What You Get
A way to build client sites, and a way to prove they worked.
Two halves arrive together. One builds the site. The other watches what the site does after it is live, so your next change is a decision instead of a guess.
The half most platforms skip
Building the site is the easy half
Any modern tool can put a page on the internet. Almost none of them tell you whether that page did anything afterward.
So you end up guessing. The client asks why traffic is flat and you open three browser tabs, squint at two dashboards, and give an answer you are not sure about. That is not a reporting problem. It is a measurement problem, and it is the part that decides whether you keep the account.
A website that cannot tell you what happened to it is a brochure. The measurement half is what turns it into something you can improve on purpose.
The inventory
Five things land in your shop
The site engine
Plain PHP and JSON files. One front controller, one settings folder, one file per page of content. No database to back up and no plugin updates to babysit at eleven at night.
The content rules
Articles, questions, glossary terms, case studies and directory listings each have a fixed shape. Fill in the shape and the page routes itself, builds its own structured data and lands in the sitemap.
The data layer
Search results, server logs and crawler visits collect in one warehouse you control. You can ask it a question about any site you run and get a straight answer with a date on it.
The nightly checks
Sitemaps rebuild, structured data is checked on every live page, and anything broken is written to a log that fails loudly. You find out from the job, not from the client.
The written standards
The build rules live in a file that travels with the system. A new site starts from the current rules instead of from whatever you remembered that morning.
Running cost
What the work looks like once it is in place
Here is the honest version. The first site takes real setup. You are installing a system, not opening an account, and that is the trade you are making on purpose.
After that the shape of the work changes. A new client site starts from the same scaffold as the last one. The checks that caught a problem on site three catch it on site thirty without anyone remembering to look.
| Job | First site | Tenth site |
|---|---|---|
| Deciding the structure | Real work. You are setting the rules. | Already decided. Copy and fill in. |
| Structured data | Wired once, in the system. | Generated. Nothing to add by hand. |
| Sitemap and endpoints | Wired once, in the system. | Rebuilt nightly by the job. |
| Finding what broke | You look. | The nightly check tells you. |
| Monthly client report | You build the query. | The query already exists. |
That is the whole argument for owning a system instead of renting one. The setup cost is paid once. The savings arrive on every job after it.
Being straight with you
What this does not do
Three things, said plainly, because finding them out later is worse.
- It does not make rankings appear. Clean structure and machine readable files help a site get read correctly. They are not a ranking button, and anyone who tells you otherwise is selling something.
- It does not run itself. AI does a lot of the typing. The decisions, the client calls and the judgment are still yours. Controlled and repeatable is the promise. Effortless is not.
- Log collection needs server access. If a client insists on hosting somewhere you cannot reach, you get the search half of the data and not the server half. Worth knowing before you promise a report.
Next step
Bring a project and we will look at it together
A walkthrough is a screen share of a live site and its data. You get to see the real thing working before any money changes hands.