Case study
Wild Sky Paragliding — making a Boulder tandem operator legible to AI search
Adventure & tourism · Boulder, Colorado · GEO audit and content build — ongoing
The situation
Disclosure first: Wild Sky Paragliding is run by Joshua Groth, the founder and CEO of NovaSapien Labs. A case study about a business our own founder owns is not an independent reference and should not be read as one. What it buys you is access — we can publish the before and after in a detail a normal client would never let us show.
Wild Sky Paragliding runs tandem flights above Boulder. By April 2026 the site was four pages on WordPress with a working Rezdy booking flow already live — the business was ready to take bookings.
It was not, however, visible. We ran the primary keyword clusters and Wild Sky did not appear in any of them. Boulder Free Flight, Red Tail, Many Feathers and the OTAs — TripAdvisor, Yelp, Viator — owned every position we looked at.
It was also absent from all ten-plus of the directory and OTA listings that dominated the results we pulled for Boulder paragliding. We checked which pages ranked and which businesses were named; we did not record which URLs the engines cited, so read that as where the category's answers were coming from, not as a citation log.
What we measured
A full audit on 30 April 2026: the live HTML of every indexable page, schema validation, SERP analysis across five keyword clusters, six direct competitors, ten-plus citation directories, and a Reddit and forum scan for how people actually ask about tandem flights.
What changed
Each pair below is a dated observation from the audit and the state of the live site when we re-checked it on September 9, 2026. Both halves are checkable.
The WordPress placeholder
Before: /hello-world/ — the default 'Hello world!' post — was still live, indexed, and linked from the footer of every page.
Now: Returns a 301 redirect. Verified live 9 September 2026.
FAQ markup
Before: The FAQ page was deployed with the content live but no FAQPage schema, so the question-and-answer structure was visible to people and undeclared to parsers. We called it the highest-leverage bug in the audit at the time and we no longer would: Google retired FAQ rich results for every site on 7 May 2026, and no answer engine has stated that it reads FAQPage when deciding who to quote.
Now: FAQPage schema is live on /faqs/. Verified live 9 September 2026.
Entity and offer markup
Before: The About page carried Joshua Groth's story but no Person schema, so the operator behind the business was invisible to engines.
Now: The homepage now carries LocalBusiness, Service, Offer, Person, AggregateRating, Review, BreadcrumbList and City nodes. Verified live 9 September 2026.
Stale booking copy
Before: The homepage still read that bookings would be 'available later this spring' while the Rezdy flow was already taking them — a staleness signal to crawlers and a conversion leak with humans.
Now: The line is gone. Verified live 9 September 2026.
What the audit found that mattered most
- Two of the competing paragliding operators the audit examined had FAQPage schema deployed; the others did not. Wild Sky already had the FAQ content written, so declaring the structure was a one-day fix. An earlier version of this page said none of the competitors had the markup, which the audit file does not support. We are not claiming that markup wins citations.
- An active brand collision with WildSky.co.za, an established South African operator, was splitting the entity: three of our search results returned the South African business rather than the Colorado one. Entity disambiguation, not more content, was the constraint.
- Five listings carried the category across the results we pulled — TripAdvisor, Yelp, the Boulder CVB, Uncover Colorado and RMHPA — and Wild Sky was on none of them. Directory presence, not page count, was the gap. Whether closing it changes what an engine says is a measurement we have not completed.
What we are not claiming
We are not publishing traffic, booking or revenue figures. The engagement is ongoing and we do not have a clean, attributable before-and-after on bookings — so we are not going to imply one.
The fixes above are verifiable on the live site. Whether they have yet moved Wild Sky into AI answers for its primary queries is a separate measurement we have not completed.
Questions
Wild Sky is run by your founder. Why publish it as a case study?
Because we would rather disclose it than have you discover it on an author page. Joshua Groth is the founder and CEO of NovaSapien Labs and he runs Wild Sky Paragliding, so this is our own operator business and not an independent client reference. What it gives you is a case study with nothing withheld: every before is dated from the audit file, every after was re-verified on the live site on 9 September 2026, and where we have not measured an outcome we say so.
Why does an FAQ page need schema if the content is already there?
Less than we used to say. FAQPage markup labels each question and its answer as a discrete pair, which makes the structure explicit to a parser rather than inferred. What it does not do is earn citations: Google retired FAQ rich results for every site on 7 May 2026 and removed the documentation that June, and no answer engine has stated that it reads FAQPage when choosing what to quote. The part that still matters is the visible content — questions phrased the way customers ask them, answered directly underneath.
What is a brand collision, and why does it matter for AI search?
It is when two unrelated businesses share a name closely enough that engines merge them into one entity. Wild Sky Paragliding in Boulder and WildSky in South Africa were being blended, so searches for the Colorado business returned the South African one. Answer engines resolve entities before they recommend, so an unresolved collision means recommendations leak to the other business. The fix is disambiguation work — consistent naming, location signals and linked profiles — not more content.
Do you only work with adventure and tourism businesses?
No. Wild Sky is an adventure operator, but the pattern is not vertical-specific: across 70 Front Range businesses in seven batches — breweries, restaurants, fitness studios, salons and spas, and law offices — half of the 66 we could sell to graded Invisible, Partial or Compromised, and the weak share ran between 40% and 56% in every one of the five verticals.