Insights
No filler. Just useful answers about search and AI visibility.
Tests, research, and working models designed to help you make a better decision, not keep you scrolling.
GetBorvis / Organic visibilityEach guide answers one specific question, shows where the evidence came from, and explains what is worth doing next. No recycled listicles. No invented ranking secrets.
Our site audits keep finding the same nine problems. Here is the fix for each.
Every audit had a broken link where buyers click and a claim with nothing behind it. Whether a crawler catches each problem, the fix, and the same checks run on our own site.
Read the guide ↗The most common problem in our audits was text nobody re-read.
Another company’s name in a title, theme instructions in published copy, one alt text on every page. Five kinds, the fix for each, and why our own search found 19 false alarms.
Read the guide ↗Every site we audited had a broken link where buyers click.
An upgrade button to a 404, a WhatsApp link without a country code, a mislabelled booking button. Six fixes, and what our own link checker got wrong.
Read the guide ↗Two of our ten top search queries are full sentences. One states a job title.
The average US Google query is about 3.5 words. Two of ours are 17 and 23. We publish the actual Search Console rows, zero clicks and page-eight positions included.
Read the guide ↗How to spot AI tracking tools in your Search Console.
Long persona questions that recur daily at one impression each and never click. Our own log, six signals, and the regex filters that separate tools from people.
Read the guide ↗What platforms measure whether AI cites your content? Six compared.
Otterly, Peec, Profound, Semrush, Ahrefs Brand Radar and Signal AI, checked at source. What each counts, which engines it watches, and when a manual audit wins.
Read the guide ↗Rebranded or acquired a business? What to do with the old website.
Our own relaunch, 27 redirect rules, old pages still in Google’s AI report weeks later, and the two rules we got wrong.
Read the guide ↗We recalculated our own 49-point map grid. The headline average was hiding two answers.
Our published case study reported an average position of 12.4. That number codes every not-found result as 21. The mean of what was actually observed is 11.23, and neither should replace a coverage map.
Read the guide ↗Why case studies rank for the client’s name and not the service.
We audited our own two case studies first and found zero contextual links to the work they demonstrate. Four query classes, a six-field worksheet, and the reasons a client-name ranking can be fine.
Read the guide ↗Why Google and AI answers still show your old menu prices.
A guide, not a study. A delivery price and a dine-in price can both be right, so comparability comes before correction. Seven fields a price needs, and who controls which source.
Read the guide ↗Can you measure whether your earned media is what AI models cite about you?
Not with the attribution chain being sold. A tracer fact gets closer, and election research shows accuracy and inclusion are different failures.
Read the guide ↗If a claim sounds certain, the evidence should be visible.
Platform documentation is treated as fact. Vendor research is labelled as research. Observed patterns are presented as interpretation, not dressed up as secret ranking factors.
Why GetBorvis works this way ↗