What CCTV buyers search before they choose a camera system

Half the demand in this niche is arithmetic. Storage is bitrate times channels times retention. A PoE budget is per-camera draw plus cable loss against the recorder's total watts, which is why an eight-port NVR rated at 80W will not carry eight cameras that each pull 15W with the IR on. Cable runs meet the 100-metre Ethernet ceiling before a switch or an extender enters the plan.

Generic content gets these approximately right and then hedges. That is the failure mode: a reader who wanted a number leaves with a paragraph, and buys where the number was.

Where your data exists, the engine writes the working from the actual figures on the datasheets for the cameras and recorders you stock, so the example matches something a customer can put in a basket. Numbers are also where a writing model is most likely to slip, which is why every draft is fact-checked by a model from a different vendor than the writer: Claude writes, Gemini and GPT check. See cross-model validation. A failed draft goes back with the specific fix, up to three rounds, and is rejected if it still fails.

Ask most CCTV sites whether a camera may overlook a neighbour's garden and you get either silence or a confident answer that is wrong. The real position has moving parts: the domestic-use exemption stops applying once cameras capture beyond your own boundary, the ICO publishes guidance for householders and separate expectations for commercial operators, signage is read differently in a home, a shop and a workplace, and audio is treated more strictly than video.

Two mechanisms keep this safe. First, your never-state list. Anything you will not have said in your name, such as a ruling that a particular setup is lawful or a fixed retention period, is enforced against every draft deterministically rather than requested politely of the writer. A draft that breaks it does not publish.

Second, articles explain how the rules work and link to the regulator for the current wording instead of freezing a figure into the page. That is the same pattern the trades pages use for grant schemes, and it is what stops a page ageing badly between reviews.

What a competitor cannot copy is your data. The manufacturer datasheets for the cameras and recorders you actually stock. The specification fields in your catalogue. The questions your sales desk answers by phone and the ones buyers leave on your marketplace listings.

Where that first-party data exists, IT Master ingests it read-only and grounds articles in it. A comparison article then quotes the sensor, lens and IR range from the datasheet of a model on your shelf; a storage article uses the bitrates your recorders actually produce; a kit article recommends parts from your own catalogue rather than a generic shopping list. Every draft is also scored for novelty against the pages currently ranking, so a draft that restates the top ten does not pass.

The live CCTV tenants on the engine are cctvline.co.uk and cctvmaster.co.uk across trade and retail, and eurocctv.co.uk on trade. Each connects on its own key, and each is grounded in its own catalogue rather than a shared pool of copy.

An installer and a homeowner search differently. The installer wants channel counts, PoE budgets, ONVIF compatibility, warranty terms and whether it is in stock. The homeowner wants to know whether the camera will read a plate at the end of the drive and whether the neighbours can object. Sites that serve both usually write for one and lose the other.

Each site is a separate tenant with its own brand voice, terminology and rules, and its data stays isolated from every other tenant. Your wording is enforced rather than suggested: if you say recorder and never DVR, or turret and never eyeball, every article follows. The same rules carry what the site must never state, which is where trade-only pricing, account terms or a discontinued range belong.

Briefs then split by audience. A trade piece can assume the reader knows what a PoE budget is and spend its words on the table; the consumer version explains it. Both link into your existing category and product pages through internal linking instead of sitting in a blog nobody reaches.

Delivery matches whatever the storefront runs on.

• WordPress: install the plugin, paste one key, and articles arrive as posts with the Yoast or Rank Math fields filled.
• Next.js: the @itmaster/sdk package, documented in the docs.
• Anything else: a read-only REST pull feed or a signed push webhook to an endpoint you name.
• Shopify, Webflow, Wix, Squarespace: Shopify and Webflow are set up by us on request; Wix and Squarespace get a copy-paste queue in the dashboard.

Every article carries JSON-LD schema, a hero image, an FAQ block and internal links, and IndexNow tells the search engines that support it on publish.

Each night the engine reads Search Console for every published article, queues a refresh for the ones decaying, and asks GPT and Perplexity the questions your articles answer to record whether you are cited. The pass rate is measured rather than promised. Across 186 Standard-tier runs, 51% of drafts passed every check first time; the rest were revised or rejected. You pay per published article from a prepaid balance, with no subscription; see /pricing. Before spending anything, run the free site check against your current site, or the AI Visibility Checker to see whether assistants name you when a buyer asks who to buy CCTV from.