Ad measurement and attribution is the software that an advertiser uses after a campaign has run, to find out what the money actually achieved. It answers three questions. Was the ad seen by a real person, in a suitable place? Which ad led to the sale? And would the sale have happened even without the ad?
Consider a mattress brand that spends ₹50 lakh on online ads in the month before Diwali. Its sales go up by a third. The marketing head now has to explain the result to the company's owners. Some of the ads may never have appeared on anyone's screen. Some may have been shown to bots, programs that pretend to be people, which is a form of ad fraud. And some buyers would have bought a mattress during the festival sale anyway.
The first question is answered by ad verification. It checks that each ad appeared on the screen, that a person rather than a bot saw it, and that it did not run beside content that harms the brand.
The second question is called attribution. It follows each buyer from the ads they saw to what they bought, and gives credit to each ad. For apps, a mobile measurement partner (MMP) does this job. It records which ad brought each new user to the app.
The third question asks whether the spending caused extra sales. This extra is called incrementality. Marketing mix modelling (MMM) estimates it with statistics on months of sales and spending. It can also compare cities where the ads ran with cities where they did not.
The third question has become more important. Attribution depends on following one person from an ad to a purchase. Data protection laws, such as India's DPDP Act, and phone makers now block much of that tracking. Marketing mix modelling looks only at totals, so it follows no one and still works.
Around all three sits marketing reporting and analytics. It collects each channel's results into one set of reports, which often end up in the company's business intelligence dashboards.
Advertisers pay these tools a monthly subscription, or a fee for every thousand ads checked. The tools must be independent of the companies that sell the ads. Otherwise the company that sold the ads would also be the one that judges whether they worked. This matters most for walled gardens, the large platforms that sell ads on their own sites and measure them with their own tools. The ads themselves are bought on the buy side and sold on the sell side of programmatic advertising.