Data-driven marketing means choosing budgets, creative, and offers based on evidence — not only habit or the loudest opinion in the room. Tools help; judgment still decides what is a leading indicator vs noise.
Learning Objectives
By the end of this chapter, you will be able to:
- Apply a five-step framework for data-driven marketing decisions.
- Compare attribution models and their effect on channel credit.
- Design a one-page weekly marketing report.
- Build habits that support a data-informed marketing culture.
Decision Framework
- Define the outcome: Revenue, qualified leads, enrolled students, booked rooms.
- Instrument: GA4 key events, UTMs, ads conversion tracking (and server-side / CAPI where relevant).
- Read the story: Which channels and landing pages create outcomes efficiently?
- Act: Pause waste, scale winners, fix broken pages.
- Re-measure: Confirm the change moved the metric.
Attribution (Keep It Practical)
Attribution assigns credit for conversions across touchpoints.
- Last-click: Simple; undervalues awareness channels.
- First-click: Credits discovery; undervalues closer channels.
- Data-driven / position-based models: More balanced when volume allows.
For Nepal SMEs with short journeys (ad → WhatsApp → sale), last-click may be “good enough,” but still note that brand search and organic often assist paid.
Worked example: one sale, four models
A customer books a Rs. 6,000 photography package after four touchpoints: Instagram ad → Google search for the brand → email newsletter → direct visit to book.
| Model | Google search | Direct | ||
|---|---|---|---|---|
| Last-click | 0 | 0 | 0 | 6,000 |
| First-click | 6,000 | 0 | 0 | 0 |
| Linear (equal shares) | 1,500 | 1,500 | 1,500 | 1,500 |
| Position-based (40% / 20% / 40%) | 2,400 | 600 | 600 | 2,400 |
Under last-click, Instagram looks useless; under first-click, it gets all the credit. Neither is “true”; each model answers a different question. Teams should know which model their reports use before cutting a channel.
Reporting That Managers Use
Weekly one-pager beats unread 40-page decks:
- Spend, key events, CPA/ROAS by channel
- Top landing pages and drop-offs
- One insight + one next test
Example weekly one-pager
| Section | Content |
|---|---|
| Headline | Enquiries up 12% week on week; cost per enquiry down from Rs. 540 to Rs. 480 |
| By channel | Spend, enquiries, cost per enquiry for search, social, and email |
| What changed | New landing page went live on Tuesday |
| Insight | Mobile users convert 30% better on the new page |
| Next test | Add WhatsApp button to the top of the page |
Culture Habits
- Agree definitions (what counts as a “lead”).
- Prefer trends over single-day spikes.
- Separate “interesting” metrics from decision metrics.
- Protect privacy: collect only what you will use.
Nepal practice: Align marketing and sales on lead quality (e.g. WhatsApp chats that share budget/location) so analytics optimizes for real customers, not form spam during festival ads.
Case study: Pausing a “cheap” campaign
CPCs looked great on a broad Match campaign, but sales said leads were students seeking jobs. Marketing added job negatives and required a “budget range” WhatsApp qualifier. CPA rose; close rate rose more. Data + sales feedback beat the CPC vanity metric.
Summary
Data-driven marketing chooses budgets, creative, and offers based on evidence. A practical framework is to define the outcome, instrument it, read the story, act, and re-measure. Attribution models assign credit differently: last-click favours closing channels, first-click favours discovery, and linear or position-based models share credit. Managers need short, regular reports with one insight and one next test. A data-informed culture agrees definitions, prefers trends, respects privacy, and includes sales feedback on lead quality.
Key Takeaways
- Outcome → instrument → read → act → re-measure.
- Attribution models change which channels look valuable; know which one you use.
- A one-page weekly report with one insight and one test beats long decks.
- Agree what counts as a lead with sales before optimizing.
Discussion Questions
- How might a manager wrongly cut a useful channel because of last-click attribution?
- Why did the cheap-CPC campaign turn out to be expensive?
- What data should a small Nepali business avoid collecting, and why?
Exam-Style Questions
Short answer
- List the five steps of the data-driven decision framework.
- What is attribution in digital marketing?
- Distinguish between first-click and last-click attribution.
- What should a weekly marketing report contain?
Long answer
- Explain attribution models in digital marketing. A Rs. 10,000 sale followed the path Facebook ad → organic search → email → direct. Allocate credit under last-click, first-click, linear, and position-based (40/20/40) models.
- Discuss how businesses can build a data-driven marketing culture, with reference to reporting, definitions, and privacy.
