A per-post marketing contract buys the pattern Google's spam updates are built to remove

For most businesses the exposure in a Google spam update is not one bad page. It is a publishing archive shaped by an invoice, and the only schedule that holds up is one where someone responsible for the brand decides whether there is anything real to say.

· 7 minute read

Google began rolling out its September 2026 spam update on September 24, according to its Search Status Dashboard, 37 days after the August update started. Two updates are not a schedule, and Google has announced none. What each one puts to a business that buys marketing by the post is a single question: does each page exist because someone had something to say, or because a contract said there would be twelve of them this month?

Google enforces its spam policies two ways, and only one of them tells you

Google enforces its spam policies through its automated spam detection systems, SpamBrain among them, which spam updates such as the September 2026 rollout improve, and through manual actions, where a human reviewer applies the penalty and the site owner is notified in Search Console. Scaled content abuse can be enforced either way. Google's spam policies state the cost of a violation plainly: sites "may rank lower in results or not appear in results at all."

The two routes end differently. A manual action names the problem, and the owner fixes it and files a reconsideration request. An algorithmic demotion names nothing. Google's spam updates documentation says improvement may come "over a period of months" once its systems learn a site complies; that describes recovery in general, not a promised path out of this rollout.

Scaled content abuse is judged by why a page exists, not how it was written

Scaled content abuse is the spam policy most business websites should read, and Google defines it as "when many pages are generated for the primary purpose of manipulating search rankings and not helping users," judged "no matter how it's created." The policy is not about AI. A person writing to a quota produces the same pattern a model does.

Google's helpful content guidance turns the idea into questions a business can ask of itself. Among them: "Are you producing lots of content on many different topics in hopes that some of it might perform well in search results?" It also warns against writing to a word count and against "changing the date of pages to make them seem fresh when the content has not substantially changed."

None of those questions asks whether the writing is good. They ask why the page exists, and a site's publishing history answers that more honestly than any single page.

A per-post contract decides content will exist before anyone decides there is something to say

A per-post marketing contract fixes the output before the month happens. Four posts ship in the month the business launched a product and four ship in the month nothing happened. In the quiet month the cheapest way to fill the slots is to cover adjacent topics, restate last quarter's piece, or refresh dates on old pages, and Google's helpful content questions name the first and the last of those moves.

No single post in that retainer looks like spam. The pattern exists only in aggregate: many pages, loosely related, published on a rhythm that tracks the invoice rather than the business. An automated system reads the aggregate. The person approving the retainer reads one post at a time.

No public Google case ties a per-post contract to a demotion, and this essay does not claim one. The claim is narrower and checkable: a quota guarantees quiet months, and quiet months under a quota produce the specific behaviors Google's own guidance lists. Whether a given archive crosses into enforcement depends on volume and proportion, and Google publishes no threshold for either.

Google's guidance credits publishing frequency with nothing, which makes a quota exposure with no return

Nothing in Google's spam policies or helpful content guidance rewards publishing more often. A skeptic reads that as proof cadence cannot matter to a spam system either.

For a per-post contract it cuts the other way. If frequency earns no ranking credit, the volume a retainer guarantees buys nothing in rankings. What volume does produce is more pages, and more pages is the one input the scaled content abuse policy names. Cadence is harmless as a ceiling and exposed as a quota.

A schedule that holds up works as a ceiling. Weekly means at most weekly, and a held week is an outcome, not a missed deliverable.

Only someone who answers for the brand can decide a month has nothing to say

The decision to publish nothing has to sit with someone responsible for the brand, because every other instrument is set before the month happens. A better agency brief is written once and cannot know that March was quiet. An editorial calendar schedules topics in advance, which is a quota with subject lines attached. A flat monthly retainer removes the per-post price and leaves the incentive: a vendor judged on visible output will not hand in an empty month and call it the work.

The person who can hold a month is the one who answers for the brand rather than for the deliverable. That person sees what actually happened this month and carries the cost if the archive drifts. The test for any arrangement is concrete: can this contract survive a month of zero without a dispute? If the answer is no, the contract is a quota, whatever it is called.

Is a small business on four posts a month anywhere near scaled content abuse?

The strongest case against this argument is scale. Google's scaled content abuse policy describes "many pages" generated for rankings, and a small business receiving four to twelve posts a month is not a content farm. On that reading the policy targets programmatic page generators and high volume networks, and a modest retainer is too small to register. A related case holds that without a calendar most businesses publish nothing at all.

Much of the scale objection is right, and the likelihood of enforcement for a small site on a modest retainer is probably low. Google publishes no page threshold, so any line is inference, and the plausible line is proportion and drift rather than raw count. As an illustration, not a case: a 30 page service site adding 12 contract posts a month holds more contract pages than original ones within three months, and its quiet months push those pages toward topics the business does not serve. At that point a small site resembles the pattern at its own scale.

Low probability does not rescue the quota, because the quota buys nothing in rankings on the other side. If frequency earns no ranking credit, a held quiet week costs close to nothing, and any real chance of months of demotion outweighs close to nothing.

The calendar objection is answered by changing what the calendar schedules. A weekly review of whether there is something real to say is a good discipline. A weekly obligation to publish whatever that review found, including nothing, is the pattern Google describes.

Audit the publishing process during a spam update, not the rankings

A business watching the September 2026 spam update roll out learns more from its publishing process than from its rankings, because the process can be inspected today and a ranking move during a rollout cannot be attributed to one cause.

First, open the Manual actions report in Search Console; if it is empty, any exposure is algorithmic and nobody will say so. Second, list every page published in the last twelve months and write beside each one the business event or customer question that caused it; if the blank lines outnumber the reasons, the archive tells the contract's story rather than the business's. Third, read the marketing contract and note whether it is priced per post, per word, or as a flat fee with an expected volume. Fourth, check whether anyone updates dates on pages that have not substantially changed. Fifth, look for pages that answer the same question in slightly different words. Sixth, name the person who can decide that this month nothing gets published, and read whether the contract lets them.

Cadence is harmless as a ceiling and exposed as a quota.

The sixth check is the one the others depend on. A business that finds exposure in the first five and has nobody with the authority to hold a month has not found a fix. It has found a description of what next month's invoice will produce.

Further reading

  1. Search Engine Journal, report on the September 2026 spam update rollout, dates and 2026 update history searchenginejournal.com/google-september-2026-spam-update
  2. Search Engine Journal, report on the August 2026 spam update and its completion searchenginejournal.com/google-begins-rolling-out-the-august-2026-spam-update
  3. Google Search Central, spam updates documentation, consequences and recovery timeframe developers.google.com/search/docs/appearance/spam-updates
  4. Google Search Central, creating helpful, reliable, people-first content, the search engine first questions developers.google.com/search/docs/fundamentals/creating-helpful-content

Every source above was fetched and a verbatim phrase confirmed on the page before this essay published. Nothing here is paraphrased from memory.