Anyone Tried Using AI To Optimize Their Campaigns?

[AmNovice]

First off, I hope everyone is well!

So, not sure if anyone has heard of using AI for ad optimization but I think it's a relatively new concept for the most part. Essentially, there's a few companies out there that have developed custom AI print tracking that incorporates machine learning to help optimize your ads and even campaigns as a whole. What comes to mind are companies like Hyros or Albert AI. I found Albert to be more professional and credible than Hyros. Why? If you go to the Hyros FAQ, there's a tonne of spelling mistakes. Very bizzare.

That being said, has anyone ran any digital marketing campaigns using AI scripts to help optimize their ads? There's big claims that it can optimize better than any human since it's AI and AI can actually sort out which of your ads are performing better than others. As you all know, having the percise info on which ad sub-groups are outperforming others is quite difficult. Even with a tonne of split testing it is hard to get such percise answers. Furthermore, the AI can also learn what time of day is most effective for your ads to be displayed. Think of it as "hyper-optimization" for your campaigns.

Lastly, has anyone used this sort of Saas as an affiliate marketer? Seems like it's mostly catered to businesses that are doing their own in-house marketing, or by agencies. I don't see why solo affiliates can't incorporate this software into their own campaigns. (provided they have the budget, of course) Anyone on here tried before?

Many thanks!

 

[T J Tutor]

The first ever AI for marketing as created in 2010 byOr Shani, Founder and CEO at Albert.

I don't know that it ever found a good footing, but remember there were some discussions in the forums about it back then.

It was bound to start coming back into the picture for us a some point, but it will be a slippery slope in some regards while being a go to tool at the same time. I think more refinement is needed presently.

 

[AmNovice]

The first ever AI for marketing as created in 2010 byOr Shani, Founder and CEO at Albert.

I don't know that it ever found a good footing, but remember there were some discussions in the forums about it back then.

It was bound to start coming back into the picture for us a some point, but it will be a slippery slope in some regards while being a go to tool at the same time. I think more refinement is needed presently.

Interesting, thanks for the response! I also learned about a company called Madgicx that does pretty much that but exclusively for FB and Google campaigns. Have you ever heard of it or used it? Seems to have overall positive reviews.

The reason why I have been keen on learning about this sort of SaaS is because I wonder if implementing it is the future. I feel like AI will get so good eventually, it would be super hard to compete with it if you're still manually optimizing. Maybe it's too early for it as you say, though.

Say I had a pretty good campaign now and I was competing against someone that has a similar campaign but is using something like Madgicx where AI can help them determine the right amount of bidding, budget, creative insights, ad copy insights, etc. I wouldn't feel too confident that I'd be able to out-optimize the AI and get better results with my campaign. So does that mean that anyone who doesn't use this kind of SaaS would eventually be at a disatvantage? Thanks agian!

 

[T J Tutor]

anyone who doesn't use this kind of SaaS would eventually be at a disatvantage?

It's always a good idea to keep up with the bleeding edge of the industry, it prepares us for what's to come and makes it quicker and easier for us to implement when the bleeding edge technologies enter the mainstreams of an industry. Acquiring a position in the forefront of the mainstream of a industry is where the smart ones implement their business practices.

 

[Honeybadger]

I don't understand the technical detail of this but the concept is very interesting and I'd like to know more

 

[Graybeard]

I think this is really machine learning mainly with the AI part perhaps matching personas or known attributes of data segments.
You would need huge amounts of advertising data and specific databases on the consumer attributes to work with.
This doesn't exist as a legitimate tool generally available to the public. I think what you are seeing is just smoke and mirrors ...

 

[T J Tutor]

I think what you are seeing is just smoke and mirrors ...

There definitely is some of that, but keep in mind brother that there are already some (not AI) automated processes already developed and deployed that seem to be getting better each year at what they do.

You and I have deeper roots than most and have a very healthy respect for time proven and honored means and methods for getting the job done. AI, and some automated processes, aren't likely to become a strong factor in most of what you and I do. Although, I know you write enough code to process some things as an auto sequence to speed up your endeavors.

 

[AmNovice]

I think this is really machine learning mainly with the AI part perhaps matching personas or known attributes of data segments.
You would need huge amounts of advertising data and specific databases on the consumer attributes to work with.
This doesn't exist as a legitimate tool generally available to the public. I think what you are seeing is just smoke and mirrors ...

Interesting. I wonder if you had a big testing budget and were able to acquire more data, if these tools would get better and better. I'd assume that's how it works. I do remember reading that one of the founders of one of those companies had to hire a data scientist to help in the development and implementation process..

 

[Graybeard]

If I do the thinking it is not AI.
Any if or else statement is not AI --it is machine code and that code author's logic.
AI is not reverse engineering an end result --that is just predictive analysis --and that is the difference.

 

[AmNovice]

Would you trust someone that is using YOUR data for THEIR PURPOSES?

Rare exception: If you are a large advertiser that can understand you data on your server you could maybe develop some form of advertising AI when used in combination with you own proprietary customer record data.

Ad networks, or traffic management, or SEO 'Gurus' claiming AI is just a red herring distraction for the real issues of fraudulent traffic and and other incompetence that, on the surface, they do very little about. In other words this all hype. They are just peddling a story.

