Intricate Art Seminars

Intricate Art teaches dry needling and spinal manipulation to licensed clinicians across the country. The teaching was excellent. The marketing around it was a calendar: add dates, hope they fill, discount what does not. We replaced that with a system.

01 · The situation

A business that could not see itself.

Intricate Art had been running live seminars for years and had built real authority in a technical field. The founder holds a genuinely differentiated clinical position, autonomic nervous system-centered reasoning, that no competitor in the category occupies.

What the business did not have was a way to see itself.

Revenue had been flat for two consecutive years while order volume climbed. Courses were being added faster than they could fill. Discounting had become the default tool for closing the gap. Analytics reported roughly double the revenue the payment processor did, so every budget conversation started with an argument about whose number was right.

02 · What we did first

We did not launch anything for six weeks.

Instead we pulled four years of order history, the full customer file and search console exports, and built the first complete economic baseline the business had ever had. Not a dashboard. A set of answers to questions nobody had been able to ask.

• The ladder was one rung long. Of the clinicians who entered through the foundational course, roughly a third ever bought again, and two thirds of those took exactly one more course and stopped. There was no third rung being sold.
• The highest-value product was invisible. Manipulation produced the highest lifetime value and the highest repeat rate of any entry point in the catalog. A handful of the flagship course’s alumni had ever bought it. It was in the catalog, not in the marketing.
• The repurchase window was knowable. Median time to a second purchase came in at 110 days, with most of the volume inside six months. That single number turned email from a broadcast channel into a timed sequence.
• The most valuable search demand was being missed entirely. State board and regulatory queries, the highest commercial-intent cluster in the category and the thing a clinician searches the week they decide to get certified, were pulling five figures in impressions at an average position on page two. The clicks were going to competitors.
• The wrong content was winning. A secondary product line unrelated to the core business was generating several times the organic traffic of every course page combined, bringing in customers worth a fraction of a seminar buyer who never crossed over.

That analysis became the strategy. Everything after it is downstream of those findings.

03 · What we built

Six things, on top of the baseline.

The site ran across five separate surfaces: primary domain, a shop subdomain, hosted checkout, a course delivery platform, and a certified-clinician directory sitting on a raw hosting URL linked from the main navigation. Every hop was an attribution seam, a trust seam, and for two of them, a total loss of search equity.

We consolidated what could be consolidated, moved the directory onto the primary domain where it builds authority instead of leaking it, rebuilt the course page template around eligibility, outcomes, approvals and logistics, and stripped years of expired dates that were burying current availability.

We also fixed the infrastructure underneath. The site had been going down on email send days, because the email platform’s tracking parameters were bypassing the cache entirely, so every newsletter click hit the origin server uncached. We identified it, resolved it, and moved hosting to a stack with proper object caching.

The category’s highest-intent search behavior is regulatory: can I do this in my state, what does my board require, how many hours. We built the state and approvals layer to own it. Structured, sourced, dated, and maintained on a review cycle, because a wrong state page in a licensed profession is a liability rather than just a ranking miss.

Answer-engine visibility got the same treatment. Course pages, approval pages and comparison content are structured so that a language model answering what do I need to get dry needling certified in Ohio has something clean and attributable to cite.

Five platforms, each with a defined job. Google Search for brand protection, certification intent, course family, and city-and-date campaigns tied to real geography and real capacity. Microsoft, which was already producing consistent monthly sales at low spend, and which told us Google was underfunded before we spent a dollar proving it. Meta, not on cold interest targeting but on lookalikes seeded from actual buyers, geo-fenced event promotion and retargeting, with separate budgets per market so one city cannot eat another’s spend. X, testing into the clinical and sports-medicine conversation. And an early implementation on OpenAI’s ads platform, which gives the account a presence in a channel most of its competitors have not found yet.

Budget moves in 10 percent weekly increments rather than jumps, so the platforms never reset their learning phases. Scaling decisions run on marginal return, not blended.

The real constraint here is not return on ad spend. It is seats. A course has a hard minimum and a hard ceiling, so we set acquisition caps per seat off contribution margin, which makes the stop rules arithmetic rather than opinion.

Content serves the two motions the business actually runs: the individual clinician choosing a course, and the clinic owner deciding to host one. Course comparisons, state requirement resources, evidence-backed clinical explainers, and employer-facing material that helps a clinician get reimbursement approved.

Everything clinical goes through a claim review before it ships. In a regulated health category, the fastest way to lose an ad account is a sentence nobody checked.

YouTube first, because technique demonstration is visual, evergreen and searchable, and it is what practitioners look for. LinkedIn second, aimed at clinic owners and the hosting pipeline. Instagram and short-form as distribution for clips cut from the library.

One idea per post, one clinical decision shown, consent documented for anyone recognizable.

This is the part that outlasts a campaign.

We built the measurement layer the business runs on: reconciled analytics against the payment processor, fixed the duplicate transaction events that had been inflating reported revenue, rebuilt tag management across the split site-and-checkout architecture, and stood up per-course reporting with defined owners and a review cadence.

Alongside it: a claim library and pre-flight review process, launch playbooks with go and no-go gates so a course gets decided on rather than drifting toward a cancellation, a host qualification and economics kit for the B2B motion, and a weekly operating rhythm of seat pacing, launch decisions and post-course retro.

We also told them what not to do. A marketplace channel they had been considering prohibits standalone digital products, and the available workaround would have read as circumvention. That was a week of work saved and a policy problem avoided.

That is Marketing Miles Per Gallon applied to a business that records everything it does anyway.

04 · Results

What the first phase produced.

Performance numbers are being reconciled across the ad accounts, Search Console and the payment processor before they go on a page. What follows is documented in the account today.

The engine, not the campaign.

Intricate Art did not need more marketing. It needed to know which marketing was working, which courses were worth running, and which customer was worth what.

We built that first. Everything running now, the paid accounts, the content calendar, the email sequences, the site itself, sits on top of it, which is why it can be scaled without guessing.