Most SEO sales pages show you a chart that goes up and to the right and ask you to trust it. This one shows you the actual domains we run and the numbers behind them, pulled live from our recommendation tracker, Search Console, and PageSpeed Insights. If a figure below moves, it is because the underlying data moved, not because someone edited a slide.
The reason to read this is simple. You are comparing $200-a-month SEO providers and every one of them promises results. Few will tell you what a month of work produces, on which sites, over what timeframe, and where it stops working. Here is our version, with the honest limits included.
The method: eight analyses, then the work
The part that separates this from a monthly audit is not the analysis. It is that implementation follows the analysis in the same week, on the same site, by the same people. Reporting on a problem is easy. Shipping the fix and confirming it went live is the job.
Every Monday, eight AI analyses run against each active domain. Each one produces a score out of 100 and a list of specific, tagged recommendations:
- Technical SEO checks crawlability, indexation, redirects, canonicals, robots rules, and security headers.
- Image looks at alt text, file sizes, formats, lazy loading, and layout shift from unsized media.
- PageSpeed reads real Core Web Vitals and lab metrics, then points at what is actually slowing the render.
- Schema covers structured data: which types are present, which are broken, and which are missing for the page's purpose.
- Content grades depth, originality, and the E-E-A-T signals that both Google and AI engines weigh.
- GEO / AI scores how citable the page is to answer engines: passage structure, crawler access, and entity clarity.
- Internal links maps how authority flows through the site and where deep pages are stranded.
- Local (for location-based businesses) reviews Google Business Profile, NAP consistency, and review signals.
A single Monday scan surfaces dozens of recommendations per domain. That is far more than anyone can responsibly ship in a week, so a person reads them. Some are wrong because the model misread the page. Some are right but low value against everything else in the queue. What survives triage becomes the week's ship list, and the rest waits in a deferred queue instead of reappearing as a fake new finding next Monday. We wrote up the full Monday-to-Friday breakdown separately in the honest weekly SEO routine.
The report a customer receives on Friday is generated from completion records. If a change was not shipped and verified, it does not appear. There is no way to pad the report with work that did not happen, because the report is not written by hand.
What the numbers look like across four domains
We run domains in different categories on the same loop: a global crypto guide, a browser-extension SaaS, a local immigration firm, and a regional wildlife-removal company. The point of showing all four is that the method is the same but the results are not, because the starting condition and the business type are not.
| Domain | Weeks live | Changes shipped | Health |
|---|---|---|---|
| Hyperliquid Guide | 21 | 585 | A (95) |
| CarSnipe | 21 | 399 | A (94) |
| Immigration Law Center | 15 | 280 | B (81) |
| Dr. Critter | 11 | 168 | C (78) |
Hyperliquid Guide is the longest engagement and the clearest read. After 21 weeks and 585 implemented changes, its Health Score sits at a grade A, and Search Console records 47,227 organic clicks against 1,210,334 impressions since onboarding. That is a site where the content was already good and the technical and GEO layers were the constraint. Remove the constraint and the existing content earns its traffic.
CarSnipe, a SaaS product, has taken 399 changes over the same 21 weeks and holds a grade A. Immigration Law Center, a local firm, is at 280 changes over 15 weeks with 26 Google Business Profile posts published, because for a local business the profile is as much of the surface as the website. Dr. Critter is the newest of the four at 11 weeks, where the early work was crawl-health and Core Web Vitals repair before anything else could compound. Its organic clicks are up 10% over the trailing 30-day window.
The honest caveat on these numbers: the engagements are young. Search Console history is shallow on the newer domains, and a percentage lift off a small base can read as more dramatic than it is. We would rather show you absolute counts you can sanity-check than a headline growth rate that flatters a two-month-old baseline. The full case studies break each domain down with its own live tracker.
What GEO actually requires
Generative Engine Optimization gets described as a mystery. It is not. Getting cited by AI Overviews, ChatGPT, Perplexity, and the rest comes down to a handful of concrete things, and most of them are technical rather than magical. We cover the full picture on the GEO page, but here is the working checklist.
Crawler access
If the AI crawlers cannot fetch your pages, none of the rest matters. GPTBot, PerplexityBot, ClaudeBot, and Google-Extended each have to be allowed in robots.txt, and your server and CDN have to actually serve them a 200 rather than a challenge page. We have watched sites accidentally block the exact bots they wanted to reach.
Server-side rendering and citability
Answer engines lift short passages, not whole pages. A section that opens with a direct two-sentence answer gets quoted. A section that buries the answer 600 words down gets skipped. The content also has to exist in the served HTML, not appear after a client-side JavaScript render, because many crawlers do not wait for it.
Schema and entity clarity
Structured data lets an engine resolve what your page is about and trust the facts on it. Article, Organization, and LocalBusiness markup, with sameAs pointing to canonical profiles, does more for AI grounding than it does for a traditional rich result. We go deeper on this in schema markup that still matters.
Snippet and freshness signals
A permissive max-snippet directive lets engines quote enough of your text to actually cite it. Consistent published and modified dates keep a page from being discounted as stale. An llms.txt file gives AI systems a clean map of what you would like them to read first. None of these is a silver bullet. Together they remove the reasons an engine passes you over.
The relationship between GEO and traditional SEO is not a rivalry, and treating it as one wastes money. The foundations overlap almost entirely, and the same well-structured page can win on both surfaces. We laid out where they split in GEO vs SEO: what changed in 2026.
The honest limits
Weekly SEO compounds, but it is not fast and it is not everything. Here is what it does not do, so you can decide with the real picture.
It does not move rankings in a week. The gap between shipping a fix and seeing it in Search Console is measured in weeks to months, because Google has to re-crawl, re-index, and re-rank, and Core Web Vitals field data runs on a rolling 28-day window. A domain that starts with a technical mess spends its first few weeks on repair before anything can compound. Dr. Critter is a live example of exactly that sequence.
It does not fix a weak business or thin product. SEO brings qualified people to your page. If the page does not answer their question or the offer does not hold up, better distribution just surfaces that faster. We would rather tell you the site needs work than take the retainer and blame the algorithm.
It does not replace the big projects. A site migration, an information-architecture rebuild, or a large content program does not fit in a three-day ship window. Those get planned across several weekly cycles with a target each week, and they run alongside the maintenance loop rather than pausing it.
And AI-citation traffic, in raw click volume, is still a fraction of organic search for most sites today. It is climbing, and the brand-mention effect of being cited tends to lead the click volume by months, but we are not going to pretend it is your main channel yet. It is a second surface worth owning early, not a replacement for the first.
Where this leaves you
If your current provider cannot answer "what shipped to production last week" with a list and live URLs, the cadence is not weekly regardless of what the contract says. That single question is the fastest audit you can run on any SEO vendor, including us.
Everything above runs on one flat plan. No tier games, no per-change billing, and you can cancel any month. If you want to see the same treatment on your own site, the pricing page lays out exactly what the weekly loop covers and what it costs.