Last month, a friend sent me two job postings. Both were from the same bank in Baku. One was titled "Business Analyst" and the other was "Data Analyst." He asked: "What's the difference?"

I read both descriptions. They were nearly identical. Same requirements (SQL, Excel, "analytical thinking"), same salary range, same team. The only difference was the title. I told him to apply to both, because whoever wrote those postings probably didn't know the difference either.

This isn't a joke — it's a systemic problem. The analytics field has become a tangled mess of overlapping titles, vague descriptions, and genuine confusion. "Business Analyst" can mean someone who writes requirements documents for software projects or someone who builds dashboards in Tableau. "Data Analyst" can mean someone who runs SQL queries all day or someone who's basically a junior data scientist training models. "BI Analyst," "Analytics Engineer," "IT Business Analyst," "Product Analyst" — the titles multiply, but the clarity doesn't.

I run BirJob, Azerbaijan's job aggregator. I see every analytics job posted across 91 websites. And I can tell you: the confusion isn't just in your head. It's baked into how companies hire.

Here's my attempt to untangle it — with data, not opinions.

The Numbers First — Because Everything Else Should Be Read Through This Lens

The U.S. Bureau of Labor Statistics projects data scientist roles to grow 36% through 2033 — one of the fastest-growing occupations in the country. Operations research analysts (which includes many analyst roles) project 23% growth — both far above the 4% average. The World Economic Forum's Future of Jobs Report 2025 lists "Data Analysts and Scientists" among the fastest-growing roles globally, estimating net creation of ~1.4 million data-related positions by 2030.

LinkedIn's Jobs on the Rise 2026 includes analytics-adjacent roles across multiple categories — AI engineer, data engineer, and business intelligence specialist all made the list. Glassdoor has ranked "Data Scientist" in its top 10 best jobs in America for nearly a decade straight.

But here's the number that actually matters for this article: when you search "analyst" on any major job board, you get dozens of different title variations. On BirJob alone, I count at least 8 distinct analyst-titled roles being actively hired for in Azerbaijan. The analytics field isn't one job — it's an entire ecosystem of roles that share a word but not always a skillset.

How We Got Here: A Brief History of "Analyst" Roles

Understanding the confusion requires understanding how these roles evolved. They didn't all appear at once — they accumulated in layers, like geological strata.

The 1990s — The Business Analyst Is Born

The original "analyst" in business was the Business Analyst (BA). Before anyone said "data-driven," companies needed people to bridge the gap between business teams and IT departments. The BA's job was clear: understand what the business needs, translate that into requirements, and make sure the tech team builds the right thing.

Tools: Word documents. Flowcharts. Requirements templates. Maybe Visio if you were fancy. The BA was fundamentally a communication role — you needed to speak both "business" and "technology" fluently, even if you couldn't code a line yourself.

The 2000s — BI Enters the Chat

As companies started sitting on more data than they knew what to do with, Business Intelligence (BI) became a discipline. BI Analysts emerged to build reports and dashboards that turned raw data into something a VP could look at and make decisions from.

Tools: SQL (now non-negotiable), early BI platforms like Business Objects, Cognos, Crystal Reports. Excel, always Excel. The BI Analyst was more technical than the BA — you needed to write queries, understand data models, and build visualizations. But you weren't doing statistics. You were answering questions like "how much did we sell last quarter?" not "what will we sell next quarter?"

The 2010s — The "Data" Prefix Explosion

Then Harvard Business Review called Data Scientist "the sexiest job of the 21st century" in 2012 (Thomas Davenport and DJ Patil), and everything went sideways. (They revisited the question in 2022 — answer: yes, but the role has evolved beyond recognition.)

Suddenly every company wanted "data people." But what kind of data people? The market fragmented:

• Data Analyst — the BI Analyst rebranded for the new decade. SQL, dashboards, reporting, but now with Python and Tableau instead of Business Objects.
• Data Scientist — supposed to be the statistical/ML expert. Builds predictive models. Uses Python, R, Jupyter notebooks, scikit-learn.
• Data Engineer — builds the pipelines that move and transform data. SQL, Python, Spark, Airflow, cloud platforms.
• Business Analyst — still exists, but now fragmented into "IT BA" (requirements/software) and "analytics BA" (data/insights)

The problem: companies didn't always know which one they needed. So they'd write a job posting that was half data analyst, half data scientist, title it "Business Analyst," and offer a salary that matched none of the above.

