Forty-five days. That is how long I ran this experiment: asking ChatGPT to help me make $100 per day online, picking the method my audience voted for, working the process every single morning, and then showing you the full truth of what happened. Spoiler: after 45 days I had made exactly zero dollars. But I also learned something far more useful than a quick commission check would have taught me, and this post is going to walk you through every step of what I tried, what the numbers actually showed, and what I am going to test next so you do not have to waste 45 days finding this out yourself.
My name is Alston Godbolt. I run alstongodbolt.com to help you make money online without the hype. A few months before this update I recorded a video where I asked ChatGPT to list free ways to make $100 per day using only the AI tool itself. ChatGPT gave several options. I let my audience vote on which one I should try first. They picked answering questions on forum sites. That decision kicked off one of the more instructive 45-day experiments I have run, and I am going to give you the unfiltered numbers right here.
What You’ll Walk Out With
- Why Reddit can block a new account within hours of posting and exactly what triggers it
- How to use a keyword research tool to find low-competition questions that still get tens of thousands of searches every month
- The exact 5-to-10-minute daily workflow: keyword research, ChatGPT prompt, Midjourney image, Facebook publish
- Why borrowing a high-authority domain like Facebook can speed up your path to ranking versus starting a brand-new website
- The real 45-day traffic numbers from this experiment and what they reveal about Facebook page reach in 2023
- A 3-version split-test plan to isolate whether your results are being limited by the platform, the page type, or the content itself
- How to figure out which online income method fits the skills you already have at finder.platformproof.com
How the Experiment Started: The Reddit Ban
The first thing I did was take ChatGPT’s suggestion and go answer questions on Reddit. Reddit is organized into sub-communities called subreddits. Think of each one like a niche Facebook group with its own culture and rules. I went to the finance subreddit and started answering questions using responses that ChatGPT helped me write. I figured this was a low-barrier way to build credibility and maybe drive people toward affiliate offers.
Within hours I was blocked from multiple subreddits. The reason, as best I can tell, is that I answered too many questions too quickly. Reddit’s moderation system flags that kind of behavior as spam. The subreddits are, as I put it in the video, prickly when it comes to people posting a lot of content in a short window. So after roughly one day of trying the Reddit approach, I was out. No commissions, no traffic, just a lesson in platform rules.
That could have been the end of the experiment. Instead, I pivoted. I took the core idea, which was using ChatGPT to write helpful answers to questions people are already asking online, and moved it to a platform where I had more control over my presence: Facebook.
The Pivot: Why I Chose Facebook Instead of Starting a New Blog
When I decided to leave Reddit, I had a choice. I could start a brand-new blog on my own domain, or I could publish on Facebook. I picked Facebook for one strategic reason: domain authority.
Every website has what is called a domain ranking or domain authority. It is a score that reflects how much trust search engines give that site based on its age, backlinks, and traffic history. A brand-new blog starts at zero. Facebook sits at the top. By publishing blog-style articles directly on my Facebook page, I was essentially borrowing Facebook’s authority so that my content would have a better shot of showing up in Google search results faster than a fresh site could.
Starting a new blog is genuinely difficult when you are starting from scratch. Google does not trust a brand-new domain. You are not getting traffic. Nobody knows you. By posting on Facebook I was trying to get in the door faster. That was the hypothesis. The 45-day results would tell me whether it held up.
The Daily Workflow: Keyword Research, ChatGPT, Midjourney, and the Affiliate Link
Before I show you the numbers, you need to understand exactly what I was doing each day. The whole process takes me between 5 and 10 minutes per post. Here is how it works step by step.
Step 1: Find a Low-Competition Keyword with Ahrefs
I use a paid keyword research tool called Ahrefs. I know Ahrefs is expensive, and I hear about it often. You can do the same research with free tools. I use paid tools because time is money, and saving time is the same as making more money. But the strategy works regardless of the tool.
I type in a seed idea. For this experiment my niche was dogs, so I might type in “can dogs eat” and let the tool show me all the questions people are asking. The keyword difficulty score runs from 0 to 100. The higher the number, the harder it is for a new page to rank. Since my Facebook page was essentially a new publisher with no track record, I set the maximum keyword difficulty to zero and filtered results from there.
