IMPORTXML not working? How to fix it, and an alternative when you can't

Open the page in an incognito window, so you see it signed out. Press Ctrl+U (Option+Cmd+U on a Mac) for View page source, not Inspect, and search for your value. Page source is what the server sends, which is what IMPORTXML sees ^{Source 1}. Inspect shows the page after scripts ran.

• In the source, inside normal tags: the formula can reach it. Fix the XPath.
• Only inside a script tag, often as JSON: IMPORTXML can return the whole script, and REGEXEXTRACT can cut your value out of it ^{Source 2, Source 3}. It breaks when the site changes its JSON, and one cell holds at most 50,000 characters ^{Source 4}, so a long script may not fit. Apps Script can parse the JSON properly, if you write code.
• Not there at all: a script loads it later, or only signed-in visitors get it. No built-in formula will find it in the page. On a public page, open DevTools, go to Network > Fetch/XHR and reload. If a request returns your value as JSON, Apps Script can fetch that URL directly ^{Source 4}.

Quotes to Scrape, a site made for scraping practice, serves the same ten quotes three ways, and the three pages look the same in your browser. Put each URL in A1 and try =IMPORTXML(A1, "//span[@class='text']").

• Normal tags. With https://quotes.toscrape.com/ in A1, the formula returns the ten quotes.
• Inside a script. With https://quotes.toscrape.com/js/, it returns #N/A. Page source shows why: the quotes sit as JSON in a <script>, in var data = [...], and a script draws them. =REGEXEXTRACT(IMPORTXML(A1, "//script[contains(.,'var data')]"), "\\u201c(.+?)\\u201d") returns the first quote. The JSON writes the curly quote marks as \u201c and \u201d. In the pattern, \\ matches one backslash, so it finds those codes and leaves them out. If the site changes its JSON, the formula breaks.
• Not there at all. With https://quotes.toscrape.com/scroll, it returns #N/A, and page source has only an empty <div class="quotes">. Open DevTools, go to Network > Fetch/XHR and reload: a request to /api/quotes?page=1 returns the quotes as JSON. Apps Script can fetch that URL with UrlFetchApp.fetch and read it with JSON.parse ^{Source 5}.

Each cause below ends with whether a change to the formula can fix it.

• The page uses JavaScript. Import functions read the HTML the server sends and don't run scripts ^{Source 4}. The cell shows #N/A and says the imported content is empty ^{Source 1}. Google's help doesn't mention this. Fixable only when the data sits in a script on the page, as above.
• The page needs a login. The request comes from Google's servers, without your browser session ^{Source 4}, so it sees what a signed-out visitor sees, usually a sign-in page. No formula fix.
• The XPath doesn't match the raw HTML. This gives the same empty result ^{Source 6}, which is why the page source test comes first. XPaths copied from DevTools describe the page after the browser changed it ^{Source 1}. Browsers add a <tbody> to tables that lack one ^{Source 7}, so a copied /tbody/ path can match nothing. Fixable.
• Could not fetch url. The URL must include the protocol, such as https:// ^{Source 8}. If it does, the site is probably refusing Google's requests, as some sites do when Sheets sends too many ^{Source 9}. No formula gets past a block.
• Result too large, or Resource at url contents exceeded maximum size. For the first, Google's fix is an XPath that returns less ^{Source 10}. The second means the page itself is over a size limit Google doesn't publish ^{Source 4}. Look for a lighter URL with the same data, such as a JSON request from the Network tab, which Apps Script can fetch. Fixable sometimes.
• Stuck on Loading: too many imports. When import functions create too much traffic, cells show "Loading data may take a while because of the large number of requests" ^{Source 10}. The quota is the creator's, shared across all the open spreadsheets they created, and Google's fix is to use fewer import functions ^{Source 10}. Guides often quote a 50-formula cap; Google's help gives no number. Fixable.
• Stale numbers: the refresh is hourly. Imports check for updates every hour while the document is open. Re-entering the formula forces a refresh; reopening the file doesn't, and NOW() or RAND() can't speed it up, because import functions can't reference them ^{Source 10}. Partly fixable.

When the value is in the page source, outside any script, try these in order.

• Test the fetch. With https://quotes.toscrape.com/ in A1, =IMPORTXML(A1, "//title") returns Quotes to Scrape. If //title fails on your page, the fetch is the problem, not your XPath.
• Skip levels. //table[@id='prices']//tr/td[2] works with or without a <tbody>. Anchor on an id or class, not a long position path.
• Match one class among several. //span[contains(@class,'price')] matches class="price sale", but also price-old and unit-price. For the exact class, use //span[contains(concat(' ', normalize-space(@class), ' '), ' price ')]. Use single quotes inside the XPath, because the formula wraps it in double quotes.
• Return one value. =INDEX(IMPORTXML(A1, "//span[@class='text']"), 1) keeps only the first quote from the practice page.
• IMPORTHTML counts from 1. =IMPORTHTML(A1, "table", 2) is the second table on the page. Tables and lists are numbered separately ^{Source 11}.

If the formula works, keep it. It's free, it lives in the cell, it needs no extension, and it runs on Google's servers, so it doesn't depend on Chrome being open.

If the value isn't in the page source at all, or you sign in to see the data, no built-in formula can reach it. Switch then, or when the site blocks Google, or when you need fresh rows at a set time, not only while someone has the sheet open. For marketplace product pages on Amazon, Walmart and similar sites, one option is Amapulse, formerly ImportFromWeb: a Sheets add-on that pulls data even from JavaScript-rendered pages, with a one-month free trial that includes 200 credits ^{Source 12}.

One thing changes when you switch to BrowseWiz. It can't write into the sheet that holds your IMPORTXML formulas, or into any sheet it didn't create. It creates a new sheet in your Google Drive and keeps writing to it, on every plan. To see those rows next to your formulas, pull them in with IMPORTRANGE and click Allow access the first time ^{Source 13}.