Should You Learn Excel or SQL First?
· 10 min read
If you are starting from zero and aiming for a data analyst job, learn basic Excel first, start SQL a couple of weeks later, and then keep both going side by side. Excel shows you every row and every result as you work, which makes it the easiest place to learn what filtering, grouping and totals actually do. SQL does the same things on data too big or too spread out for a spreadsheet, and it is the skill analyst interviews tend to test hardest. If you already use spreadsheets at work, skip ahead and start with SQL now.
So the honest answer to "should I learn Excel or SQL first" is: it depends on where you start, and it matters less than people think, because the two tools teach the same thinking. The rest of this guide shows why, with one business question answered in both.
The short answer by starting point
| Where you are now | Learn first | Why |
|---|---|---|
| Never used formulas | Excel | You see the data and the result in the same place, so mistakes are visible |
| Use spreadsheets, know SUM and filters | SQL | You already think in rows and columns; SQL is the bigger gap |
| Comfortable with SUMIFS, lookups and pivots | SQL, seriously | Most of what you do in Excel has a direct SQL equivalent |
| Studying for a specific job | Whatever the job ad lists first | Many analyst ads ask for both; read the ad, not the internet |
| Aiming for data engineering or software | SQL | Spreadsheets matter less in those roles |
None of these paths means skipping the other tool. Analysts use both, often on the same day.
Why Excel first works for most beginners
You can see everything. In a sheet, the data, the formula and the answer are on one screen. When a total looks wrong, you click the cell and look. In SQL, the result is a new table and the rows that went into it are hidden unless you go looking. That makes Excel a forgiving place to learn what "filter, then add up" means.
Feedback is instant. Type a formula, press Enter, see a number. Change one input and every dependent cell updates. That loop teaches cause and effect faster than writing a query, running it and reading an error message.
It is still where a lot of analyst work ends up. A common question from people moving into data work is why analysts still use Excel when databases exist. The answer is that the last step of many analyses happens in a spreadsheet: a quick check of an export, a small model a manager can change, a table someone outside the data team can open. You will be asked to work in Excel even if you prefer SQL. Our Excel practice exercises list the formulas that come up most.
Why SQL first can be the better choice
The data lives in databases. At most companies with a data team, orders, customers and events sit in a database or warehouse, not in a file. SQL is how you get them out. If you cannot write a query, you depend on someone else to export the data for you.
It scales where spreadsheets struggle. A full year of order lines, joined to customers and products, is easy in SQL and slow or error-prone in a sheet. Joins in particular are clearer in SQL than chains of lookups.
It is what gets tested. Analyst interviews often include a live SQL task. If your spreadsheet skills are already solid, every hour on SQL is worth more than another hour on Excel. See SQL for data analysis for the skills that matter most.
The same question in Excel and in SQL
Here is why the order matters less than it seems. Brightline Supply Co. sells office, kitchen and cleaning supplies to businesses in four regions. Revenue counts completed orders only, and each order line's revenue is quantity × unit price × (1 − discount). The question: how much completed revenue did each region bring in from 3 to 5 March 2025?
In Excel, you calculate revenue per line in column G, then add it up per region with SUMIFS, keeping only completed orders:
| A | B | C | D | E | F | G | H | I | J | |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Order | Status | Region | Qty | Price | Discount | Revenue | Region | Revenue | |
| 2 | 50182 | completed | South | 1 | 9.90 | 0% | 9.90 | East | 2,019.03 | |
| 3 | 50182 | completed | South | 5 | 24.00 | 0% | 120.00 | North | 507.00 | |
| 4 | 50182 | completed | South | 1 | 22.00 | 5% | 20.90 | South | 835.14 | |
| 5 | 50183 | completed | East | 5 | 21.50 | 5% | 102.13 | West | 1,823.45 | |
| 6 | 50184 | completed | South | 6 | 16.50 | 0% | 99.00 | Total | 5,184.62 | |
| 7 | 50184 | completed | South | 1 | 17.50 | 0% | 17.50 | |||
| 8 | 50185 | completed | North | 14 | 15.00 | 0% | 210.00 | |||
| 9 | 50186 | completed | East | 11 | 59.00 | 10% | 584.10 | |||
| 10 | 50186 | completed | East | 13 | 84.00 | 10% | 982.80 | |||
| 11 | 50186 | completed | East | 14 | 12.50 | 0% | 175.00 | |||
| 12 | 50186 | completed | East | 7 | 25.00 | 0% | 175.00 | |||
| 13 | 50187 | cancelled | East | 14 | 19.90 | 0% | 278.60 | |||
| 14 | 50188 | completed | North | 22 | 13.50 | 0% | 297.00 | |||
| 15 | 50189 | cancelled | North | 9 | 12.50 | 5% | 106.88 | |||
| 16 | 50189 | cancelled | North | 6 | 24.00 | 0% | 144.00 | |||
| 17 | 50190 | completed | South | 4 | 5.50 | 0% | 22.00 | |||
| 18 | 50190 | completed | South | 4 | 15.00 | 5% | 57.00 | |||
| 19 | 50190 | completed | South | 5 | 23.50 | 0% | 117.50 | |||
| 20 | 50191 | completed | South | 6 | 6.20 | 5% | 35.34 | |||
| 21 | 50191 | completed | South | 4 | 84.00 | 0% | 336.00 | |||
| 22 | 50192 | completed | West | 4 | 11.00 | 0% | 44.00 | |||
| 23 | 50192 | completed | West | 3 | 15.50 | 0% | 46.50 | |||
| 24 | 51110 | completed | West | 29 | 24.00 | 10% | 626.40 | |||
| 25 | 51110 | completed | West | 24 | 36.00 | 10% | 777.60 | |||
| 26 | 51110 | completed | West | 17 | 21.50 | 10% | 328.95 |
