchasing cycle, since customers buying expensive products may take considerably longer to convert than shoppers purchasing everyday items.
There are several practical approaches to incrementality measurement. Customer-level holdout tests are suitable when advertisers can reliably separate users into experimental groups. Geographic experiments compare results across selected regions where advertising activity differs, making them useful for campaigns that cannot easily be controlled at an individual level. Businesses can also use marketing mix modelling to examine how advertising expenditure, seasonal demand, pricing and other factors relate to sales over time. Each method has limitations. Randomised experiments generally provide stronger evidence of causality, while geographic testing and statistical modelling require particular care to distinguish advertising effects from differences in market conditions.
Calculating Incremental Lift, Additional Revenue and Campaign Returns
Incremental lift expresses the improvement in business results caused by advertising compared with the expected outcome without it. Consider an experiment involving two randomly assigned groups of 10,000 customers. The advertising group records 660 purchases, while the control group records 550. Their conversion rates are 6.6% and 5.5%, respectively. The absolute improvement is 1.1 percentage points. Relative incremental lift is calculated by subtracting 550 from 660, dividing the result by 550 and multiplying by 100. This produces an estimated lift of 20%. In practical terms, the campaign generated an estimated 110 additional purchases among the 10,000 customers assigned to the advertising group. These figures are illustrative, and a real experiment must also assess statistical uncertainty.
Once the additional purchases have been estimated, marketers can calculate incremental revenue. If the 110 additional orders have an average value of £90, the campaign produces approximately £9,900 in incremental revenue. Suppose advertising expenditure totals £4,500. Dividing incremental revenue by advertising expenditure gives an incremental return on advertising spend of 2.2. This means that every pound invested generated an estimated £2.20 in additional sales revenue. However, revenue is not the same as profit. If the business retains a 35% contribution margin before advertising costs, the additional orders contribute £3,465 towards covering campaign expenditure. Under these simplified assumptions, the advertising investment would not recover its full cost through immediate sales alone.
Incrementality calculations should always be interpreted alongside the reliability of the experiment. A reported lift of 20% does not guarantee that future campaigns will produce the same result. Customer demand can change because of seasonality, competitor activity, pricing and product availability. Statistical confidence intervals help indicate the range of advertising effects compatible with the observed results. If that range includes zero, the evidence may be insufficient to conclude that the campaign increased sales, even when the estimated lift is positive. Such an outcome does not necessarily prove advertising was ineffective. The experiment may have involved too few customers, an insufficient number of purchases or excessive variation to detect the true effect accurately.
How to Use Advertising Incrementality to Improve Marketing Performance
Incrementality measurement becomes commercially valuable when the findings influence advertising decisions. A business may discover that campaigns targeting unfamiliar customers generate considerably more additional purchases than remarketing advertisements aimed at recent website visitors. Another advertiser might find that branded search campaigns report excellent conversion figures but contribute relatively little additional revenue because existing customers would have reached the website through unpaid search results. These findings can help marketers reconsider audience targeting, bidding strategies and budget allocation. However, a weak result from one experiment should not automatically lead to cancelling an entire advertising channel, since performance can vary between audiences, products and campaign objectives.
Businesses can also use incrementality findings to identify the point at which additional advertising expenditure becomes less productive. Increasing a campaign budget does not always produce a proportional increase in sales. Initial investment may reach customers with strong purchasing potential, while subsequent expenditure reaches people who are less likely to respond or who have already encountered several advertisements. Repeated experiments at different spending levels can help identify these diminishing returns. For example, a retailer may find that increasing weekly advertising expenditure from £5,000 to £7,500 generates worthwhile additional revenue, while a further increase to £10,000 produces only a small improvement. Such evidence supports more disciplined spending decisions.
Incrementality should form part of a broader marketing measurement approach rather than replace every existing metric. Attribution reports remain useful for monitoring campaign interactions, while customer research helps explain purchasing motivations. Marketing mix modelling can provide a wider perspective on how advertising expenditure and other commercial factors relate to sales. Controlled experiments add evidence about causal effects. When these methods produce different results, marketers should examine the assumptions behind each measurement instead of selecting whichever figure appears most favourable. A channel that performs strongly in attribution reports but produces limited incremental sales may be capturing existing demand rather than creating substantial additional demand.
Common Incrementality Testing Mistakes and How to Avoid Them
A common mistake is assuming that any increase in sales following an advertising campaign was caused by that campaign. Revenue may rise because of seasonal demand, promotional discounts, changes in competitor pricing or improvements in product availability. Comparing sales before and after advertising begins does not automatically separate these influences. Another frequent problem is comparing customers who saw advertisements with those who did not without considering differences between the groups. People who frequently visit a retailer’s website may be more likely both to encounter advertisements and to purchase products. Random assignment helps reduce this selection bias and creates a more credible basis for estimating the effects of advertising.
Insufficient sample sizes and poorly chosen testing periods can also undermine results. A campaign generating only a small number of conversions may not provide enough information to distinguish genuine advertising effects from ordinary variation. Before beginning an experiment, marketers should estimate the number of participants and conversions required to detect an improvement that would matter commercially. They should also account for purchasing delays, especially when selling products that require extensive research. Changing targeting settings, prices or major promotional offers during the study can complicate interpretation. Similarly, if customers assigned to the control group still receive the tested advertisements through another route, the measured difference between groups may become smaller and less informative.
Finally, businesses should avoid treating a single experiment as a permanent assessment of campaign effectiveness. Advertising results can change as customer awareness grows, competitors adjust their strategies and market conditions evolve. A campaign may create substantial additional demand during a product launch but deliver a smaller effect once the product becomes familiar. Short-term experiments may also miss repeat purchases or longer-term customer relationships, although these benefits should be measured rather than assumed. A reliable incrementality programme involves repeating important tests when conditions change, applying consistent measurement definitions and recording the limitations of each result. The purpose is to establish which advertising investments create genuinely additional sales, which mainly capture existing demand and how marketing expenditure can contribute more effectively to sustainable business growth.