Highlighted posts

We keep reading from Google what needs to rank, but in the end game, I think we know it's going to be their ads. I remember the good old days when you could spam the hell out of Google, but now you need a huge budget for content, backlinks, and making sure your site is fast and well-designed. My huge question is, what other update can they give us that we don't know already?

Zwielicht said:

Okay, so @Impulse and I looked into this.

It is possible for us to have a public one, but there are some caveats with it, one of them being that this is a black hat forum, and those in charge of enforcing those laws might just assume we're breaking the law, even if we prohibit certain kinds of discussions.

We're going to have to decline this one for now, but it is something we can add in later on, especially once we move to our new servers.

Click to expand...

You need some sort of cheap, legal attorney/lawyer to help you with these matters.

Like Michael Scott had that guy who focused on motorcycle head injuries and diet-pill lawsuits or whatever.

Once we've written these up, we'll either add it in as its own thread here or we'll add a new chapter to The Trading Post rules.

Rem said:

I received a review copy
bery gud
#1 trusted sell3r
AAA+
---
ok jokes aside
i do agree on this

To me this is a sale, a trade
just not with your body money but rather, with your opinion on it.
We have a rule saying fake reviews are a no-no
but what about people who never actually give the feedback?
to me, that is a scam.

Click to expand...

That is something I've always meant to take care of. I see it as being similar to scamming if someone takes a review copy and doesn't provide a review. The conundrum with enforcing this before was that you can't force them to leave feedback after that, because they might just provide negative feedback out of spite.

I do have a couple of ways to go about this. One is to create a list of users who are banned from accepting review copies in the Beta Testing Hub.

This would just be a simple list of members who have abused this feature and are no longer allowed to use it. I could even take it a step further and ban those people from even viewing the section using a custom usergroup).

What do you think about this idea? If I proceed with going this route, it will be fully explained in the guidelines for review copies as well.

I’ve been meaning to write a guide like this for a while. There are a lot of ways AI can improve day-to-day work, but I’ve seen plenty of people try it once, expect it to do everything for them, and then write it off as useless when they don’t get what they wanted out of it. This is a guide for those people, along with anyone else who wants to make repetitive, annoying work go faster using AI.

In my previous job, I was stuck in a support role with a heavy workload and basically no backup. Then new management came in and added a pile of tedious processes in the name of ‘efficiency’, which only made everything slower and more frustrating. They also tried introducing various systems to track the work that everyone was doing, but it was incredibly inaccurate. So, on top of having to work with these new processes, I also needed to figure out how to game everything so I could look good enough to pass whatever arbitrary standard they were using.




I needed a way to get things done without losing my mind, and that’s where AI came in.

Once I started using ChatGPT, the difference was immediate. I had been stuck doing more work than ever at a lower standard, and AI helped flip that around. It cut down on repetitive tasks, helped me communicate clearly with difficult people, made complex jobs easier to manage, and kept my writing consistent and professional. Even when I wasn’t feeling particularly professional myself.

In this guide, I’ll break down the ways I used AI to work faster and smarter, and hopefully give you a few ideas on how it can help you do the same.

Disclaimer: I had access to the premium version of ChatGPT, which definitely improves the quality of responses. That said, everything in this article can still apply to the free version or other decent AI tools, so you don’t need to spend money to make it work.

---

1. Common Tasks AI Can Handle​

The real key to using AI well is knowing what to hand off to it. You won’t be able to replace your entire job, but you can figure out which parts of your work are dull enough to outsource without causing problems. That’s where it shines: boring, repetitive stuff you’d rather not waste energy on. Things like answering the same basic question ten times in a day, summarising a long block of text, or cleaning up messy notes.



Even if the output isn’t perfect, getting a rough draft or basic structure from AI can save you a lot of time, and you can always tweak the result if it needs polish.

Here are a few examples of tasks that are easy to speed up with AI:

1. Content creation (guides, FAQs, documentation)
   Obvious, but worth including. This is one of the most common uses for AI, and with good reason. The key is to avoid lazy prompts. If you give it vague instructions, you’ll get vague content. You’ll get better results if you provide clear goals, structure, and context (more on that in the next section).
   