In SQL, the same steps are clauses. WHERE keeps completed orders in the date range, the multiplication works out line revenue, and GROUP BY adds it up per region. Region lives in other tables, so the query joins orders to their lines, customers and regions:
SELECT rg.region,
ROUND(SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct)), 2) AS revenue
FROM orders AS o
JOIN order_items AS oi ON oi.order_id = o.order_id
JOIN customers AS c ON c.customer_id = o.customer_id
JOIN regions AS rg ON rg.region_id = c.region_id
WHERE o.status = 'completed'
AND o.order_date BETWEEN '2025-03-03' AND '2025-03-05'
GROUP BY rg.region
ORDER BY rg.region;| region | revenue |
|---|---|
| East | 2019.03 |
| North | 507 |
| South | 835.14 |
| West | 1823.45 |
Same four numbers. Now delete the date line from the query and run it again. You get the whole year, all 1,020 completed orders, in the same amount of time:
SELECT rg.region,
ROUND(SUM(oi.quantity * oi.unit_price * (1 - oi.discount_pct)), 2) AS revenue
FROM orders AS o
JOIN order_items AS oi ON oi.order_id = o.order_id
JOIN customers AS c ON c.customer_id = o.customer_id
JOIN regions AS rg ON rg.region_id = c.region_id
WHERE o.status = 'completed'
GROUP BY rg.region
ORDER BY revenue DESC;| region | revenue |
|---|---|
| West | 100293 |
| South | 90503.1 |
| North | 84436.87 |
| East | 60763.76 |
That is the trade-off in one example. Excel let you see all 25 lines and check the cancelled order by eye. SQL needed four tables joined correctly, but then the year took no extra effort. The thinking was identical: which rows count, how is one row's value worked out, and what do you group by. Whichever tool you learn first, you are learning that thinking, and it carries straight over. The syntax for the SQL side is in SQL GROUP BY and joining two tables; the Excel side is in SUMIFS with multiple criteria.
Python or SQL first?
For analyst roles, SQL first. Most analyst work starts with pulling and shaping data, which is SQL's job, and SQL is small: a few clauses cover most of what you will write. Python is a general programming language. It is worth learning later for automation, statistics and work that does not fit in a query, but it is a bigger first step and you will still need SQL to get your data. If you are aiming at data science or machine learning, the order can flip, but you will want both in the end.
Can you learn Excel, SQL and a BI tool at the same time?
Excel and SQL, yes. They reinforce each other: a SUMIFS is a WHERE plus a GROUP BY, a lookup is a join. Spending a few days on each, alternating, works well.
Adding a BI tool such as Tableau or Power BI at the same time usually spreads you too thin. A dashboard is only as good as the numbers behind it, and producing correct numbers is the Excel and SQL part. Learn to get the right totals first, then learn to chart them. If you are weighing the options, see Power BI vs Excel.
A learning order that works
- Excel basics. Cells, formulas, relative and absolute references, SUM, AVERAGE, COUNT.
- SQL basics.
SELECT,WHERE,ORDER BY. Look at one table at a time. Start this while you are still on step 1. - Summaries in both. SUMIFS and COUNTIFS in Excel;
GROUP BYand aggregates in SQL. Answer the same question both ways, as above. - Lookups and joins. XLOOKUP or INDEX MATCH, then
JOINandLEFT JOIN. - Checking. Reconcile totals, count rows before and after a join, test a formula on rows you can work out by hand.
- The next layer. Pivot tables and dynamic arrays in Excel; CTEs and window functions in SQL.
Step 5 is the one most courses skip, and it is the one employers notice. More on SQL practice in SQL practice online.
Common mistakes
Waiting to "finish" Excel before starting SQL. There is no finish line. Once you can write a SUM and a SUMIFS, start SQL.
Learning syntax on toy tables only. A one-table example with clean data teaches the keywords but not the judgment. Real data has statuses that should not count, missing values and tables that need joining. Practise on something that looks like a business.
Skipping Excel because it seems basic. Analysts are expected to be quick in a spreadsheet. An interview task in Excel is common, and a slow or messy one makes a bad impression even if your SQL is strong.
Collecting tools instead of skills. Excel, SQL, Python, a BI tool and statistics all at once usually means none of them sticks. Two at a time is plenty.
FAQ
Is SQL harder than Excel?
The basics are not harder, just less visual. SQL has fewer moving parts than Excel's hundreds of functions, but you cannot see the rows behind a result unless you query them, so mistakes are easier to miss.
Can I get a data analyst job with only Excel?
Some roles, especially in finance, operations and small companies, are mostly spreadsheet work. Many analyst roles also ask for SQL, so learning it widens the jobs you can apply for.
Should I learn Python or SQL first?
SQL first for analyst roles. It is smaller, you will use it daily, and you need it to get data anyway. Add Python once you can write joins and GROUP BY queries comfortably.
Can I learn Excel and SQL at the same time?
Yes, and it often helps. Solve the same question in both tools and compare: the overlap makes each easier.
Do I need to be good at maths?
Not for this. Everyday analysis is arithmetic, percentages and averages. Being careful matters more: knowing which rows should count and checking that totals add up.
Practise both
You can start both tools on the same company's data. Practisheet's first five Excel exercises and first four SQL exercises are free, with no sign-up. Begin with Type a number in Excel or Your first query in SQL, then try the SQL exercises in the practice section below.
Practise this on a realistic dataset
Solve it yourself on one company's data, with instant feedback that points at what's wrong without giving away the answer.
Try “Your first query” →