Question 1

What is marketing mix modelling and why is it coming back?

Accepted Answer

**Marketing mix modelling (MMM)** is a statistical technique estimating how different marketing activities contribute to sales, using aggregate historical data rather than tracking individuals.

**How it works.** A regression model relates sales over time to spend across channels, plus external factors — seasonality, pricing, promotions, competitor activity, weather, economic conditions. The output attributes a share of sales to each input, and estimates diminishing returns and **carryover** (the lasting effect of spend after it stops).

**Crucially, it requires no user-level data at all.** It works on weekly or monthly aggregates: spend by channel, sales, and context.

**Why it is returning after decades out of fashion.** MMM was the standard approach before digital tracking. It fell out of favour because click-level attribution appeared to offer precision MMM could not. Several things reversed that:

**Tracking degraded.** Third-party cookie restrictions, Apple's App Tracking Transparency, browser protections and consent requirements have substantially reduced the coverage and reliability of user-level attribution. The precise measurement turned out to be neither.

**Privacy regulation** makes aggregate methods legally simpler.

**Attribution's blind spots became obvious.** Click-based models cannot measure television, radio, out-of-home, sponsorship or the long-term effects of brand advertising — so they systematically credited the channels they could see. MMM measures everything that could plausibly affect sales.

**Computing and open-source tools** made it accessible to organisations that could never afford a consultancy engagement.

**Its limitations are real:**

**It needs a lot of data** — typically two or three years of history.

**Correlation, not causation.** Spend often rises when sales are expected to rise anyway, which biases results unless handled carefully.

**It cannot optimise in-flight**, reporting at aggregate cadence rather than daily.

**Results are sensitive to specification**, so two analysts can produce different answers.

**The emerging consensus** is a triangulation of MMM, incrementality experiments and attribution together.

Question 2

What is a retail media network, and why is every retailer building one?

Accepted Answer

A retailer selling advertising inventory against its own shopper data — on its website, in its app, in physical stores, and increasingly on other websites entirely. It has become the fastest-growing

Question 3

How do ad blockers affect marketing measurement?

Accepted Answer

They remove a non-random slice of your audience from the data, which distorts more than the raw blocking rate suggests — and the browsers themselves now do much of the blocking, independently of

Question 4

Does demographic targeting actually work?

Accepted Answer

Less well than the industry's reliance on it suggests. Demographic categories are weak predictors of purchasing behaviour, and the evidence for this has been accumulating for years without changing

Question 5

What is the difference between first-party, second-party, third-party and zero-party data?

Accepted Answer

Who collected it and from whom — a distinction that became central as third-party tracking was restricted. First-party data. Collected by you, directly from your own customers and visitors, through

Question 6

What is marketing attribution and why do platforms all claim the same conversion?

Accepted Answer

Attribution is the process of deciding which marketing touchpoints deserve credit for a conversion. It sounds administrative and it is one of the hardest problems in the discipline. Why the numbers

Question 7

How does affiliate marketing actually work, and who gets paid for what?

Accepted Answer

A performance model where a third party — a publisher, creator or comparison site — sends traffic to a merchant and is paid only when a defined action occurs, usually a sale. The mechanics are

Related Questions