When I searched “can dogs eat chicken nuggets” with that zero-difficulty filter, I could see that the search term gets 79,000 queries per month and there are 1,000 related keywords I could target. That means a lot of people are searching for answers about dog food safety, and the competition to rank for those questions is almost nonexistent. That is the sweet spot this method hunts for.
Step 2: Feed the Keyword Into ChatGPT
Once I have a keyword, I go to ChatGPT and type a prompt like: “Write a 3,000 word blog post on ‘can dogs eat chicken nuggets.’ Do not include word count.” That last instruction keeps ChatGPT from padding the piece with filler about how long the response is. I then copy the full output.
While ChatGPT is writing, I move to the next tool.
Step 3: Generate an Image with Midjourney
I use Midjourney to create a header image for the post. Midjourney is also a paid tool. I type something like “imagine dog with chicken nuggets” and let the AI generate a set of images. In the video I showed an image of a dog surrounded by chicken nuggets that looked genuinely photo-realistic. That image becomes the featured graphic for the Facebook article, which gives the post a visual identity that can help it stand out and may also signal quality to search algorithms.
One note from the 45-day experiment: I had only recently started adding Midjourney images toward the end of the test period. Images may be part of what was missing from my earlier posts.
Step 4: Post to the Facebook Page and Add the Affiliate Link
I paste the ChatGPT blog post into my Facebook page as an article. The longest part of this step is sometimes Facebook just gets stuck on the upload and I have to refresh and try again. That aside, the publish step takes less than a minute.
Embedded in each post is an affiliate link to a product on Amazon. For this experiment I chose a dog probiotic that retails for $20. My reasoning was simple: if someone is reading about whether their dog can eat a particular food and the answer suggests the dog might have a sensitive stomach, a probiotic is a natural next step. Pet owners tend to buy things for their animals without much hesitation, especially at a $20 price point. We actually use this product at my house, so I was comfortable recommending it.
Beyond probiotics, I had a longer list in mind if this started working: pet insurance, telehealth vet services, specific dog food brands, training tools. The probiotic was just the starting point.
The 45-Day Results: Real Numbers
Now for the part you probably came here for.
Over the roughly 90-day window I looked at in the analytics, my Facebook page had received about 100 visits total. Not 100 per day. One hundred total. And honestly, some of those visits were probably me checking my own content. The reach numbers were not good.
The best-performing post in terms of search interest was “can dogs eat eggplant.” According to Ahrefs, that keyword gets about 4,000 searches per month and has a keyword difficulty of 2. That should be easy to rank for. When I checked who was in the top 10, I found sites with domain rankings as low as 11 and 15. One site was just a 705-word blog post on a domain with a ranking of 31. Those are small, low-authority sites. My Facebook page, borrowing Facebook’s domain authority, should have been able to outrank them. But I was not in the top 10.
Total money earned after 45 days: zero dollars.
I want to be clear about something. I am not telling you this to discourage you. I am telling you this because honest results are how you learn. The experiment has not failed yet. It has just not worked yet. There is a difference.
Not sure which online income method to try first?
Answer 7 questions and get a personalized match based on your time, skills, and goals at finder.platformproof.com.
Why It Might Not Be Working Yet
I thought through several possible explanations for the low traffic and zero commissions. Here is what I came up with:
Facebook suppresses pages. Facebook does not give organic reach to pages the way it used to. A personal profile or a group might get seen by more people than a business page with the same content. That is a platform-level reality in 2023, and it could be dragging down my rankings or my click-through rates even if Google does index the posts.
Raw ChatGPT output may not be enough. I wanted to test whether purely AI-generated content, with minimal editing, could rank. The honest answer so far is: probably not without some modifications. I should be adding original information, unique angles, and better formatting to these posts. I kept them close to raw ChatGPT output intentionally, but that may be costing me in search quality.
Images came late. I only recently started adding Midjourney images to the posts. It is possible that the earlier posts without images were already indexed but ranked lower because they lacked a visual element. Google tends to give some preference to pages that include relevant images with proper descriptive text.