The 2020s — Analytics Engineer and the Modern Stack

dbt Labs coined the term "Analytics Engineer" around 2019 to describe a role that had been emerging organically: someone who combines the SQL skills of a data analyst with the engineering practices of a data engineer. According to dbt's State of Analytics Engineering survey, the role has grown an estimated 40–60% year-over-year in job postings since 2022. Analytics engineers build the transformation layer — they write the SQL that turns raw data into clean, tested, documented datasets that analysts can actually use.

This role filled a real gap. Data engineers were building pipelines but didn't always understand the business questions. Data analysts understood the questions but couldn't build production-grade data transformations. The analytics engineer sat in the middle.

By 2026, the modern analytics ecosystem has at least seven distinct roles that companies commonly hire for — and that's before you add industry-specific variants like "Product Analyst," "Marketing Analyst," "Financial Analyst," and "Revenue Operations Analyst."

The Seven Roles — What Each Actually Does

Let me be specific about what each role does on a Tuesday morning at 10am. Not the job description. The actual work.

1. Business Analyst (BA) — The Translator

What they do: Sit in meetings. Lots of meetings. They interview stakeholders to understand business problems, document requirements, map processes, and write specifications that developers can build from. They're the bridge between "the business wants X" and "the engineering team builds X."

Day-to-day: Requirements gathering, user story writing, process mapping, stakeholder management, UAT (user acceptance testing), Jira tickets, Confluence documentation.

Tools: Jira, Confluence, Lucidchart/Miro (for flowcharts), Excel, PowerPoint. SQL is a bonus, not a requirement.

Key skill: Communication. A great BA can explain a complex technical constraint to a non-technical stakeholder and make it sound obvious. They can also take a vague business wish ("we need better reporting") and turn it into actionable specifications.

Salary range: $55,000–$95,000 (US). 1,500–3,000 AZN (Azerbaijan).

2. IT Business Analyst — The BA Who Speaks Code

What they do: A specialized BA who works specifically on software development projects. They understand system architecture, APIs, databases, and can read (if not write) code. They translate business requirements into technical specifications — ERDs, API contracts, data flow diagrams.

Day-to-day: Technical requirements documents, system design sessions with developers, data modeling, integration specifications, sprint planning, testing coordination.

Tools: SQL (required), UML tools, API documentation tools (Swagger/Postman), Jira, sometimes basic Python scripting.

Key skill: Bilingualism — they speak both business and engineering at a technical level. Not "I understand technology" but "I can tell you exactly which tables need to be joined and what the API response should look like."

Salary range: $65,000–$110,000 (US). 1,800–3,500 AZN (Azerbaijan).

Why people confuse it with BA: Because many companies just call it "Business Analyst" without the "IT" prefix. The same title, completely different work.

3. Data Analyst (DA) — The Question Answerer

What they do: Pull data, analyze it, and turn it into insights that drive decisions. They answer questions: "Why did sales drop last month?" "Which marketing channel has the best ROI?" "What's our customer churn rate and what predicts it?"

Day-to-day: Writing SQL queries, building dashboards in Tableau or Power BI, cleaning messy datasets, creating Excel models, presenting findings to stakeholders, ad-hoc analysis requests.

Tools: SQL (essential), Excel (essential), Tableau/Power BI/Looker (at least one), Python or R (increasingly expected), basic statistics.

Key skill: Storytelling with data. Anyone can run a query. The good data analyst knows which question to ask, how to frame the answer, and how to present it so someone actually changes their behavior based on it.

Salary range: $55,000–$90,000 (US). 1,200–2,500 AZN (Azerbaijan). BLS reports median pay of $83,640/year for operations research analysts (the closest BLS category).

4. BI Analyst/Developer — The Dashboard Builder

What they do: Build and maintain the reporting infrastructure. Where a data analyst answers specific questions, a BI analyst builds the systems that let everyone answer their own questions. They design data models, create self-service dashboards, manage the BI platform, and ensure data quality.

