Meta's New Commercial Terms: What Advertisers Must Do

Meta rewrote the contract you accept every time you run an ad. Buried in the Meta Commercial Terms update is a clause that lets its bots, and test logins, go through the website behind your ads. You can't decline it and keep advertising. So

Meta rewrote the contract you accept every time you run an ad. Buried in the Meta Commercial Terms update is a clause that lets its bots, and test logins, go through the website behind your ads.

You can't decline it and keep advertising. So the useful question is what to change on your side.

Most coverage says the new terms start on October 30. Meta's own pages say something different, and we'll get to that. First, the clause.

The clause is short and it reaches your whole website

We read the primary text, not the summaries. On Meta's Commercial Terms preview page, section 1 says you authorize Meta, "or any entity acting on our behalf", to use automated software (scrapers, bots, crawlers) and to create test credentials to "access, index, cache, analyse, collect, retrieve or crawl" the domains and pages associated with your products or services.

It also lets Meta pull updates from those pages periodically.

The stated purpose is "to provide, develop, improve, research and maintain advertising and business products." That includes products that help you create ads and "provide users information about your products and services."

You authorise us (or any entity acting on our behalf) to use automated software and create test credentials to access, index, cache, analyse, collect, retrieve or crawl the domains, web pages or URLs that are associated with your products or services.

— Meta Commercial Terms (preview), section 1

Read that as a business owner. Your product pages, your pricing page, your blog, your landing pages. All of it is "associated with your products or services."

The date depends on which Meta page you believe

Trade coverage, including The Register's report on the clause, puts the effective date at October 30, 2026. The footer of Meta's preview pages agrees.

But the banner at the top of both the Commercial Terms preview and the Self-serve Ad Terms preview reads: "These updates will come into effect on 20 November 2026." The current live Self-serve Ad Terms carry the same banner.

We can't tell you which date Meta enforces. Two dates on one page is Meta's inconsistency, not ours.

Plan for October 30. If it turns out to be November 20, you've lost nothing by being early, and you have a full pre-Black Friday window to check your ad creative.

Advertisers already agreed to a version of this in 2025

This is where the headlines overshoot. The current Self-serve Ad Terms, effective August 22, 2025, already say that by placing an Order you grant Meta the right to "access, index, cache, analyse or crawl" your domains. The purposes listed there include "research and training."

So for people who run ads, crawling rights are not new. What changed is more specific:

1. The right now lives in the Commercial Terms too, which apply to any business use of Meta products, not just ad purchases.
2. It adds test credentials and periodic re-pulls.
3. The word "train" is gone from the new text. Digital Trends reported that Meta revised the language after criticism, and the preview pages we read match that.

Narrower on paper, broader in reach. Both are true.

The business that only manages a Page is now covered

The Commercial Terms define business purposes widely: running ads, listing products, creating a catalog, managing a Page, managing a group for business purposes, or using Ads Manager or Meta Business Suite.

That means a salon owner who never spent a dollar on ads but runs a business Page is covered too. The Page is the trigger. The website the Page links to is the target.

Meta's reply to the criticism, as relayed by PPC Land, is that crawling is not changing and that the update consolidates language scattered across terms. A spokesperson pointed to robots.txt. Meta's Andy Stone said on X that the reports were false and that the change helps Meta pull accurate product information for recommendations.

Fair. But a company's statement on social media is not a clause in your contract. The contract is what binds you.

The AI rules in the ad terms matter more than the crawler

Here is our opinion, and it's not the popular one: the crawl clause is the headline and the smaller operational risk. Almost no small business will be harmed by Meta reading its pricing page. The AI creative rules are where an ordinary ad account can actually get hurt.

The new Self-serve Ad Terms preview adds three things that weren't in the August 2025 version:

• You own the AI's output. Meta says it may offer features that "optimise, enhance or create Ad Creative", that AI features "may make mistakes", and that you are "solely responsible for and assume any and all risks" associated with those features and the resulting ad creative.
• No stripping provenance. You "must not alter or remove any watermark or content-authenticating metadata generated by AI in connection with Ad Creative."
• Tests that can move your results. Meta may run tests on its ad systems that can "affect your use and experience... including campaign performance." That line existed before, but the new text adds "ad creative" to the list of things being tested.

If you've turned on Meta's automatic creative enhancements, you've handed a machine permission to edit your ads. The terms say the consequences are yours. We covered the product side in our post on Meta's AI ads assistant, and the rules for labeling in what you have to disclose on AI-generated ads.

A robots.txt that matches what you actually want

Meta's crawler documentation lists the bots and what each is for. The ones that matter to you:

• Meta-ExternalAgent collects web data for uses such as training AI foundation models or improving products.
• Meta-ExternalAds collects web data for uses such as improving ads and other business-related products.
• Meta-WebIndexer supports Meta AI search. If you allow it, Meta can cite and link to your content in Meta AI answers.
• FacebookExternalHit fetches a page when someone shares a link. Meta says it may ignore robots.txt for security and integrity checks.
• Meta-ExternalFetcher fetches links at a user's request and may ignore robots.txt.

Meta's documented example blocks meta-externalagent and meta-externalads by name. It also says changes can take up to 24 hours to apply.

Two honest caveats. Robots.txt compliance is voluntary, and critics quoted in the coverage note that Meta's terms don't promise to honor it. And Meta doesn't say what blocking Meta-ExternalAds costs you in ad product quality, so that block is a test, not a given.

Our default for a typical small business: block Meta-ExternalAgent if you don't want your content feeding model training. Leave Meta-WebIndexer open if you want Meta AI to be able to cite you. Treat Meta-ExternalAds as a decision to make deliberately, ideally on a day you aren't launching a sale. For the wider trade-off between training bots and citation bots, see should you block AI crawlers from your website.

What to check this week

You don't need a lawyer for most of this. You need an afternoon.

1. Open your robots.txt. Add a deliberate line for each Meta crawler, even if the decision is to allow it. An unexamined default is the actual mistake.
2. List what's behind login or paywall. The terms mention test credentials. Meta declined to say whether those would be used to get past gated pages. If you sell memberships or run a client portal, make sure nothing sensitive is reachable without real authentication.
3. Turn off automatic creative changes you didn't choose. In Ads Manager, review which AI enhancement toggles are on at the ad level. If you can't explain why one is on, turn it off.
4. Test one uploaded asset for metadata. Take an AI-generated image, run it through your normal editing and export steps, and inspect the final file. If the metadata is gone, your pipeline is the problem, not the generator.
5. Put a human approval step before launch. If Meta says you carry the risk of AI-edited creative, someone with authority has to look at what actually runs. We cover how a compliance pre-check slots into a high-volume pipeline in AI-native ad creative at 50 variants per week.
6. Screenshot the terms and note the date. Terms change and pages disappear. If Meta's two dates stay inconsistent, you want your own record.

If your ad creative runs through several tools and nobody owns the export step, our content team can build the pipeline and the approval layer for you.