The affiliate link might be the wrong conversion point. I am wondering whether dropping a direct Amazon affiliate link into the post is the right call, or whether I should be sending people to a landing page and building an email list first. A landing page could let me promote a dozen products over time instead of one probiotic. I might test replacing the affiliate links in some posts with a lead capture form to see if that changes anything.
Honest Drawbacks of This Method
I want you to go into this with realistic expectations if you decide to try something similar. Here is what I have found to be the real friction points.
The method is not truly free. ChatGPT has a free tier, but Ahrefs and Midjourney are both paid tools. Ahrefs in particular is expensive. You can substitute free alternatives for keyword research, and you can use free image sources or skip AI image generation entirely, but the experiment as I ran it involved real costs.
Facebook’s organic reach for pages is limited. This is not new information, but it is something people underestimate when they try to use Facebook as a traffic source. If you are not running ads or posting in groups, the reach from a business page is often very low.
Pure ChatGPT content may not rank long-term. Google’s helpful content updates have been aimed at exactly this kind of material: AI-generated answers to common questions without original insight. That does not mean AI-assisted content cannot rank. It means pure, unedited AI output will struggle unless you are adding something that cannot be easily replicated by other sites doing the same thing.
Traffic takes time. Even if every post I published was perfectly optimized, SEO takes months to show results. 45 days is a short window for organic search traffic to build. I am aware of that. But I also think 100 visits in 45-plus days is well below what the keyword data suggests is possible, which tells me there is something else going on beyond just needing more time.
The 3-Version Split Test: What I Am Trying Next
Instead of guessing which variable is the problem, I am going to run a controlled test. Here is the plan.
I will pick one keyword, something like “can dogs eat chocolate,” and create three versions of a blog post on that topic using ChatGPT. Each version will be slightly different. I will then publish one on my Facebook page, one inside a Facebook group dedicated to dog owners that I own and manage, and one on my own personal blog or a new standalone pet website.
Then I wait and watch. If none of them rank, the problem is the content itself. ChatGPT is not producing posts that Google finds useful enough. If my personal blog ranks but neither Facebook option does, the problem is the platform. Facebook’s domain authority is not helping the way I theorized. If the Facebook group ranks but the Facebook page does not, the problem is specific to how Facebook treats business pages versus community groups.
This is how marketers actually work. You isolate one variable at a time, test it against a control, and let the data tell you what is true. Guessing is expensive. Testing is cheap. I would rather run a clean three-way test than spend another 45 days wondering which piece of the puzzle is broken.
The Real Lesson From 45 Days of Zero Dollars
I know this update might not be what you were hoping for. You came here wanting to hear that ChatGPT made me $100 per day and here is how you can copy it. Instead I am telling you: the method did not work yet, but I know why it might not be working, I have a plan to test each possible cause, and I am still in the game.
That is actually the most valuable thing I can share with you. Most people quit after one try. They set up a system, it does not produce income in week one or week four, and they walk away convinced that making money online is impossible or that it only works for people with big audiences. That is not true. It just means you have not found the combination that works for your niche, your platform, and your content style yet.
This experiment is still running. The process takes me under 10 minutes a day. I might accidentally earn a commission any day. And if I figure out the right combination through testing, I will scale it fast. The point is not to quit. The point is to keep adjusting until you find what works.
Find Your X
ChatGPT-assisted content writing works for some people in some niches on some platforms. It did not hit $100 per day in my first 45 days. Maybe it will in the next 45. But if you are sitting here wondering whether this specific method is even the right one for you to try, that is a fair question. The best path to your first $100 online is not always the one that worked for someone else. It is the one that matches your existing skills, your available time, and what you actually know how to talk about.
If you want help figuring that out, visit finder.platformproof.com. Answer seven questions about your background, your schedule, and your goals, and the tool will match you with the income method that fits you specifically. No guessing, no copying someone else’s niche. Just a clear starting point built around what you bring to the table.
Frequently Asked Questions
Can you really make $100 per day using only ChatGPT for free?