Day-to-day: Building data models (star schemas, dimensional modeling), creating dashboards, writing complex SQL, managing BI platform administration, ETL/ELT pipeline maintenance, training end-users.

Tools: SQL (advanced), Tableau/Power BI/Looker (expert level), data modeling tools, sometimes dbt, sometimes Python for automation.

Key skill: Data modeling. A great BI analyst can take a chaotic production database and build an elegant, performant data model that makes self-service analytics possible.

Salary range: $65,000–$105,000 (US). 1,500–3,000 AZN (Azerbaijan).

Why people confuse it with Data Analyst: Because many companies use the titles interchangeably. In practice, a DA is more of a generalist analyst; a BI analyst is more of a builder/engineer who happens to work with analytics tools.

5. Analytics Engineer — The Modern Middle Ground

What they do: Own the transformation layer between raw data and clean, usable datasets. They write the SQL (usually in dbt) that transforms messy source data into well-structured, tested, documented tables that analysts and BI teams can trust.

Day-to-day: Writing dbt models, version-controlled SQL transformations, data testing and documentation, data quality monitoring, collaborating with data engineers on pipeline design and with analysts on data needs.

Tools: SQL (expert), dbt (the defining tool), Git, cloud data warehouses (Snowflake, BigQuery, Redshift), Python for scripting, sometimes Airflow/Dagster.

Key skill: Software engineering practices applied to analytics. Version control, testing, documentation, code review — the stuff that data analysts often skip and data engineers consider basic.

Salary range: $90,000–$140,000 (US). 2,000–4,000 AZN (Azerbaijan, rare but growing).

Why this role is confusing: It didn't exist before ~2020. It borrows from data engineering and data analysis but isn't quite either. If you haven't heard of dbt, you probably haven't heard of this role.

6. Data Scientist — The Model Builder

What they do: Build statistical and machine learning models to predict future outcomes, classify data, or find patterns invisible to human analysis. They go beyond "what happened" (data analyst territory) to "what will happen" and "why did it happen."

Day-to-day: Exploratory data analysis, feature engineering, model training and evaluation, A/B test design and analysis, statistical analysis, Jupyter notebooks, presenting results to stakeholders.

Tools: Python (primary), SQL, Jupyter notebooks, scikit-learn, pandas, NumPy, TensorFlow/PyTorch (sometimes), statistical libraries (statsmodels, scipy), visualization (matplotlib, seaborn).

Key skill: Statistical thinking. Not just "I can run a regression" but "I know when a regression is appropriate, when it isn't, and what the confidence interval actually means for the business decision."

Salary range: $90,000–$150,000 (US). 2,500–5,000 AZN (Azerbaijan). Glassdoor reports an average of ~$120,000.

7. Product Analyst — The Metric Owner

What they do: Embedded within a product team, they measure how users interact with the product, define key metrics, run A/B tests, and provide the data that drives product decisions. They're data analysts who specialize in product.

Day-to-day: Analyzing user funnels, measuring feature adoption, designing and evaluating A/B tests, defining KPIs, building product dashboards, presenting data at product reviews.

Tools: SQL, product analytics platforms (Amplitude, Mixpanel, Google Analytics), Python, Tableau/Looker, experimentation frameworks.

Key skill: Product intuition backed by data. Knowing not just what the metrics say but why they matter for the product's success.

Salary range: $80,000–$130,000 (US). 2,000–4,000 AZN (Azerbaijan, mostly at tech companies).

The Cheat Sheet — Because You're Going to Need It

Why Companies Get This Wrong

I see it every day on BirJob. A company posts a "Business Analyst" role that requires Python, machine learning experience, and "building predictive models." That's a data scientist — not a BA. Another posts a "Data Scientist" role where the responsibilities are "create weekly reports in Excel and maintain dashboards." That's a data analyst. Sometimes a BI analyst. Definitely not a data scientist.

Why does this happen?