   
2. Professional communication (with employers, work colleagues or clients)
   Have you ever had to respond to someone professionally but don’t know how to get the tone right? ChatGPT’s writing style tends to be very appealing to people that are more corporate or formal. It’s also useful if you’re trying to explain something without sounding overly blunt.
   
   
3. Preparing daily task summaries and updates
   Sometimes, you may be asked to provide summaries of the work you have done over a period of time. If you have had the misfortune of needing to do this tedious task, you can ask an AI bot to write this up for you. You just need to provide a very rough draft of the kind of work you do and ask it to fill in the blanks as best as it can.
   
   
4. Text extraction (OCR alternative)
   I haven’t seen much discussion about this, but AI tools like ChatGPT can work as an alternative to traditional OCR software by extracting text from images. Instead of installing a separate program just to copy text out of an image, you can use AI to handle the whole process. Just upload a screenshot, and it can pull out the text, clean it up, and even summarise or rephrase it for you.
   
   
5. Improved search queries
   AI is also useful for figuring out how to search for something when you’re not sure what keywords to use. It can answer questions directly, rephrase what you’re looking for, or help break down more complex topics. There are a lot of AI tools now that have browsing capabilities and can even search the web for you and summarise results.

---

2. How I Used AI​

Now that I’ve covered general use cases, here are some of the specific ways I used ChatGPT in my day-to-day work. Most of my job role was within the realms of customer support, so if you’re in a similar line of work, you’ll probably find these examples useful.

Support Ticket Responses​

A big chunk of my time was spent responding to support tickets from users.

The most obvious use case for using ChatGPT in the support desk was for writing responses. It was great for responding to tickets from members and giving them answers that were detailed, professional, and polite – especially when I didn’t have the patience for it myself.




And I wasn’t the only one doing this either. A few other support staff were clearly using ChatGPT too. Funnily enough, our manager clearly had no idea, which kind of proves how easy it is to get away with using AI when the people above you don’t really understand the tech.

I also used it to create canned responses for some of our more common questions and queries. I’ll go into more detail on how I prompted it later, but the short version is: I gave ChatGPT the tone I wanted, the kind of result I was aiming for, and let it figure out the phrasing.

Marketplace Moderation​

Alongside support, I handled marketplace moderation: reviewing sales thread content, processing thread edit requests, and sending invoices.

I found that I could use ChatGPT to speed up certain marketplace processes; Namely the thread edit process, which would involve tediously checking the two sales threads to see what was changed and whether or not I would be required to charge them extra for changing too much. I was able to provide ChatGPT with the thread editing guidelines to remember, and then I could compare the two different sales copies to summarise the changes that were made. This worked especially well with text-based sales threads, but could even be done with image-based sales copies too (although it’s a lot more limited with images).

The added benefit of using AI to analyse the changes was that it also meant I could quickly copy and paste the summary to sellers who wanted to know what had changed and why they were being charged extra.

Internal Team Communication​

Communication between team members at the old forum was… not great. There were issues with conflicting orders, goalpost-shifting and team members spending a lot of their time going through the work that was done by others to look for any mistakes. In a hostile work environment like this, what was the best way I found to talk to others without invoking their ire? AI.

I noticed people were far more receptive to my messages when they were written by ChatGPT. Turns out, it’s very good at writing in a way that’s hard to misinterpret, which is especially useful when dealing with people who aren’t exactly known for their reading comprehension skills.

Content Creation and Editing​

I also wrote support content to keep things running smoothly. This included seller guides, canned responses, and internal docs which all existed to make sure the work team weren’t all saying different things.

For dry or low-effort content, like policy threads or announcements, ChatGPT did most of the heavy lifting. I’d give it the general idea and tone, let it generate a draft, and then do a quick pass to make it less bland and more human. It saved time, lowered my mental load, and let me knock out boring writing tasks without sacrificing quality.

I also used it to review content I had written myself – especially when I wanted to tweak phrasing or shift the tone depending on who I was talking to.

---

3. Writing Good Prompts​

Anyone who uses AI regularly will tell you that prompting is everything. It’s not exactly a groundbreaking insight, but it’s true. It can be hard to pinpoint exactly what makes a prompt good, but here’s what I’ve learned works best:

Don’t worry about your writing​

Yeah, I know it sounds stupid for me to say “my tip for writing good prompts is to not worry about how good your writing is” but just bear with me. ChatGPT is pretty good at deciphering typos, awkward phrasing, and walls of text. If you’ve got a bunch of rough thoughts, don’t even think too much about it – just throw them in and let the AI tool do the cleanup.