ChatGPT itself is free, but the full method Alston tested also used Ahrefs (a paid keyword research tool) and Midjourney (a paid image generator). The core workflow could be replicated with free tools, but you would be trading cost for time. After 45 days of this experiment, $0 was earned, which does not mean the method cannot work, but it does mean it did not produce results in this time window with these specific variables.
Why did Reddit ban the account so quickly?
Reddit subreddits have active moderation and community rules around promotional or rapidly repeated posting. When Alston answered too many questions in the finance subreddit in too short a time, the mods flagged it as spam behavior and blocked the account from those communities. Reddit is generally not a welcoming platform for anyone who looks like they are running a volume content strategy.
What keyword difficulty score should you target with this method?
Alston set his maximum keyword difficulty to zero in Ahrefs. That is the most conservative filter because a new Facebook page or blog has no domain trust built up. Even at a difficulty of zero, the “can dogs eat chicken nuggets” keyword still had 79,000 monthly searches and 1,000 related keywords to work with. Starting at zero keeps you from competing with established, high-authority sites before you have built any credibility.
Is Ahrefs the only tool that works for this kind of keyword research?
No. Ahrefs is Alston’s preferred tool because it saves time. He acknowledged in the video that free tools can accomplish the same thing. Options like Ubersuggest, Google’s free Keyword Planner, or even manual searches using Google’s autocomplete and “People Also Ask” section can surface low-competition questions without a paid subscription. The strategy is tool-agnostic. You are looking for high-search, low-competition questions in a niche with buyers.
How long does it actually take to produce one post using this method?
Alston said the entire process takes him 5 to 10 minutes per day. The longest single step is sometimes a Facebook upload hiccup where the platform gets stuck and requires a refresh. Keyword research takes a minute or two with a saved filter. The ChatGPT prompt and copy takes a minute. The Midjourney image generation runs while ChatGPT writes. Publishing takes under a minute. It is genuinely a low time commitment once you have the workflow set up.
Why did Alston choose a $20 dog probiotic as his affiliate product?
The probiotic matched the content: people reading about whether their dog can eat a certain food are already thinking about their dog’s digestion. A probiotic is a logical next purchase at a price low enough that it is not a hard decision. Alston also mentioned that his household actually uses this product, which meant he could recommend it honestly. He had a longer list of potential products in mind if this method began generating traffic: pet insurance, telehealth vet services, and other pet supplies.
What were the actual traffic numbers after 45 days?
Alston showed approximately 100 total visits over a 90-day window in his Facebook page analytics. He noted that some of those visits were likely his own views checking the content. The best-ranking post was “can dogs eat eggplant,” which targets a keyword with 4,000 monthly searches and a difficulty of 2. Despite those favorable conditions, the post was not ranking in the top 10 results on Google. Total revenue from affiliate commissions in this period: zero dollars.
What is Alston planning to test next to troubleshoot the results?
He outlined a 3-version split test: take one keyword (“can dogs eat chocolate” was the example he mentioned), create three ChatGPT-written posts on that topic, and publish one on his Facebook page, one in a Facebook group he owns, and one on a standalone personal blog or pet website. Whichever version ranks tells him where the problem is: if none rank, it is the content; if only the personal blog ranks, it is a Facebook platform issue; if the group ranks but the page does not, it is specific to how Facebook treats business pages.
Read Next
If you want to see how I approach finding income streams that actually match where you are right now, this post walks through a method that has worked for people starting from zero without a product, a big audience, or a paid tool subscription.
Check out: I Tried Making Money Online for 10 Years
Sources
- Original video: UPDATE: I Asked ChatGPT To Make Me $100 Per Day | 45 Day Results (YouTube, Alston Godbolt)
- Ahrefs keyword research tool: ahrefs.com
- Midjourney AI image generation: midjourney.com
- Amazon Associates affiliate program: affiliate-program.amazon.com
- Facebook Pages organic reach data referenced in the video from Alston’s own analytics
Helping 1 million working adults make their first $3,000 online with the skills they already have. Alston Godbolt, Platform Proof.