HR departments write job postings, not hiring managers. The hiring manager says "I need someone who can work with data." HR translates that into whatever title they've heard is trendy. In 2015, that was "Data Scientist." In 2020, it was "Data Analyst." In 2026, it's probably "AI Analyst" or "Analytics Engineer." The title follows the hype cycle, not the actual work.

Small companies can't afford seven different analysts. In a 50-person company, you're not hiring a separate BA, DA, BI analyst, analytics engineer, and data scientist. You're hiring one person and calling them whatever fits. That person will do requirements gathering on Monday, SQL queries on Tuesday, a dashboard on Wednesday, and a "predictive model" (which is really a pivot table with conditional formatting) on Thursday.

The roles genuinely overlap. A data analyst who builds dashboards is doing BI analyst work. A BI analyst who writes dbt models is doing analytics engineering. A data scientist who mostly does exploratory analysis is doing data analyst work. The boundaries aren't crisp because the work isn't crisp.

The Skills That Actually Matter — Regardless of Title

After looking at thousands of analytics job postings, here's what I've noticed: the tools and titles change, but certain skills appear in every posting.

SQL — The Non-Negotiable

Stack Overflow's 2025 Developer Survey consistently shows SQL as one of the most-used languages globally. For analytics roles, it's not optional — it's the entry requirement. If you can't write a JOIN, a window function, and a CTE, you can't work in analytics. Period.

Every single one of the seven roles above uses SQL (except the pure BA, where it's a strong plus). It's the one skill that connects all analytics work.

Excel — Still Alive, Still Essential

I know this annoys people who want analytics to be "cool." But the reality is that stakeholders live in Excel and Google Sheets. You can build the most beautiful Tableau dashboard in the world — your CFO will still ask for the data in a spreadsheet. Advanced Excel skills (pivot tables, VLOOKUP/XLOOKUP, conditional formatting, basic VBA) remain valuable across all analyst roles.

Python — The Career Accelerator

Python separates the analyst who can only answer questions they've been asked from the analyst who can automate, scale, and build. It's required for data scientists, expected for analytics engineers, and increasingly expected for data analysts. The Kaggle ML & Data Science Survey found Python is used by 84% of data practitioners globally.

Visualization — The Delivery Mechanism

Tableau, Power BI, or Looker — you need at least one. The specific tool matters less than the ability to turn complex data into a clear, actionable visual. Gartner's Magic Quadrant for Analytics consistently shows Microsoft (Power BI) and Salesforce (Tableau) as leaders, but the best tool is whichever one your company uses.

Communication — The Multiplier

The most underrated analytics skill. You can write the most elegant SQL query, build the most accurate model, create the most insightful dashboard — if you can't explain what it means and why anyone should care, it doesn't matter. Every analytics role, from BA to data scientist, ultimately exists to change decisions. You change decisions by communicating clearly.

The Entry-Level Bottleneck — Is Analytics "Oversaturated"?

You'll hear this on Reddit, on Twitter, and in every analytics Discord: "Data analytics is oversaturated. Don't bother."

The data tells a more nuanced story. The BLS projects 36% growth for data scientists through 2033. Lightcast (formerly Burning Glass), which tracks job postings, reports ~300,000 unique data analyst postings in the US in 2024. The field isn't shrinking — it's growing.

But here's the catch: the Google Data Analytics Certificate alone has had 2.5+ million enrollments. Coursera, DataCamp, and dozens of bootcamps produce hundreds of thousands of "entry-level ready" candidates annually. Hiring managers report 200–500+ applications for a single junior data analyst position at well-known companies.

The oversaturation is real — but only at the entry level. Companies consistently report difficulty finding experienced (3–5+ year) analysts. The problem is a funnel: too many juniors with identical Google/IBM certificates and no domain expertise, not enough mid-level and senior analysts who can actually drive business decisions.

What differentiates you at the entry level: domain expertise (healthcare, finance, logistics), portfolio projects with real data (not Kaggle tutorials), and SQL fluency demonstrated in a technical interview — not another certificate.

Will AI Replace Analysts?

Let's address this directly, because every analytics professional is thinking about it.