Clearly define your goal​

Be clear about the result you want from your prompt. If you want something written for a specific purpose, you should specify what that is. For example, don’t just say, “write a reply.” Say what kind of reply. Is it formal? Blunt? Empathetic? Is it for a customer, a coworker, or someone you’re mildly annoyed at? A lot of AI responses default to being overly polite and vague unless you steer it in the right direction.

Give it the context it needs​

If you’re dealing with something complicated, feed the bot all the background information. Long ticket history? Drop the relevant parts in. Ongoing team drama? Add a quick summary. And you can just copy and paste anything if it needs to be low effort. AI tools will take any context you give it to provide a more appropriate response for you.

Requesting specific formats or styles​

Make sure to give more specifications to the AI if you need a specific tone and style. For instance, you can tell it that you want to sound less formal. If you need a certain writing style or language setting (like UK or Australian English), just tell it. You can even ask it to avoid formatting you don’t want, like bullet points, which tend to be an instant ‘this-was-written-by-AI’ giveaway.

Use follow-up prompts to refine responses​

AI works best when you treat it like a tool you’re working with, not just something you throw tasks at. One of the most helpful habits I developed was using follow-up prompts to refine responses. If something was too long, too vague, or didn’t have the tone I needed, it was simple to tweak the result with follow-up clarifications. You can try asking for a shorter version, a more professional rewrite, or a more direct answer.

---

4. Avoiding Common Pitfalls​

AI is useful…. sometimes a little too useful. Which makes it easy to become lazy with it, and that’s when you stop getting good results. Whether it’s blindly trusting the output or just sounding like a tech demo instead of a human, these are the mistakes I’ve seen (and made) that are worth avoiding.

Over-relying on AI without fact-checking​

AI will give you clean, professional-sounding nonsense if you’re not paying attention. Even if you’ve given it a lot of information, there’s always a chance it can write something in error or simply misinterpret something that you thought was clear. A good practice is to always verify anything factual, especially if it’s going to be published or sent to someone like a client.

Not giving it enough context to work with​

Giving it vague instructions will only get you vague output. If you don’t explain what you’re trying to do or who it’s for, you’ll get something generic, half-baked, or just completely wrong. A few extra seconds of context in your prompt saves way more time than trying to fix bad output later.

Letting the tone go completely off the rails​

Tone is one of the easiest ways that AI-generated content can mess up, and it usually misses the mark unless you’re clear about what you want. It can take some fine-tuning and figuring out what descriptors get the exact vibe that you are going for.

Also, try to avoid telling AI to write “in a casual tone”. I tried that when I first started out with ChatGPT and what I got was writing that sounded like this:

Sounding like an AI wrote it​

If your content reads like it came straight out of a chatbot, people will notice, and they won’t take you seriously. You don’t need to rewrite everything, but keep in mind which aspects of AI content are dead giveaways to people, and try to edit out those dead giveaways: overuse of bullet points, unnatural emoji use, and way too many em dashes.

---

Thank you for reading all of this. I hope this will help some people see how they could improve their work and get AI to do the menial work that we all hate doing.

AI isn’t going to do your job for you, but it can make your job suck a little less. Whether you’re stuck writing policy threads no one reads or having to explain the same thing for the tenth time in one day, AI tools like ChatGPT can save you time, energy, and brainpower.

If you are reading this and have used AI to improve your work in any way, please feel free to leave a comment and share what you’ve done. I’d love to see what tricks other people have found, and it might help someone else find a smarter way to use it too.

CapCut has many good subtitle features; its speech-to-text function is also very good.

Zwielicht said:

The new section is now live and the prefixes have been added in for Google, Meta, and Linkedin to start.

Click to expand...

Can you add a prefix for Tiktok Ads and another one for Native Advertising (that would contain threads about Taboola (incl Realize), Outbrain, Gemini...)?

I'm rethinking the Meta Ads prefix, maybe it would be better to create just a Social Ads prefix that would include Meta, Tiktok, Snapchat, Twitter, etc.. otherwise it'll never end.

The reason I think Google and Bing should keep their prefixes is that each has tons of practices that are unique to them as well as many different features that would require different prefixes in the same thread, like Search, Display, Video...

