Getting Started With Data-Driven Marketing When You're Not a Data Person
Start marketing with data without coding: ask the right questions, choose metrics that reflect the business, run small tests and avoid vanity numbers.
Every Monday morning your marketing team opens a long report full of reach, likes, clicks and nice charts. But when someone asks "which channel should get more budget next month?", everyone answers with gut feeling. There's plenty of data, but it isn't used to decide. Data-driven marketing doesn't start with more data. It starts with using the data you have to answer questions that matter.
What data-driven marketing is
It's planning, deciding and adjusting marketing based on evidence from data rather than personal preference or guesswork: choosing audiences, channels, messages and budgets. The goal is for every decision to have an explainable reason, and to learn from real results faster.
You can start without coding
Most marketers already have data: reports from Meta Ads or Google Ads, website data in Google Analytics, sales from the back-office system and customer data from LINE OA. The first skills you need are asking questions and reading numbers. Excel or Google Sheets is enough to begin.
Four steps to marketing with data
1. Start with a business question
Don't start with "what data do we have?" Start with "what do we need to decide?" For example: which channel brings new customers most cost-effectively? Why do people add items to the cart but not pay?
2. Choose metrics that answer it
If the question is about value for money, the metric is cost to acquire one customer, not clicks. If it's about repeat buying, look at repeat purchase rate and time between purchases.
3. Run small experiments
An A/B test compares two versions that differ in only one thing, such as two email subject lines or two ad images, to see which performs better. Change one thing at a time so you know what caused the result.
4. Decide, record and repeat
Sum up the result in one sentence, such as "Product-in-use images got more orders than white-background images, so we'll make them the main approach next month," and add it to the team's experiment log.
A real-world example
An online cosmetics shop saw website visits rise while sales stayed flat. Looking at the data, the team found many mobile users leaving at the payment page. They tried cutting the number of form fields and adding a QR payment option, then compared results with the previous period. The interesting part: the answer wasn't in the ads at all, but in the experience after the click.
Write reports people can decide from
A good report isn't long. It answers three things clearly: what happened, why, and what we should do next. Try turning a ten-page weekly chart deck into one page with no more than five key metrics, each with a short observation and recommendation. Executives will read to the end, and the team will start discussing the same numbers.
Common traps
- Vanity metrics: likes or views that look good but don't relate to sales or goals.
- Concluding from too little data: deciding after a two-day test, when the result may be chance.
- Confusing correlation with cause: sales rose while ads ran, but it may also have been a festive season.
- Mismatched data between systems: ad platform and back-office numbers often differ, so understand how each system counts.
A caution on customer data
Using customers' personal data must follow Thailand's PDPA and your company's policies. Use only what's necessary, and be careful about putting customer data into external tools or AI tools your company hasn't approved.
Data doesn't decide for you. It helps you make fewer wrong decisions and notice sooner when you do.
How to start this week
- Write down one business question your team needs to decide this month.
- Choose 1 or 2 key metrics that answer it.
- Gather the data you already have into one Google Sheet.
- Design one small A/B test that changes only one thing.
- Sum up the result in one sentence and share it with the team.
To know which metrics to watch, read marketing KPIs you should know and marketing analytics basics. To practise using data hands-on, see courses from Plants.work learning partners.