The WEF Future of Jobs Report 2025 lists "Data Analysts and Scientists" among the fastest-growing roles, but also notes that 39% of existing skill sets will be transformed by 2030. McKinsey estimates that 60–70% of current analytics tasks could be augmented by generative AI — but that leads to analysts doing more high-value work, not unemployment. The Goldman Sachs estimate of 300 million jobs globally affected by AI includes many analytical functions.

Here's what AI can already do in analytics:

• Write SQL queries from natural language. ChatGPT, Claude, and GitHub Copilot can generate SQL from plain English descriptions. For simple queries, they're faster than a human.
• Build basic dashboards. Tools like ThoughtSpot and Power BI's Copilot let non-technical users ask questions and get visualizations.
• Summarize data. Upload a CSV to Claude or ChatGPT, and it'll give you summary statistics, identify outliers, and suggest patterns in seconds.
• Generate boilerplate reports. Weekly status reports, standardized analyses, KPI summaries — AI handles these faster than any human.

Here's what AI can't do (yet):

• Know which question to ask. AI answers questions brilliantly. Knowing which question matters for the business — that's a human skill.
• Understand context that isn't in the data. "Sales dropped because our main competitor launched a new product and our sales team was at an offsite" — that context lives in people's heads, not in databases.
• Navigate organizational politics. Getting a VP to actually act on data requires understanding their incentives, concerns, and communication style. AI can't do that.
• Design experiments. Deciding what to A/B test, how to structure the experiment, what confounders to control for — this requires judgment that AI assists but doesn't replace.
• Handle ambiguity. "We need better analytics" is the kind of vague request that BAs and analysts field daily. Turning ambiguity into clarity is fundamentally human work.

My honest take: AI will eliminate tasks within analytics roles before it eliminates the roles themselves. The analyst who spends 80% of their time writing repetitive SQL queries is in trouble. The analyst who spends 80% of their time framing questions, designing analyses, and communicating results is safe — and will be more productive than ever with AI handling the grunt work.

The role that's most at risk isn't any specific analyst title — it's the analyst who never moves beyond mechanical execution. If your job is "someone gives me a question, I write SQL, I make a chart," AI is coming for you. If your job is "I figure out what we should be measuring, why it matters, and what to do about it," you're fine.

Career Paths — How to Navigate the Maze

One of the most common questions I hear: "I'm a data analyst. What should I become next?" Here's how the career paths typically flow:

The hidden career path: Many of the best analytics careers don't follow a straight line. The data analyst who understands marketing becomes a marketing analytics leader. The BI analyst who understands finance becomes a CFO's right hand. The domain expertise you develop alongside your analytics skills is often what defines your career trajectory.

Certifications — Which Ones Actually Help

The certification landscape for analytics is crowded. Here's what I see employers actually recognize:

Honest advice: A certification opens a door. A portfolio project walks you through it. The Google Data Analytics Certificate alone has had 2.5+ million enrollments — it's no longer a differentiator, it's a baseline. If you're choosing between spending 6 months on a certificate and 6 months building a real analytics project that answers an interesting question — do the project. Or better yet, do the certificate in 2 months and spend the other 4 on projects. See our full guide to free certifications →

What I See on BirJob — Azerbaijan's Analytics Market

Since I have the data, let me share what's actually happening with analytics hiring in Azerbaijan:

• Banks dominate analytics hiring. Kapital Bank, PASHA Bank, ABB, and AccessBank are constantly hiring analysts. The roles are usually titled "Data Analyst" or "Business Analyst" but the actual work varies wildly between banks.
• "Data Analyst" is the most common title. It's become the default title for any analytics-adjacent role, even when the job is really BI development or business analysis.
• Analytics Engineer roles are appearing. Slowly, but they're appearing — primarily at tech-forward companies and international firms like Deloitte and Andersen.
• SQL + Power BI is the Azerbaijani standard. Most local companies use the Microsoft stack. Tableau is less common. Python is increasingly expected but not always required.
• Salaries are rising. A good data analyst in Baku earns 1,500–2,500 AZN. A senior analyst at a bank can earn 3,000–4,000 AZN. These numbers were lower two years ago.