I bought a fat-ass book for meditation and I haven't opened it since (more than a year).

I keep postponing.... the consistency and effort it takes is too much for me.

No it won't but anyone who doesn't use AI would surely be left behind

I am a big fan of human touch

Think of it as your assistant but mind you it can do things that will blow your mind

AI is here and it's still in its early stages it's better to get well accustomed to it and reap it's benefits

Happy Easter to you all.
Already eaten too much chocolate!

travelling this easter

I'm of course not as cool as @Zwielicht to saddle one of these bad boys:



but I love my 2015 Chevy Malibu

Four days off, nice huh. Happy Easter guys.

Anyone doing anything fun this Easter?

Other that that coke vibing bunny @Yes

TheVigilante said:

When you say it just got harder what has got harder?

Click to expand...

There's more competition and the prices of services/backlinks are much higher than before. Markets are more saturated, organic resuls are barely above the fold nowadays.

So there's much smaller piece of the pie to be had and it costs much more to get it.

But the absolutely biggest issue is the frequency and randomness of Google updates. You can't just decide to do everything by the book and be fine, you will get hit again and again for no particular reason whatsoever.

And by the time you figure out why you were hit (if you even figure it out), you will be hit by another update.

I have some very useful data from my attempts over the last decade to get a job.

I did manage to get a job back in 2018 (technically, I had 3 offers in total, but I didn't take the other two). I know it's not easy because I had been applying for about 2 years to so many different positions. Really, from my experience, this whole system is more about who you know rather than you know, but you can still get jobs the regular way.

Anyway, it was actually through my last job that I reviewed many applications myself for both employees and volunteer moderators. It gave me more insight as to why this process is so off for many of us.

Because I was ultimately unhappy with my last job, I did also occasionally continue submitting applications over the years, so this is something I have quite a lot of experience in.

Before I go any further, just be aware that this is just my experience applying for jobs in the United States and reviewing applications for a UK-based company, so it might be different where you live.

Self-Employment
Look, if you care about getting a job, never mention self-employment.

It was only when I put down that I was an employee of my own company that I started getting any calls back. Anytime I put self-employment back on, it was radio-silence. A decade of this is pretty telling, if you ask me.

It's not just me, either. Many of my friends who are self-employed have mentioned facing similar issues.

Politics or Religion
I'm not kidding, I've seen this before. Personally, it didn't bother me because I'm very accustomed to people of different backgrounds, but there are some people who review applications who will be biased and might even reject your application based on your beliefs.

Multiple Short-Term Jobs
This is one I've seen get questioned when reviewing applications. If you look like a job hopper, you're less likely to be selected for an interview because, unless a company explicitly states that they're looking for a short-term employee, most don't want to take the chance on you.

Employment Gaps
These almost always get questioned as if you're a lazy bum who's not interested in working.

Irrelevant Skills
I possess a lot of skills. For example, did many of you know that I'm a very good mechanic? After all, I did much of the work on my classic Mercedes-Benz for nearly a decade. I've also rebuilt engines and restored a couple of cars (a Geo Storm and an MG Mdiget).

But if I'm applying for a desk job, then what my skills as mechanic have to do with the job? Sure, you can rationalise this and say that it's a good demonstration of your soft skills with working independently or something like that. But look, the people looking at your application are very unlikely to put two-and-two together there. They likely have hundreds of applications to go through and not as much time to think about the nuances of yours.

Irrelevant Jobs
This goes back to the last point. I know it's harder to leave out an irrelevant job, especially if it creates an employment history gap, but it's better to find a way to fill it in with something relevant or changing what you did for the company if it's no longer around.

Conclusion
I hope these tips help some of you out. I know it's tough out there right now with how flooded job market is, but I want to keep sharing my experience to help my fellow office outlaws out.

Hello,

Here are all the official social profiles from Office Outlaw:

BlueSky: https://bsky.app/profile/officeoutlaw.bsky.social
Twitter: https://x.com/Office_Outlaw
Instagram: https://www.instagram.com/officeoutlaw
Facebook: https://www.facebook.com/profile.php?id=61574922642454 (Currently not available. Will be fixed soon)
TikTok: https://www.tiktok.com/@officeoutlaw
Pinterest: https://www.pinterest.com/officeoutlaw/
Youtube: https://www.youtube.com/@OfficeOutlaw
Tumblr: https://www.tumblr.com/officeoutlaw
Reddit: https://www.reddit.com/r/officeoutlaw/
Threads: https://www.threads.net/@officeoutlaw

Still a work in progress. However, hopefully within the next few days we will have everything up and running smoothly.

