MKTG3600 · Marketing in Practice
Execution: Data & Technology
Week 10 covers data and technology across the marketing mix, delivered by a digital-first brand. It defines data and tech, shows how each touches all four Ps, explains programmatic buying as a four-step flow, and introduces experimentation and incrementality (control vs exposed) as the way to prove marketing actually caused an outcome. The recurring line: "fundamentals don't change, the tools do." In the exam expect "explain how data and tech change the marketing mix" or "what is incrementality and why does it matter"; it supports the measurement and execution logic of the live brief.
What this chapter covers
- 01Data = capturable information used to understand and improve marketing effects; Tech = the tools and platforms that enable marketing at scale
- 02John Wanamaker (~1919) — "half the money I spend on advertising is wasted; I don't know which half" — data now shows which half
- 03Data & tech across the 4Ps — dynamic pricing (Price), data-led distribution (Place), personalisation (Product), automated right-message-right-moment (Promotion)
- 04Programmatic media (4 steps) — User → Signal → Match → Ad Served
- 05Incrementality / experimentation — Control (no ads) vs Exposed (shown ads); uplift over control = true ROI
- 06"Were you going to buy anyway?" — the core incrementality question; lookalike audiences and bidding more for high-value users
- 07Frameworks the team relies on — CEPs, the marketing funnel, experimentation and continuous learning, Media Mix Modelling
- 08Emerging trends — AI, LLMs as channels, less PII, consent/regulation, retail media; fundamentals stay, tools change
Short-answer: explain incrementality and control-vs-exposed testing (15 marks)
- +3LAYER 1 — Define. Incrementality is the extra outcome that marketing CAUSED — sales that would not have happened otherwise. State the core question: "were they going to buy anyway?" Distinguish it from attribution, which can credit sales that would have occurred regardless.
- +3LAYER 2 — Apply the experiment. Describe a control-vs-exposed design: a control group sees no ads, an exposed group does; the difference in conversion (uplift of exposed over control) is the true, causal effect. This is the clearest way to measure causality, and scaling it across channels leads to Media/Market Mix Modelling.
- +3LAYER 2 (cont.) — Connect to data & tech. Show how data (captured signals) and tech (platforms) make the test possible — audiences can be split, exposure controlled, outcomes tracked; lookalike audiences let you bid more for likely high-value users. Data shows "which half" of Wanamaker's spend worked.
- +3LAYER 3 — Application & examples. Work it concretely: a launch campaign run as a geo split (some regions exposed, a matched region held out); the sales gap over the holdout is the incremental effect, not the raw sales in exposed regions. Show why raw attributed sales overstate impact.
- +3LAYER 4 — Critique. Weigh the limits: clean control groups are hard to hold in practice; experiments need scale, time and discipline; privacy rules and less PII make audience splitting harder; results can be channel- and context-specific. Incrementality is the gold standard for causality but is costly and not always feasible — so it complements, not replaces, other evidence.
Key terms
- Data
- Capturable information used to understand marketing effects and improve them — customer records loaded into a platform for targeting (input) and the clicks, sign-ups and reach that result (output).
- Technology (marketing tech)
- The tools, systems and platforms that enable marketing and advertising at scale — front-end (what customers see: apps, search, social) and back-end (ad serving, data pipelines, dashboards).
- Programmatic media
- Automated ad buying in four steps: User (lands on a site) → Signal (data captured) → Match (a bid decision plus ad copy) → Ad Served (the right ad at the right moment).
- Incrementality
- The extra outcome that marketing actually caused — sales that would not have happened otherwise. It answers "were they going to buy anyway?" and is distinct from attribution.
- Control vs exposed
- An experimental design: a control group sees no ads, an exposed group does, and the uplift of exposed over control measures the true, causal ROI. Scaling across channels leads to Media Mix Modelling.
- Wanamaker's problem
- The old lament that half of ad spend is wasted but you don't know which half. Data now reveals which half works, and technology provides the pipes to act on it at speed and scale.
Execution: Data & Technology FAQ
What is the difference between incrementality and attribution?
Attribution assigns credit for a conversion to the touchpoints that preceded it, but it can reward marketing for sales that would have happened anyway. Incrementality asks the harder, causal question — "were they going to buy anyway?" — and answers it with a control-vs-exposed experiment. In the exam, showing you understand that attributed sales overstate true impact is the key application point.
How does a control-vs-exposed experiment work?
You split the audience: a control group is deliberately shown no ads while a matched exposed group is; the difference in outcomes (the uplift of exposed over control) is the effect marketing actually caused. It is described as the clearest way to measure causality and incrementality. Scaling the same logic across many channels over time is what Media/Market Mix Modelling does statistically.
How do data and technology change the 4Ps?
Price moves from set-and-hope to dynamic and elasticity-modelled; Place goes where the data shows the audience is (e.g. new commerce surfaces); Product gets personalised per user; and Promotion becomes automated right-message-right-person-right-moment delivery through programmatic and dynamic creative. The framing is that the fundamentals of the mix don't change — the tools that execute them do.
Can Sia help me with data and tech topics in MKTG3600?
Yes — Sia can explain incrementality, control-vs-exposed testing, programmatic's four steps and how data and tech touch each P, and check that your answer separates causal from attributed effects. It teaches the method and checks your reasoning; it does not complete your graded live brief or exam, and University of Sydney academic-integrity rules apply. Confirm set materials on Canvas.
Exam move
Anchor this chapter on one big idea — incrementality — and be able to define it, contrast it with attribution, and describe a control-vs-exposed experiment cold, because that is the most exam-able concept in the week. Learn the two supporting structures as ordered lists: programmatic's four steps (User → Signal → Match → Ad Served) and how data/tech changes each of the 4Ps. Use Wanamaker's line to frame why measurement and data matter. Keep the critique ready — clean controls are hard, experiments need scale and time, and privacy/less-PII constrains targeting — and remember the recurring thesis that fundamentals stay while tools change. This chapter connects tightly to Week 11 measurement, so study them together. Confirm exam format on Canvas.
Working through Execution: Data & Technology in MKTG3600? Sia is AskSia’s AI Marketing tutor — ask any MKTG3600 Execution: Data & Technology question and get a clear, step-by-step explanation grounded in how MKTG3600 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