What I Actually Think

The title confusion is a feature, not a bug. The analytics field is still young enough that roles haven't fully crystallized. This is frustrating when you're job hunting, but it's actually an opportunity — you can shape your role more than in established professions. A "Data Analyst" who starts writing dbt models becomes an analytics engineer. A "Business Analyst" who learns Python becomes something more valuable than either title alone.

The most valuable analyst is the one who can do multiple things. Pure specialization works at Google, where there are enough people to have seven distinct analyst roles. At most companies — especially in Azerbaijan — the analyst who can gather requirements, write SQL, build a dashboard, and present it to executives is worth more than three specialists. Breadth wins in small markets.

Don't obsess over the title. Obsess over the skills. If you know SQL well, can clean and visualize data, and can communicate clearly — you can get hired as a DA, BI analyst, or BA. The title on the offer letter matters less than the work you'll do and the skills you'll build. Focus on becoming genuinely good at a few things rather than collecting titles.

AI will make analysts more powerful, not obsolete. The analyst who embraces AI tools (ChatGPT for SQL generation, Claude for data summarization, Copilot for Python) will be 3–5x more productive than the one who doesn't. That productivity gap will widen every year. The question isn't "will AI replace analysts?" It's "will analysts who use AI replace analysts who don't?"

The field is still growing. Despite the confusion, despite the AI hype, despite the overlapping titles — the BLS projects 23% growth through 2033. The WEF says data analysts are among the fastest-growing roles globally. Every company is drowning in data and desperate for people who can make sense of it. The demand is real. The confusion about what to call those people doesn't change that.

If You're Trying to Break In Right Now

Here's the shortest path into each role:

• Business Analyst: Get CBAP-adjacent training, learn to write requirements documents and process flows, practice stakeholder communication. No coding required.
• Data Analyst: Learn SQL (seriously — Mode's SQL tutorial is free and excellent), pick up Tableau or Power BI, take the Google Data Analytics certificate, and build a portfolio project analyzing a real dataset.
• BI Analyst: Master SQL + one BI tool deeply. Learn dimensional modeling. The Kimball methodology is still the gold standard for BI data modeling.
• Analytics Engineer: Learn SQL, then dbt. dbt's free courses are the best starting point. Contribute to open-source dbt packages. This role values engineering discipline as much as analytics skill.
• Data Scientist: Python + statistics + a portfolio of ML projects. The IBM Data Science certificate is a good foundation, but you'll need to go beyond it with personal projects.

Sources

• U.S. Bureau of Labor Statistics — Operations Research Analysts Occupational Outlook
• World Economic Forum — Future of Jobs Report 2025
• LinkedIn — Jobs on the Rise 2026
• Harvard Business Review — Data Scientist: The Sexiest Job of the 21st Century (2012)
• dbt Labs — What Is Analytics Engineering?
• Stack Overflow — 2025 Developer Survey
• Goldman Sachs — How Will AI Affect the Global Workforce
• Glassdoor — Best Jobs in America
• Glassdoor — Data Scientist Salary
• Kaggle — ML & Data Science Survey
• Gartner — Magic Quadrant for Analytics and BI Platforms
• Google Career Certificates — Data Analytics
• Kimball Group — Dimensional Modeling Techniques
• U.S. Bureau of Labor Statistics — Data Scientists Occupational Outlook (36% growth)
• Harvard Business Review — Is Data Scientist Still the Sexiest Job? (2022 follow-up)
• McKinsey — The Economic Potential of Generative AI
• dbt Labs — State of Analytics Engineering 2024
• Lightcast (formerly Burning Glass) — Job Posting Analytics
• Robert Half — 2025 Salary Guide
• IIBA — Business Analysis Certifications (CBAP/CCBA)

I'm Ismat, and I build BirJob — Azerbaijan's job aggregator. I scrape 91 job sites daily and see every analyst posting in the market. If this article helped you make sense of the analytics landscape, support the platform at birjob.com/support.

You might also like

• The Data Engineer Shortage Nobody Talks About: Why It Pays More Than Data Science Now
• Data Analyst vs Data Scientist vs Data Engineer: The 2026 Decision Guide
• AI Engineer vs ML Engineer: What Actually Changed and Why It Matters
• DevOps vs SRE vs Platform Engineer: The Infrastructure Title Mess, Explained