Advertising Effectiveness

An advertisement can receive a click shortly before a purchase without having caused that purchase. The central question in effectiveness research is what customers would have done without the advertising. Randall Lewis, Justin Rao, and David Reiley’s research overview distinguishes the abundance of digital reporting metrics from the harder task of measuring causal effects on consumer behavior. Its abstract supports that methodological distinction; detailed findings from the full paper are not asserted here. s1

A useful evaluation therefore begins with a defined outcome and period. Reporting exposure, visits, or transactions answers different questions from estimating additional purchases. Treating these as interchangeable can make a campaign appear successful merely because it reached people who were already likely to buy. s1s2

[s1] [s2]

Google’s media effectiveness guide distinguishes attribution, incrementality experiments, and marketing mix modelling. Attribution assigns credit among recorded interactions on a conversion path. It can support frequent campaign decisions, but assigning credit does not itself establish causation. s2

Incrementality experiments compare outcomes between groups assigned to receive advertising and groups from which it is withheld. Marketing mix modelling instead uses aggregate information to examine media investment alongside other influences on sales, including seasonality, prices, and economic conditions. The guide presents the methods as complementary and acknowledges limitations: experiments concern particular strategies and periods, while models depend on assumptions and suitable data. Google sells advertising, so this is a vendor-authored methodological guide rather than independent evidence of results in Iran. s2

An eBay field-experiment abstract provides a concrete warning about selection. Tom Blake, Chris Nosko, and Steven Tadelis reported that conventional estimates overstated paid-search returns and that responses differed between customer groups. These findings concern the studied eBay campaigns; they do not establish the returns of Iranian search advertising or imply that every brand campaign is ineffective. s3

[s2] [s3]

Reporting from Iran illustrates why changes in access belong in a campaign’s evidence record. In January 2026, Associated Press described an internet shutdown disrupting businesses that depended on online advertising and communication with customers. Interviewees reported difficulty reaching audiences on the services they normally used. These accounts document disruption, not a representative estimate of advertising returns. s4

The measurement implication is an inference: a fall in recorded visits during an outage cannot automatically be attributed to weaker advertising. Both the route to the customer and the route by which activity is measured may have changed. An evaluation should establish which services were accessible before comparing periods. s4

[s4]

A credible effectiveness claim identifies the outcome, explains the comparison, and states what the method cannot resolve. Attribution reports, aggregate models, and experiments should retain their different meanings when summarized for readers. Evidence from another country can explain a method without supplying an Iranian benchmark. s1s2s3

No inspected source establishes a representative annual Iranian effectiveness series. The defensible conclusion is methodological: more detailed reporting can improve description, but a claim of additional impact still requires a justified causal comparison. s1s2

[s1] [s2] [s3]