Table of Contents
In 2019, I started testing an affiliate campaign with two or three Twitter accounts. I posted manually, changed the timing and copy, and watched what happened. Once the concept worked, I began automating it.
As of 12 September 2026, the experiment has generated $160,000 in cumulative revenue. That is the total before costs, accumulated since the project began.
When I first wrote about it in October 2021, the figure was close to $85,000. This is the updated story: how the system worked, what took the most effort, and why maintaining it became a project of its own.
The numbers, with their dates attached
| Measure | What it covers |
|---|---|
| 2019 | The year I started the campaign |
| 2–3 accounts | The initial manual test |
| About 2,700 accounts | All accounts created over the project |
| At most 900 accounts | The largest number running together |
| 1–1.5 million tweets | Posting volume during a twelve-month period |
| Close to $85,000 | Cumulative revenue reported in October 2021 |
| $160,000 | Updated cumulative revenue as of 12 September 2026 |
The tweet count covers twelve months; the revenue total covers the life of the project.
I tested the boring part manually
Before building the automation, I needed to see whether the campaign worked on a small scale. With two or three accounts, I could test posting schedules, different kinds of tweets, links and frequency without having to manage hundreds of moving parts.
That stage was basic, but it gave me something concrete to automate. I had a working process before I started trying to multiply it. A script can repeat an action thousands of times; it cannot tell you whether the original idea was worth repeating.
The engine was an RSS feed and IFTTT
The core setup was simple: an RSS feed produced a new item, IFTTT detected it, and a connected Twitter account published a tweet.
I wrote the RSS generator in PHP. Each account had its own feed, and the feed supplied the text that would become the tweet. The original system used variations of prewritten copy, affiliate links and hashtags. IFTTT handled the connection between the feed and Twitter.
Splitting the system this way made its parts easier to understand. The generator decided what to publish. The automation service passed it along. The Twitter account was the distribution channel.
I also described WordPress’s built-in RSS feeds as a possible alternative in the original article. The campaign itself used my custom PHP generator.
Setting up the accounts took the most time
Publishing a tweet automatically was only one part of the job. Creating an account, preparing its profile, connecting it to IFTTT and checking the result was repetitive work. At the time, I estimated ten to fifteen minutes for each setup. Verification requests could interrupt the process too.
The pictures from that period show some of the hardware I discussed alongside account setup: a prepaid SIM listing and a multi-SIM adapter.
I eventually outsourced that work. In my experience then, an account created and connected to IFTTT cost around $2.50 to $3.00. Those are historical prices from this project. They do not tell you what the entire campaign cost: development, supervision and ongoing maintenance were separate concerns.
Delegating also meant documenting the process clearly. Someone else had to understand what a finished setup looked like and where a connection might fail. The instructions and the checks became part of the system alongside the code.
At 900 accounts, management became its own problem
With hundreds of accounts active, keeping track of everything became difficult. Accounts could be banned or lose visibility. I built a custom dashboard to follow the accounts and the problems I was encountering.
Back in 2019, an account might last only a few days, while others lasted six to eight months. Some were still active when I first published this story in 2021.
Automation reduced the repetitive posting work. It still left me with a changing collection of accounts and a need to know what was working. The dashboard was a response to that practical problem.
The setup belongs to its time
Even the 2021 version of this story said the method needed changes because Twitter had altered its anti-bot measures. Treating that article as a current installation guide would miss that context.
There are concrete differences today. IFTTT’s current documentation says affiliate URLs are not supported in the text of its Twitter posting actions. X’s automation rules prohibit duplicate or substantially similar posts across multiple accounts, as well as automated accounts serving duplicative purposes. These points were checked on 12 September 2026. See IFTTT’s Twitter integration and X’s automation rules.
The old interface walkthrough, free-plan limits and account-creation tactics would therefore make a poor set of instructions for a new project. The useful part of this account is the actual experiment: a small manual test, a simple publishing system, a large increase in volume, and the maintenance that followed.
The updated result: $160,000
Updating the figure from roughly $85,000 to $160,000 puts a new number on an old experiment. The mechanics were simple; maintaining hundreds of accounts was where it became real work.
If you have questions about the experiment, leave a comment below. I will do my best to answer.
This case study was first written in October 2021. Revenue updated on 12 September 2026.
