Reconsideration request builder

01Cannabis sites get hit from two directions at once

Most industries have one dominant manual action risk. Cannabis has two, and both of them trace back to the same structural fact: paid advertising is closed to you on the major platforms. That single constraint shapes how dispensary marketing behaves, and it pushes operators toward the exact behaviours Google penalises.

Because you cannot buy visibility with ads, you cover every market with location pages, and at scale those pages tend toward a template. And because you cannot buy visibility with ads, you buy links instead, which is the other lever available when the obvious one is switched off. Scaled content abuse and unnatural inbound links are therefore the two actions that dominate this niche, and it is common to see a site carrying risk on both fronts simultaneously.

The same constraint that makes organic search essential in cannabis is the one that pushes operators into the behaviours that get it taken away.

02The content side: thin and scaled

Thin content with little or no added value and scaled content abuse are closely related findings, and on a dispensary chain they usually describe the same set of pages. Dozens of location pages, built from one template, differing only by city name and address. To the team that shipped them it looked systematic. To a reviewer it looks like pages produced to rank rather than to help anyone.

There is a cannabis specific aggravating factor that most recoveries miss entirely. If your product menu loads inside an iframe or through JavaScript, the genuinely unique content on each page never reaches Google. Strip the menu out and what remains is the templated wrapper, which is why dispensary pages score far thinner than the team believes them to be. Rewriting the wrapper without fixing the menu delivery means the pages stay thin and the request gets denied, and the team concludes Google is being unreasonable when the real problem was never addressed.

A credible content recovery has three outcomes across the affected set, not one. Some pages get rewritten from genuinely local facts. Some get consolidated into stronger single pages with redirects. Some get removed entirely, because closed locations and stub pages will never carry value. A recovery where every page survived and simply got longer reads as an audit that was not honest.

03The link side: bought, blasted, and inherited

Unnatural links to your site is the second dominant action here, and cannabis operators arrive at it by four routes. Some buy links outright. Some run guest posting at a volume and quality that stops being editorial and becomes link building. Some blast directory and citation submissions until the profile looks manufactured. And a great many inherit the problem from a previous agency whose work was never audited.

The last case is the most common and the most frustrating, because the current team is cleaning up something they did not do. It makes no difference to the outcome. The site owns the links regardless of who built them, and a request that leans on blaming a predecessor lands badly. Acknowledge it happened, take responsibility for not auditing it, and describe the cleanup.

On remediation, the ordering matters more than people expect. Google expects genuine removal attempts first and disavow second. A request that shows only a disavow file, with no outreach log, reads as taking the shortcut, and denials say so explicitly. Contact the worst sources, log every attempt including the ones that never reply, then disavow what will not come down. The outreach log is often the single most persuasive thing in the whole submission.

And the campaign has to have stopped. Cleaning up links while continuing to acquire more is the clearest possible signal that nothing has changed. Paused is not the same as ended, and reviewers read the difference.

04The other actions worth knowing about

Cloaking is the cannabis specific one, and it is almost always accidental. A well meaning developer builds the age gate server side and adds logic to let crawlers through, or serves them a simplified page. That is a different experience for Googlebot than for visitors, which is cloaking regardless of intent. The fix is to rebuild the gate as a client side overlay sitting on top of a fully rendered page, so the HTML never varies by who is asking.

Site reputation abuse is newer and growing quickly. If a third party publishes on your domain to borrow your authority, whether that is a rented subfolder, a coupon section run by an outside partner, or a syndicated content deal, that now carries its own action. Noindexing it is not a fix. It has to come off the domain.

User generated spam, hacked content, and structured data issues round out the list. Each has its own remediation, but they share one requirement: the fix has to include prevention, not just cleanup. Removing spam without adding moderation, or cleaning injected pages without closing the vulnerability, means the problem returns before the review even completes.

05What reviewers actually reward

A reviewer reads a lot of these. They are not looking for persuasion, and they are certainly not looking for length. They are looking for one thing: evidence that you understand what was wrong and have genuinely fixed it. Everything else is noise.

That means admitting the problem plainly rather than framing it as a misunderstanding. It means stating what you did in numbers, because thirty four pages inventoried and twenty two rewritten is checkable in a way that we improved our content is not. It means linking documentation the reviewer can actually open. And it means not arguing, not making excuses, and not blaming a previous agency or a competitor.

The tone that works is a professional owning a mistake and showing the fix. The tone that fails is a marketer defending their work. That distinction alone has flipped denials into approvals more often than any other single change.

06Why requests get denied

Five reasons account for the large majority. Submitting before the fixes are live, so the reviewer sees the problem still there. Fixing only the examples Google named while leaving the rest of the pattern untouched. Rewriting so lightly that pages are still recognisably the same. Removing pages without proper redirect or status code handling, creating a wall of errors. And writing a defensive request instead of an accountable one.

This builder blocks on the ones it can detect, and it does so deliberately. A denial is not a neutral outcome you can simply try again after. Each one makes the next review slower and more sceptical, and there is a point at which a reviewer stops reading carefully. Waiting a week to finish the work properly is almost always cheaper than spending a review cycle proving you were not ready.

What this produces is a starting point built on your answers. Read it through, put it in your own words where it does not sound like you, and never include anything you cannot evidence. Everything runs in your browser, so nothing you enter here is uploaded or stored.

Scans a page the way a search engine sees it. Finds embedded menus Google cannot read, age gates blocking the crawler, and…

Detects which menu platform you run and whether its products actually reach the HTML Google reads. If your menu loads in an…

Reads your sitemap, finds every location page, and measures how similar they are to each other. This is the scaled content risk…