The 'Medic' Update: Google Starts Grading Credibility

Google confirmed a broad core algorithm update rolling out from August 1, and offered its standard non-guidance. No specific fixes. Nothing is ‘wrong’ with the sites that moved.

The industry’s tracking data promptly did something core updates rarely do so clearly. It clustered. Health, medical, finance, and wellness sites absorbed movement so lopsided that the industry named the update ‘Medic’ within the week.

The loss reports assembled a recognizable profile. Sites publishing high-stakes advice with anonymous authors, unsourced claims, and ad-heavy layouts: hit hard. What to take, what to invest, what to do about the symptom. Credible institutional publishers in the same fields: largely holding or gaining.

our protocol for update coverage, verify your own impact first, per the Fred-era protocol, stands as always. But this update earns more than protocol, because the clustering points somewhere specific. At the framework Google has published openly for years in its Search Quality Rater Guidelines, and at the moment that framework visibly became rankings. The guidelines’ vocabulary is about to take over the industry’s, so let us install it precisely, and then do what we actually get paid for. Translate it into content operations our institutional readers can run.

E-A-T and YMYL, installed precisely

Two acronyms from the rater guidelines, defined from the source rather than the webinar stage.

E-A-T: expertise, authoritativeness, trustworthiness. The qualities Google’s human quality raters are told to assess when judging whether pages deserve their rankings. Does the content show real knowledge? Does the author and site carry recognized standing in the topic? Can the operation be trusted: accuracy, transparency, honest purpose?

YMYL: ‘Your Money or Your Life.’ The guidelines’ category for content that can affect a person’s health, money, safety, or wellbeing. There, the raters are told to hold E-A-T to its highest bar, because the cost of bad information is highest.

The mechanics deserve one clarifying sentence, because the industry chronically garbles it. The raters do not directly adjust rankings. Their grades train and check the systems that do. The guidelines are best read as Google publishing its intended destination in plain language, with the algorithms following at whatever pace the machine-learning progress we track permits.

August’s clustering is exactly what that following looks like when it takes a visible step. Content whose consequences are highest, graded hardest, on the credibility measures the guidelines have described all along. For the healthcare systems, financial firms, and universities we serve (YMYL publishers, every one) that is not a threat. It is the algorithm finally moving toward ground they already hold. If, the operative if, their sites make it legible. Show your experts. If a real person with real credentials stands behind the page, say so on the page.

Translating raters’ criteria into content operations

• Visible authors with real credentials, on high-stakes content. The clinical page reviewed by the named physician with the linked profile. The financial guidance carrying its advisor’s credentials. The research explainer with its faculty author. The habit of publishing as ‘staff’ gives up precisely the asset the update just priced, and the fix is editorial policy, not technology.

• Citations to real sources. Claims sourced, sources linked, per the pattern every credible profile in August’s data shares. Unsourced assertion is the content-farm tell, and high-stakes content that reads like it will be graded like it.

• Editorial review that exists and is described. The review process running in fact and stated in public. Who checks clinical accuracy, how often, against what standards. The ‘medically reviewed by’ line. The editorial-standards page. The dated review stamps we already prescribe for accuracy reasons.

• Institutional trust signals, unglamorous and essential. Contact transparency. Real about pages that say something. Credentials stated. Policies findable. The competence signals the trust research has always stressed, now doubling as rankings infrastructure.

• The content audit, YMYL-weighted. The site’s high-stakes content inventoried and held against the list above. The thin, dated, and orphaned tier (the 2013 health tips, the stale financial guidance) gets updated or retired per the quality-over-volume rule. Core updates grade the whole body of work, and the site’s worst high-stakes page is now part of its argument.

The honest reading of ‘nothing to fix’

Google’s own advice on core updates deserves an honest paraphrase, because it is both true and chronically misread. ‘Nothing to fix’ does not mean nothing to do. It means the update is not a penalty with a specific remedy but a re-grading of what deserves to win. The useful response is not forensic tweaking but becoming, visibly, the kind of source the re-grading favors.

That reading converts August from anxiety into strategy for exactly our readership. The sentence we keep repeating in client meetings this month belongs in the post. Institutional credibility is precisely what our clients have that content farms do not. The work is making it machine-legible, and it was worth doing before the algorithm asked.

The hospital’s clinicians. The university’s faculty. The bank’s fiduciary duties. These are real E-A-T, built over decades. August’s losers were largely sites performing expertise they did not possess.

The unfairness cases, credible small publishers caught in the update’s breadth, and there are real ones in the reports, get our sympathy and the same advice. The legibility work above is also their best available response. Google has signaled core updates will now recur regularly. Recalibrations will keep arriving, with recoveries possible at each pass for sites that did the work between them.

Where quality enforcement goes next

The forward note, because August opens a question we will follow for years.

Search’s quality systems are visibly moving up the ladder. From counting links, the PageRank era. To grading content, the Panda era. To judging sources, this month. Asking not just ‘is this page good’ but ‘who says so, and why should anyone believe them?’

That path has obvious next stops. Author-level signals. Cross-site reputation. The entity understanding the knowledge-graph coverage tracks. And it converges, from a different direction, on the same destination as the machine-mediated surfaces file. A web where being a trusted source, in machine-readable form, is the decisive asset. Whether the machine is ranking pages, pulling snippets, or speaking answers aloud.

Our readers’ institutions spent decades building exactly that trust in the physical world. The strategic project of the next several years, named plainly by an August update, is finishing its translation to the digital one. Named experts. Sourced claims. Described standards. Transparent operations.

Grade your own high-stakes content the way a skeptical rater would, this quarter. The algorithm is now doing it either way, and it will be back.