Data Analytics: What It Is, 4 Types and Where to Start
In short
Data analytics is the process of collecting, organizing and analyzing data to make better decisions. There are four types of analysis (descriptive, diagnostic, predictive and prescriptive), and the path starts with a business question, not a tool.
Data analytics is the process of collecting, organizing and analyzing data to answer business questions and support decisions. It turns loose numbers, such as clicks, sales and leads, into information someone can use to decide what to do next week.
In practice, most companies don't lack data. They struggle with data scattered across tools that don't talk to each other, metrics nobody owns and reports nobody uses. This guide shows how to organize it, from the concept to the first step.
The 4 types of data analysis
Each type answers a different question. The further right in the table, the more value and the more difficulty.
| Descriptive | Diagnostic | Predictive | Prescriptive | |
|---|---|---|---|---|
| Question | What happened? | Why did it happen? | What is likely to happen? | What should I do? |
| Marketing example | Leads generated this month | Why cost per lead went up | Which leads are most likely to buy | Where to reallocate the media budget |
| Foundation needed | Reliable, collected data | Segmented data that can be cross-referenced | Long history and statistical models | Models and clear business rules |
Most of the value a company gets early on comes from the first two types. Predictive models only work well on top of a solid descriptive base.
Data analytics, BI, data science and marketing analytics
The terms get mixed up, but each has a different focus:
- Data analytics: analyzing data to answer specific questions.
- Business Intelligence (BI): recurring dashboards and reports that track business indicators.
- Data science: statistical and machine learning models to predict or automate.
- Marketing analytics: analysis applied to channels, campaigns and the customer journey, usually with Google Analytics 4, a CRM and ad platforms.
The steps of the process
- Define the question. Without a question, it becomes a report. Example: "which channel brings the leads that most often become sales?".
- Collect. Bring together data from the site, paid media, CRM, email and sales.
- Clean and organize. Standardize names, remove duplicates and set a single source for each metric.
- Analyze. Segment, compare periods, cross sources and look for the reason behind the number.
- Visualize. Build a simple dashboard with few indicators and an owner for each one.
- Act and measure again. Analysis only counts when it changes a decision. Then you measure the effect.
Where to start
Order matters, because each step prepares the next.
- Pick 3 to 5 business questions. For example: where customers come from, how much it costs to win one and which funnel stages lose the most people.
- Define the metrics and who is accountable for each. Write the formula. "Lead" must mean the same thing to marketing and sales.
- Secure the collection. Install Google Tag Manager, set up GA4 with the key conversion events (form, WhatsApp click, purchase) and respect user consent.
- Connect the sources. Join site, media and CRM with standardized UTM parameters. This is where most projects get stuck.
- Build a lean dashboard. One page, few numbers and a comparison with the previous period.
- Create a routine. A short, recurring meeting to look at the numbers and decide one action.
Tools by category
There is no single right tool, only the combination that fits your stage:
- Collection: Google Tag Manager and GA4.
- Storage: spreadsheets at first, then a data warehouse such as BigQuery.
- Transformation: SQL and data-cleaning routines.
- Visualization: Looker Studio, Power BI or Tableau.
- Activation: CRM and automation, to use the data in communication and media.
Common mistakes
- Starting with the tool instead of the question.
- Measuring everything and prioritizing nothing.
- Having different numbers for the same metric in each team.
- Not standardizing UTMs, which breaks lead source tracking.
- Ignoring data quality and trusting a beautiful dashboard built on a wrong base.
- Delivering a report with no recommended action.
How to know it is working
Three simple signs: teams agree on the numbers, decisions cite the data and you can explain why a result went up or down. If reports still raise doubts about where a number comes from, the foundation still needs work.
Quick questions
What is data analytics in one sentence?
It is the process of turning data into useful information for decisions, through collection, organization, analysis and action.
What is the difference between data analytics and BI?
BI tracks indicators on a recurring basis, in dashboards and reports. Data analytics is broader and includes on-demand analysis to answer specific questions, such as investigating the cause of a drop.
Do I need a data warehouse to start?
No. Many companies start well with GA4, spreadsheets and a Looker Studio dashboard. A data warehouse makes sense when volume grows or when you need to cross many sources.
Where do I start a career in data analytics?
With the fundamentals: Excel or spreadsheets, SQL, basic statistics and a visualization tool. Practicing with real business data speeds things up more than any single course.
Want to organize your marketing data? Also read SEO, AEO and GEO: the difference and where to start or get in touch.
Want to apply this to your business?
Talk to me and tell me about your situation.