by /
My Google Knowledge Panel disappeared without warning. My entity in Google’s Knowledge Graph had split into scattered, conflicting fragments with really low confidence scores, 24 in the case of the one that should have been me, down from over 200. Nothing on my end had changed; I was publishing more consistently than ever. The repair took one focused pass over my Wikidata, and the score is back at 208 with the panel live again. Here is the exact process, because we have since run it for clients and seen the same kind of results.
Dennis Yu wrote the definitive guide on how to fix your Wikidata, which is the article that triggered this repair. It’s the same process behind our Trenton Sandler Wikidata build and Brooke Lance’s Wikidata and schema. If the scoring concept is new to you, start with what the confidence score means.
What a panel collapse looks like
A Knowledge Panel rides on the Knowledge Graph’s confidence that one consolidated entity exists. When that confidence drops, the panel does not shrink gracefully. It fragments.
“I was down to absolutely nothing. It was just this. There was a bunch of conflicting entities. Mine had split into all these other ones with really low confidence scores.”
The confusing part was the timing. I had been repurposing videos into articles and doing everything right, and the panel still vanished. That’s the lesson for anyone running entity work: content volume cannot compensate for a weak or fragmented entity record.
The repair: every claim, referenced
Wikidata is a giant database of entities and facts that Google and other search engines pull from when they build Knowledge Panels. Every entity has its own Q number, and every fact is stored as a subject, a property, and a value. My entity is Q138557395.
The pass itself is systematic. If the person has no entity, create one; I had to create mine originally, and most of our clients start from zero too. If one exists, go statement by statement: confirm each claim points to the right thing, attach a reference to back it up, and fill in every missing property you can support. On my entity, the Minnesota Dunk Squad is listed as a sports team I created, with the claim referencing the team’s own website. Occupations, birthday, languages, location: everything supportable goes in, and everything gets a receipt.
Person schema makes the site part of the same entity
The second half of the fix lives on the personal brand site. Person schema with a sameAs property points to the Wikidata URL, LinkedIn, Instagram, and the rest of the profiles, so Google reads the website, the Wikidata entry, and the social accounts as one identity. Claude applied it through Rank Math on my site, and any schema method works as long as the sameAs list is right. Claude can handle essentially all of this work, the Wikidata pass included.
The result isolates the variable
I changed nothing else, which is what makes this case study useful. Wikidata plus schema took the entity from fragments scoring 24 back to a consolidated 208, near my previous peak of around 250, with the panel showing the overview, age, and social profiles again. We’ve implemented the same thing for a lot of our clients at Local Service Spotlight, giving them Wikidata and applying it as schema on their websites, and seen some very good improvements.
Run it on yourself
Find your entity first. The Knowledge Graph Explorer we built searches Google’s Knowledge Graph and returns the entities on a name, each with its confidence score and its KGMID.
“Just an identifier, kinda like your Social Security number on Google, as we like to say.”
From there, the prompt I give away in the video is the whole SOP: check whether I have a Wikidata entity, optimize it if I do, create one if I don’t, then apply it as schema on my personal brand site. Hand that sentence to Claude along with Dennis’s article and let it work.
This repair is one half of entity authority. The other half is volume: the Claude pipeline repurposing my 260+ videos keeps corroborating the entity that the Wikidata work consolidated. My first-person account of the scare is on my site, along with the repurposing system walkthrough. If your panel has vanished, or never appeared at all, check the graph before you change anything else.
Dylan Haugen
Dylan Haugen is a professional dunker, content creator, and editor at the Content Factory, where he transforms podcasts and interviews into strategic brand assets. He collaborates with Dennis Yu to support young entrepreneurs and business owners in building their personal brands through education, transparency, and effective content marketing. As the host of the Dunk Talk podcast and a dedicated advocate for establishing dunking as a recognized sport, Dylan combines athletic expertise, storytelling, and digital strategy to help elevate the next generation of creators.