How to Scrape a Website With Google Sheets or Excel, No Code

The short answer to how to scrape a website without code: let a spreadsheet do it. If the data sits in a table or a list on a public page, Google Sheets pulls it in with one formula, IMPORTHTML, and Excel does the same through Power Query's From Web command. Both come with software you already have, both refresh, and neither asks you to program anything. They stop at pages that build their content with JavaScript, pages behind a login and jobs that span hundreds of pages. That is where an agent with a real browser takes over.

How to scrape a website with a spreadsheet you already have

Scraping means copying data from web pages into rows and columns you can sort, filter and compare. A spreadsheet can do the copying itself, because both Google Sheets and Excel can fetch a page and read its structure. The whole method is four moves: find the page, find the table or element that holds the data, import it, then freeze a copy when you need a record.

Check one thing before you start. Open the page, right-click and view the page source, then search for a value you can see on screen, such as a product name. If the value is in the source, a spreadsheet can read it. If it is missing, the page loads its data after opening and the free route will return nothing useful. Our guide to web scraping explains why pages behave so differently.

The spreadsheet route fits small, steady jobs: one supplier price table, a public directory, a schedule, a league table, a list of store locations. It suits anyone who wants to scrape data from a website a few times a month and would rather not learn a new tool for it.

Google Sheets: IMPORTHTML for tables, IMPORTXML for the rest

IMPORTHTML takes three arguments: the page address, the word "table" or "list", and a number that says which table or list on the page you want, counting from 1.^[1] Tables and lists are numbered separately, so a page can have a table 1 and a list 1.

1. Open a blank sheet and paste the page address into cell A2, including the https:// part.
2. In A4, type =IMPORTHTML(A2,"table",1) and press Enter.
3. If the wrong table arrives, change the 1 to 2, then 3, until the right one appears.
4. If the data is a bulleted or numbered list, replace "table" with "list".
5. Leave the cells below and to the right empty. The import fills them.

When the data is not in a table, use IMPORTXML. It takes the address and an XPath query, a short expression that describes where the data sits in the page. With the XPath from Google's own sample, =IMPORTXML(A2,"//a/@href") returns every link on a page. In the same way //h2 returns every second-level heading and //span[@class='price'] returns each element marked with that class name. This is the part people search for as importxml google sheets, and it rewards ten minutes of trial and error.

Sheets keeps the import alive. IMPORTHTML, IMPORTXML and IMPORTDATA check for updates every hour while the document is open, and reloading the tab does not force a refresh; deleting the formula and entering it again does.^[2] The imported cells are a live view, not a record. To keep today's numbers, copy the range and use Paste special, Values only, on another tab.

Scrape a website to Excel with Power Query

Excel's route is a menu instead of a formula. Microsoft's steps for Excel for Microsoft 365 are short.^[3]

1. Open a workbook and select the Data tab.
2. In the Get & Transform section, select From Web.
3. Paste the page address into the box and select OK.
4. In the Navigator pane, click each detected table. A preview appears on the right.
5. Select Load. Excel places the data in the workbook as a table.

The result stays connected to the page, so refreshing the query pulls the current version. Power Query also cleans as it loads. You can open the editor from the Navigator, remove columns, change types and filter rows, and those steps repeat at every refresh.

When the Navigator finds no table, look for Add table using examples. Microsoft's Power Query documentation describes it this way: you type a couple of sample values from the page into a grid, and Power Query extracts the other values that fit the pattern.^[4] It is the closest thing to point-and-click scraping in software you already have. It handles product cards and listings that are consistent but not laid out as a table.

Three supplier pages, one comparison sheet

Picture a small landscaping company that buys from three suppliers. Each supplier publishes a price table on a public page. The owner wants them side by side every Monday.

She creates one tab per supplier. On each tab, A2 holds the page address and A4 holds =IMPORTHTML(A2,"table",1). Two tables land on the first try. The third needs index 2, because the page opens with a small layout table. A fourth tab, Compare, uses ordinary lookups to line up the items she buys most.

On Monday she opens the file, waits for the imports to settle and pastes the Compare tab as values into a tab named with the date. Total setup: under half an hour. She also compares five rows against the live pages by eye. A site redesign can shift a column without any error message.

Site rules and server load come before the first import

Free web scraping is still collecting from someone else's site. Read the site's terms and its robots.txt file, which lists the areas the owner asks automated visitors to stay out of. Skip personal data you have no right to use: names, emails and phone numbers of private people are not raw material. And keep the load light. One import per page, refreshed hourly at most, is polite. Dozens of formulas hitting the same small site are not.

Five signs the spreadsheet has run out of road

The free route breaks in specific, recognizable ways.

• The cells come back empty. The page builds its content with JavaScript, and import formulas read only the HTML source.
• The data is behind a login. A formula cannot sign in for you.
• The list runs over many pages. Every page needs its own formula, and "load more" buttons cannot be clicked.
• You need hundreds of pages. Google throttles heavy use of import functions across all the spreadsheets you created.
• You need a schedule. Sheets refreshes only while the document is open.

One of these is an annoyance. Two or more mean the job needs a browser that an agent can drive. The agent opens pages the way a person does, waits for them to render, clicks through and writes rows.

Gumloop suits a recurring job on public pages that the whole team relies on, such as a weekly price or listings sheet where someone will ask where a number came from. Its setup is covered in our guide to Gumloop web scraping. If your pages sit behind an account you already use every day, the next option starts from a different place: your own browser.

That makes Manus AI the natural step for a solo professional pulling data from portals they already log in to, with the option to grow into batches later. Our walkthrough of the Manus AI Browser Operator shows the setup. Not every job is a batch, though. Some come up once and never again.

Whichever agent you pick, keep the habit from the spreadsheet days: compare a handful of rows against the live pages before the data goes into a report or a system of record. Agents misread pages too. For the full set of parts a collection pipeline needs, and how more products cover them, see our comparison of web scraping tools.

One table tonight, an agent only when it breaks

Pick the single page you check most often and import it with IMPORTHTML or From Web before you close this tab. If the table arrives, you have a working scraper and a reason to add the second page. If the cells stay empty, or the page wants a password, you have learned in five minutes that this job belongs to a browser agent. Either way you know which side of the line you are on.