Research insights
Evaluating Marketing Effectiveness with Big Data: Methods and Pitfalls
Evaluating marketing effectiveness with big data means using large volumes of customer, media, web and sales data to measure what a marketing activity actually changed: awareness, engagement, purchases and repeat buying. Big data shows what people did at scale, but it rarely explains why, and it often misses offline behaviour and competitors. The most reliable evaluations combine behavioural data with consumer research and a clear test design.

On this page
What is big data in marketing?
Big data refers to data sets that are too large, fast-moving or varied to handle with simple spreadsheets. It is often described with the “5 Vs”: volume, velocity, variety, veracity and value. Typical marketing sources include:
- First-party data: CRM records, loyalty programmes, website and app behaviour, e-commerce orders.
- Media and platform data: impressions, clicks and conversions from advertising platforms.
- Social data: posts, comments and reviews collected through social listening.
- Sales and market data: sell-out data, retail measurement and consumer panels.
Which KPIs should you use to evaluate marketing?
| Goal | KPI | Typical data source |
|---|---|---|
| Brand building | Awareness, consideration, brand image | Brand health tracking survey |
| Engagement | Visits, time spent, interactions, content views | Web and platform analytics |
| Conversion | Conversion rate, cost per acquisition | Ad platforms, e-commerce data |
| Sales | Sales uplift, penetration, purchase frequency | Sales data, shopper panel |
| Loyalty | Repeat purchase, retention, share of category spend | CRM, shopper panel |
Choose KPIs from the campaign objective, not from what is easiest to collect. A brand-building campaign judged only on clicks will look weak even if it worked.
How do you evaluate a marketing strategy with big data?
- Set the objective and KPIs before launch. Agree what success means and how it will be measured.
- Collect and clean the data. Remove duplicates and bot traffic, align definitions across sources and check for gaps.
- Build a comparison. Use a control group, test and control regions, or before-and-after periods so that the campaign effect can be separated from seasonality and other activity.
- Analyse. Common techniques include regression to estimate the contribution of each driver, cluster analysis to find customer segments, and attribution or marketing mix models to compare channels.
- Explain the numbers. Use surveys and in-depth interviews to understand why customers responded or not.
- Adjust and re-measure. Change budget, targeting or creative, then track the same KPIs again.
Why combine big data with consumer research?
Behavioural data has blind spots. Platform data usually covers only that platform; CRM data covers only your own customers; online data misses purchases in traditional trade, which remains important in Vietnam. Big data also cannot tell you what people think of the brand or why they chose a competitor.
Consumer research fills these gaps. Surveys measure awareness, image and intent across the whole target market, including non-customers. Qualitative research explains motivations. Shopper panel data such as i-Panel, which records shoppers’ purchases from receipts and barcode scans across channels, shows category purchases, brand switching and competitor activity that your own data cannot see.
What are the main challenges?
- Data quality: incomplete, duplicated or inconsistent data leads to wrong conclusions.
- Integration: sources use different IDs, time periods and definitions.
- Attribution bias: last-click models give too much credit to the final touchpoint.
- Skills: analysts need both statistical skills and marketing understanding.
- Privacy: collecting and combining personal data must follow applicable data protection rules.
What are common mistakes?
- Confusing correlation with effect. Sales may rise during a campaign for other reasons; use a comparison group.
- Trusting platform self-reports alone. Each platform measures its own contribution; independent measurement is needed to compare channels.
- Measuring only short-term sales. Brand effects build over time and need tracking.
- Ignoring the “why”. Without research into motivations, it is hard to know what to change next.
FAQ
How do you measure marketing effectiveness?
Set KPIs from the campaign objective, compare results against a control group or baseline, and combine behavioural data (sales, conversions) with survey measures of awareness and consideration.
Is big data enough to evaluate a campaign?
Usually not. Big data shows what customers did in the channels you can track, but it misses offline purchases, non-customers and motivations. Surveys, qualitative research and panel data fill these gaps.
What is the difference between attribution and marketing mix modelling?
Attribution assigns credit for conversions to digital touchpoints at the user level; marketing mix modelling uses aggregated data over time to estimate the contribution of all channels, including offline media and price.
Evaluate your marketing with better data
INTAGE Vietnam is a market research company in Vietnam and part of the INTAGE Group, Japan. We help businesses evaluate marketing with brand health tracking, surveys, qualitative research and shopper panel data. See our market research services or contact us to plan a study.
Related reading
Latest posts
- Effective M&A (Mergers & Acquisitions) Opportunity Research
- Market Research: Building a Market Research Report for Investment Funds
Survey Sample Size: How Many Respondents Are Enough?
Survey Data Collection Methods in Vietnam: How to Choose- Technology Adoption in Enterprises: How to Research Employee Behavior


