Work · Property

A year of content produced over a weekend, and the rankings moved.

The portal already ranked 2nd to 6th on Google for more than four thousand property keywords, and was producing content one article at a time: one freelancer, two weeks, around $200 each.

Client
A Zimbabwean property portal
Our part in it
Jon Dallas, freelance
When
October 2024 to January 2025
The size of the problem

The portal this ran on.

LargestOnline property portal in the country
6,869Listings live today
156Agencies with a profile on it
4,000+Keywords it already ranked for

The market position is Bloomberg's own editorial description, 9 May 2025, rather than the company's claim about itself. Listings and agencies were counted from the live site on 26 August 2026 and drift by around ten percent. The keyword count is the client's own baseline going into the work, and it describes what they already had rather than anything this produced.

What we built

A six-stage pipeline. It picks the keyword, scrapes the competing pages, queries the portal's own listings database for figures nobody else has, assembles the expertise signals Google looks for, generates the article, and QA-checks it before it reaches the CMS.

Each stage feeds the next, and the last one feeds the first. What ranked decides what gets written next, so the pipeline gets better at choosing rather than only faster at producing.
Step by step

How an article gets written without anybody researching it first.

The pipeline

The pipeline was prototyped in under a fortnight and the prototype became the production system. Each stage below is a real step with a real tool behind it.

Scored, then chosen
TermRank nowWorth writing
A suburb plus a property type4Yes
A broad national term38No
A suburb plus a price question6Yes
A term with no volume2No
Stage 1

The keyword is chosen by what is winnable, not by what sounds good

Search Console and ahrefs are scored together across more than four thousand terms the portal already ranked for: search volume, competition, the position it currently holds, and what that position is actually worth in clicks.

A term sitting fourth with real volume is worth writing for. A term sitting fortieth is not, however interesting it is, and no writer chooses that way on their own.

  • Google Search Console
  • ahrefs
1 of 6
The production board, including the column for what failed the check.
A drawing of the interface, not a capture. The production board, including the column for what failed the check.

What happened once it was running

450 articles were researched, written and checked in 48 hours and went into the content system as drafts for the client to review and publish. The year that followed shows it: the portal published 133 articles in 2025 against 53 the year before. The process it replaced was one freelancer, two weeks, and around two hundred dollars an article.

450Articles in 48 hours
4,000+Keywords already ranking
~8%Flagged by QA for revision
+2Average positions, six months on
Measured from the delivered pipeline run

The 450 figure and the 48 hours are the client engagement's own record. The published article counts were counted from the portal's live sitemap on 26 August 2026: 53 articles dated 2024 and 133 dated 2025. The 450 were drafts for review, and the entry says drafts because that is what they were.

Why this page is short

Every figure on this page traces to something we measured.

This page once carried claims that described the work inaccurately. They came down in an audit of our own copy, and each one went back up only when its numbers could be traced to something measured. A case study you cannot check is worth nothing to you, and one that turns out to be wrong costs us more than it ever earned.