Got you, so you don't think that a company like Madgicx can add value in campaign optimization?

 

[Honeybadger]

If I do the thinking it is not AI.
Any if or else statement is not AI --it is machine code and that code author's logic.
AI is not reverse engineering an end result --that is just predictive analysis --and that is the difference.

Love the new sigil - is that House Greybeard?

 

[Graybeard]

Would you trust someone that is using YOUR data for THEIR PURPOSES?

Rare exception: If you are a large advertiser that can understand you data on your server you could maybe develop some form of advertising AI when used in combination with you own proprietary customer record data.

Ad networks, or traffic management, or SEO 'Gurus' claiming AI is just a red herring distraction for the real issues of fraudulent traffic and and other incompetence that, on the surface, they do very little about. In other words this all hype. They are just peddling a story.

 

[Kamari Cart]

If everyone use it , Then it is Difficult!!

 

[VortexAlpha]

I haven't used too many AI generators except Jarvis for some content writing automatization, but I never liked the idea of only relying on AI no matter if you're using it for writing or for optimizing ads and campaigns. These tools can be really useful and helpful but only if you use them the right way, as if you leave all the work to them and don't check things manually, it can turn out that you placed online something that doesn't look right. So, yes, AI can be great but only if you put your own effort as well.

 

[petrderbikov]

First off, I hope everyone is well!

So, not sure if anyone has heard of using AI for ad optimization but I think it's a relatively new concept for the most part. Essentially, there's a few companies out there that have developed custom AI print tracking that incorporates machine learning to help optimize your ads and even campaigns as a whole. What comes to mind are companies like Hyros or Albert AI. I found Albert to be more professional and credible than Hyros. Why? If you go to the Hyros FAQ, there's a tonne of spelling mistakes. Very bizzare.

That being said, has anyone ran any digital marketing campaigns using AI scripts to help optimize their ads? There's big claims that it can optimize better than any human since it's AI and AI can actually sort out which of your ads are performing better than others. As you all know, having the percise info on which ad sub-groups are outperforming others is quite difficult. Even with a tonne of split testing it is hard to get such percise answers. Furthermore, the AI can also learn what time of day is most effective for your ads to be displayed. Think of it as "hyper-optimization" for your campaigns.

Lastly, has anyone used this sort of Saas as an affiliate marketer? Seems like it's mostly catered to businesses that are doing their own in-house marketing, or by agencies. I don't see why solo affiliates can't incorporate this software into their own campaigns. (provided they have the budget, of course) Anyone on here tried before?

Many thanks!

To be honest I don't know anything in this area and I'm not into that sort of thing, but I have a friend who is very knowledgeable about it, if you still need I can contact her and give you an answer!

 

[Graybeard]

Cthulhu Brigade

Got you, so you don't think that a company like Madgicx can add value in campaign optimization?

I really don't know --but it takes more than claims on some webpage to impress me.
I would have to see a sample of the real data they are using --to make any definitive statement beyond 'skepticism'

 

[Gamerseo]

Rather top too soon for such things.

 

[Graybeard]

Real AI is going to be costly.
It will be possible to scale.

Right now it appears to me that AI is still learning
the database content is being crowdsourced
--that why is free or low cost at the moment

The language of innovation: How AI is improving the translation experience for marketers

Having a translation partner leverage AI to improve quality can make a real difference.

www.marketingdive.com

This is the top of the food chain and this is why AI won't be cheap enough for 'everyone' to use in competition functionally.
Here is an ETF of people in Quantum Computing. Quantum computing is what real AI will be using in the near future

 

[Graybeard]

A lot of people are not going to like AI when it makes them obsolete.
I used an AI paraphrasing tool the other day;

What I did was:

1. excerpt the text content from the top SERP position websites to my query
2. aggregate the test in the same order that the sentences were scraped
3. fixed the '(ticks) and other characters that would be problematic
4. removed any brand quotation that would be of no benefit ex;(replacing example.com with a generic theSite)
5. API posted the text and cURL and printed the reply to a file.

I then edited the text a bit and started to plagiarism check the result
As I edited I achieved a 81% to 100% unique content. Using 4 online tools and varying text ...

So, yes some human editing will be needed and this has to be test for SEO and reedited to be used as a CTA type landing page format.
Not fair to assess the time expended as this is a learning curve working with the API also.

I created a document and edited it

In a 'short time' if this was not a learning curve test
Denial is not a river in Egypt.

As far as optimizing advertising campaigns:

• Most of what I have seen is really sophisticated machine learning
  • if it is to be called AI
  • it is Monkey AI.
• Just binary conditionals based on results and not predictive at all, totally retrospective.

 

[Graybeard]

Give me enough ad campaign / result data all the way through the funnel --I have, or could create, datasets to make conclusions and predictions. So, with (ML) Machine learning and AI the algorithms *could be automated*

There are businesses spending hundreds of thousands a month, or in some cases over a million per month --I am sure they have noted this too and are internally developing this for their own use.

If you developed a "secret weapon" that would give you a market advantage, you would not make it public. Eventually, it will leak out and be filtered down. Just like the nuclear bomb did.