Get it
fixed.
Three ways in, all priced on this page. You'll know exactly what you're getting before you pay, which shouldn't be remarkable and somehow is.
The Launch Roast
A 15-minute recorded video teardown of your landing page, app store listing, or launch - back within 48 hours. Blunt, specific, and scored against the rubric below so you can see the reasoning rather than just the verdict.
What you send
- The URL (or store listing), and one sentence on who it's for.
- What you think the problem is - so I can tell you if you're right, which is roughly half the value.
The rubric, in full
| # | What gets checked | What a pass looks like |
|---|---|---|
| 01 | The four-second test | A stranger can say what it is and who it's for after four seconds |
| 02 | Competitive alternative | The page names what you replace - the spreadsheet, the agency, the four tools |
| 03 | Specificity of the promise | An outcome with an edge, not "streamline your workflow" |
| 04 | Proof placement | Evidence appears before the ask, not in a testimonial slab at the bottom |
| 05 | One primary action | A single obvious next step, repeated - not five competing buttons |
| 06 | Friction in that step | Nothing asked for that isn't needed to deliver value |
| 07 | Machine readability | Content in the raw HTML, title and description doing real work, crawlable |
| 08 | Search and AI surface | The page answers a query someone actually types or asks a model |
| 09 | Mobile reality | The promise and the action both visible without scrolling on a phone |
| 10 | Measurement | The one action is instrumented, so the next change can be judged |
What you get back: the video, a written list of the three things to change first in priority order, and an honest note on whether your problem is even on this page - sometimes the page is fine and the problem is upstream. I'd rather tell you that than sell you a fix you don't need.
The GEO Audit
Does ChatGPT recommend you? For a growing share of buyers, "best tool for X" is now a question asked of a model, not a search engine - and the same handful of names comes back every time. This audit establishes whether you're in that set, why not, and what changes it.
What's included
- Citation baseline. Your category's real buying prompts, run across the major assistants, recorded with dates. You get the raw results, not a score out of a hundred.
- Share of voice. Who does get named for your prompts, and what those pages have that yours doesn't.
- Entity and extractability fixes. Whether models can tell what you are, what you do, and who you're for - and the specific structural fixes that change it.
- Information-gain content plan. The pages that would earn citations rather than compete for clicks. First-party data, comparisons, direct answers - the formats models actually quote.
- A 90-day recheck. Same prompts, same method, dated. Because an audit that never gets graded is just an opinion in a PDF.
Agencies doing this work start around $1,500 a month, and the monitoring platforms are built for enterprises. Meanwhile the discipline is young enough that models still reward topical depth over domain authority - which means a small site can win citations in weeks, not the years classic SEO authority takes. That window is open now. It won't be in two years.
The rubric is published in full. Every check, how it's tested, and what a pass looks like: read the GEO audit rubric. Most of it is genuinely self-serviceable in an afternoon - what you're paying for is someone running the citation baseline properly and telling you which three things to fix first.
The "Now What" Roadmap Sprint
Two weeks, working through the whole thing: what you've built, who it's for, why nobody's finding it, and precisely what to do for the next 90 days - in order, with numbers attached.
What happens
- Week one - diagnosis. Positioning against real alternatives, the search and AI landscape you're actually competing in, your funnel's leak points, and what you can't currently see because it isn't instrumented.
- Week two - the roadmap. A sequenced 90-day plan: which channel first and why, the content and pages that earn compounding traffic, the measurement setup, and the specific claims we expect to be true by day 30, 60 and 90.
- The claims get graded. Every prediction in your roadmap has a metric, a threshold and a date. You'll know whether the plan worked, which is more than most strategy documents ever tell you.
Two per month, maximum - it's two weeks of real work, not a deck. Audit or roast graduates first; the audit fee comes off the price.
Ongoing
The audits are a photograph. This is the film: your Search Console and your first-party analytics collected every morning, twenty-one rules run against them, and every change we make graded on a date decided before it was made.
Three tiers, and they are permissions rather than packages
Each one is a setting on your site in the console, enforced in code rather than promised in a contract. The tier below cannot do the thing the tier above does, whatever anybody types.
- Watch - $149/mo. A login to the console. Daily collection from Search Console and the first-party tag, what you rank for and where, which pages earn it, the queue of what the rules flagged, and the change log with its verdicts. Nothing is touched.
- Propose - $499/mo. Everything above, and the queue gets worked. Findings investigated against your own repo and docs, briefs written for the gaps, changes drafted with a hypothesis and a measurement date attached. You decide what happens to each one.
- Ship - $1,500/mo. Everything above, and the changes arrive as pull requests on your repository. One lever per pull request, so the measurement means something. Nothing is ever pushed to your default branch, which makes every change reviewable before it is live and revertible after.
What you get that is unusual
- A verdict, not a report. Every change carries a number and a date. On that date it is graded improved, no-change, worse or inconclusive, from your own Search Console data. "No-change" is a common answer and you will be told it.
- The reasoning, in writing. Every queued finding gets a recorded decision - done, dismissed or snoozed - with why. A month where nothing was worth doing says so, rather than inventing work to look busy.
- Levers that get retired. A change that did not work is recorded as settled, and the same lever does not get re-run on your site six weeks later by somebody who has forgotten.
What it is not
- Not a ranking guarantee. Nobody can sell you one, and the ones who imply it are describing a hope with a price attached.
- Not content production. Briefs are written; the pages are yours to write, or a separate conversation.
- Not a dashboard you have to check. The queue is the product; the console is where you go to disagree with it.
You have no traffic. Under roughly a hundred impressions a day there is nothing for the rules to read and nothing for a change to move - every verdict would come back inconclusive and you would be paying monthly for that sentence. Get found first: the free URL check and the 30-day playbook are the whole of what is useful at that stage, and they cost nothing. Come back when Search Console has something to say.
Two grants make it work and both are yours to give and withdraw: our service account added as a restricted user on your Search Console property, and one script tag on the site. Ship additionally needs a repository we have confirmed - never one matched by name - and consent recorded against the engagement rather than inferred from having the access.
Book it
Everything above, priced. The roast and the audit are instant pay because there is nothing to scope first - you send a URL and get work back. The sprint is two weeks of work, so it gets a conversation before an invoice.
When not to buy any of this
Genuinely: if you're not measuring anything yet, don't buy anything. Install analytics, verify Search Console, define one activation event, and wait two weeks. That's a two-hour job you can do yourself today, and doing it first makes everything I'd charge you for meaningfully better.
Same if you launched three days ago and the silence is new - that's not a failure signal yet, it's just week one. Read the receipts to see what normal actually looks like before you conclude anything is broken.
And if what you need is someone to execute for months rather than diagnose and hand over a plan, I'm the wrong shape - that's an agency or a fractional hire, and I'll say so rather than take the money.