In celebration of the new monetization subforum, I created a list of affiliate networks for various niches. While I didn't have experience with all those affiliate networks, I can assure you that I only picked the top-performing ones. This list may or may not be edited at some point, and there may be some affiliate programs listed twice (because they serve not just one vertical but multiple). This was a lot of work for me to put together, as I had to manually check every single affiliate network and put the information together to make sure they were legit. Nevertheless, before you register on an affiliate network, make sure to verify that they are legit yourself, as OfficeOutlaw cannot be held responsible for an Affiliate network's actions. To visit any of these affiliate networks, simply click on their name in the list. - Here is my list:

E-Commerce:​

Dating/Adult: ​

Sweepstakes:​

Gambling/Forex/Trading/Finance/Crypto:​

Nutra/Health:​

Survey/mVAS/Mobile:​

Gaming/iGaming:​

Miscellaneous:​

This list may or may not be updated in the future.

I looked into the first one. There used to be an add-on for it back in 2015, but it's no longer available.

As for the second, I re-enabled alert improvements. Let me know if there are any issues with it or around the forum.

To be honest a lot of what always worked still works, it just got harder.


When you consider Diggity and the Authority Hacker guys, both shout down their courses and have basically reneged on the lifetime access, it put a lot of people off doing affiliate stuff. I think with affiliate stuff now, Ai is such a game changer.

INCC said:

Any solid advice how to deal with the competition that uses fake click software to drain your budget?

I tried Google PPC a few years ago. My competition killed me instantly.

Tried to look for advice, got none. Some said Google gives you money back at some point but that's not certain any way.

Click to expand...

You should look into software like Clickcease. It doesn't really make much of a difference (and very hard to prove), but if I thought I was targeted, this would be my go-to.
Google has become much better at identifying it in real time, and it's also true that they pay you back (more accurately, they do most of it instantly and the rest over a few more days.This money goes back directly to your budget or ad credit if a day has already passed).

These days, it's rare to see a campaign running on Manual CPC, but why is it?
Of course, using automated bidding strategies, such as Maximize Conversions or Maximize Conversion Value (on Google) or their equivalents on Bing and Meta, is much simpler, but do the conveniences they add outweigh the extreme control we get from Manual CPC bidding?
Let’s find out.

Before diving in, I want to add an important note to the relatively simple bidding strategy question. This is, by the way, the most interesting question on this subject: How does changing Google Ads’ entire bidding eco-system towards automated bidding strategies affect how each of us manages our campaigns? Campaigns do not run in a vacuum, after all. And who earned from all this? Hint: Google, but we too.

Price Per Click in Different Bidding Strategies

Up until a few years ago, every campaign ran on Manual CPC (with very very few exceptions), campaign managers used to know the ranking and actual CPC formulas by heart.

Ad Rank = (Quality Score * CPC bid) | The ard are ranked in the SERP by their Ad Rank - The higher it is, the higher our ad appears on the results page.
Actual CPC = (Ad Rank of the ad below us / Quality Score) + 0.01

Both are still in use today (so they’re worth memorising if you don’t know them) and recalculated in every auction. Or at least supposed to.. But that’s a rant for another day.

These days, most advertisers don’t even use manual CPC (even though we still pay for clicks, mostly). How does everything keep running so smoothly?

This is because even though we use different methods to bid more efficiently, Google still normalizes our bidding strategy to the same base metric: CPC bid.

CPC Bids in Different Bidding Strategies
This is important, so read it twice if you didn’t get it.

Maximize Conversions
When running on Maximize Conversion (especially with tCPA - target CPA), our conversion rate is taken into account when entering an auction. A tCPA of $100 with CvR of 8% will get us clicks at ~$8 - and the bid that would get us an actual CPC of ~$8 is the bid we’re going into auctions with (remember the actual CPC formula?).

Look at it as a general guide because each user has a different chance of converting. To determine a user’s chance of converting Google use the search term, location, and time of day, together with what Google already knows about the user, such as their interests, what other searches they performed, and other metrics used to find the overall likelihood of each query for a conversion. This results in a broader range of CPCs for the same keyword.

Maximize Conversion Value
In Maximize Conversion Value, there's an additional "layer" for the likelihood of a user converting: its worth.

Google estimates the users' value (taking into account its likelihood to hit that) and divides that by the tROAS. The higher the value or lower your tROAS, the higher the CPC bid you enter the auction with.

A Higher Conversion Rate Becomes an Interest of Google
Since an automated bidding strategy is tied to a campaign's success, the higher your conversion rate, the more volume you should expect for a given tCPA bid.

Let’s take the same $100 tCPA bid from earlier. If you manage to increase the CvR from 8% to 12%, it "allows" Google to get clicks at $12 instead of $8 and still hit the CPA goal. Assuming you’re not limited by budget, you’ll probably get significantly more volume, as if you raised your Max CPC by 50%, from $8 to $12.
If you read my marginal CPA guide, you should know that getting such an increase in volume without affecting your CPA is something worth doing.

A higher conversion rate has always been important for scaling up campaigns, but it's more important now than ever.

Real Time Bid Adjustments
Google indeed uses machine learning and its signals to do the same thing that we can do manually and achieve the same goal. That being said, a bot is more likely to win when it comes to understanding signals in real time.

Another interesting observation is that Google advises new accounts to start by creating a campaign using Maximize Conversions. This means that Google may even use signals from other similar advertisers, and not just from within an account, at least to kick off things


Automated Bidding vs Manual CPC
Let's say that two advertisers are running on the same keyword with the same Quality Score:

Advertiser #1 is using manual CPC with a bid of $5.
To make it simpler, let’s assume they pay what they bid.
Advertiser #2 is running tCPA bid of $40 with average CvR of 12.5%.

Overall, it’s pretty simple - they’ll both pay $5 per click on average.

However, advertiser #2 will pay more for users who are more likely to convert. This means that a user that Google thinks has a 20% chance of converting will still have a $5 CPC bid from advertiser #1, but advertiser #2 will bid $8 and have a higher chance of getting the click (thanks to a higher Ad Rank).

A user that is only likely to convert 6% of the time would get a lower CPC bid from advertiser #2, $2.4, while, as earlier, advertiser #1 will bid $5 for them.

This means that advertiser #1 will have a competitive edge over advertiser #2 for all the users who are less likely to convert. Sucks, isn’t it?

The Biggest Earners From The Shift to Automated Bidding Strategies
So, let's look at what happens these days, where the vast majority of advertisers are competing against each other using conversion-guided bidding strategies.

Three advertisers are running on similar keywords with roughly the same tCPA bid:
In each search, Google determines the CPC bid for these advertisers based on their tCPA or tROAS and the user's likelihood to convert for each. This practically ranks the ads in order of the user's likelihood to convert for each.

So when there are multiple searches for which all advertisers are qualified to show their ads, each is bidding more competitively for users who are more likely to convert, resulting in a much higher conversion rate for all of them. This is even though they did nothing but use an automated bidding strategy. Their average CPC is most likely increased by roughly the same rate, not changing their actual CPA.

Who Earns From the Use of Automated Bidding Strategies?
The advertisers are happy: their ads target users who are likely to buy their services or products rather than waste resources like free trials and sales representatives' time, their remarketing lists are better because of the same reason, there are less users for the same amount of conversions.
Google is happy: the average CPC rises, making each search more valuable.

----------------------------
Even though I focused solely on Google Ads in this article, Microsoft is doing the exact same thing on Microsoft Advertising, so the same principle applies, just a bit less accurately and with some nuances.

Automated bidding strategies are worth trying if you haven’t given it a shot already, and if you have and it didn’t work, it’s in your best interest to figure out why. It’s true that Google is not hitting the goal 100% of the time, but it’s important to remember that it takes some time and that tCPA and tROAS are calculated over 30 days.

---------------------------
My Google rep talked me into testing broad match keywords about three years ago instead of my beloved phrase match ones. I’m sorry to say that the broad won the test by a mile. After thinking about it, it was obvious why automated bidding worked so well with the most liberal match type. It’s just using the power of machine learning and the data Google has to your advantage. And what does Google get from it? Higher CPC. Everybody wins.