How hard is creating ads on Facebook, Google and Instagram for you?

Do you do it?

I am asking because we are developing an AI-based tool which makes online advertising extremely simple for small business owners. So we are thankful for any market response, feedback or any words at all :)



Hi peter,

By AI I assume you mean Artificial Intelligence, which can mean a lot (neural networks, inference engine, machine learning, natural language processing, etc)

Can you explain a little more what you mean by AI-based tool, and how it can be utilized to be better marketers?

I'm 6 months from my launch; so, I can't give you my feedback yet.
But, I recently thought about using ML against existing customer database to identify customer pattern/segments, which can make me more effective in marketing campaigns.
Last edited:
Hi Edvin,

Yes, it is an artificial intelligence based tool. We are using it to create online advertising ads on Facebook, Google and Instagram. I am not a machine learning expert so I can't really comment on how it works on the tech level. However, it makes online advertising very simple and AI can manage campaigns automatically, so it is a hands-off solution, very convenient for people who want to be effective with online marketing and they even save money on PPC experts.

Your application sounds very interesting too. Is it aimed at a database which is used in email marketing?


Hi Peter, AI driven ad engine sounds very cool.

Take a look at @zyrakate pain point, which is not knowing the ideal bidding budget/limit. A tool that can adjust bids based on numerous factors (conversion rate, customer lifetime value, time of day, sudden change in conversion, etc) can be valuable.

Another challenge marketers face is to come-up with appropriate search terms. Third-party providers like SpyFu can help marketers come-up with key terms; but, that also means that we will be bidding for premium terms. There are strategies and tools for finding long-tail keywords; but, I'm not familiar with them.

As for my previous post ...
I was planning on investigating Azure ML because it has a suite of tools for both an analyst and developers (for SaaS API). Marketers/analysts can generate and validate a mathematical model, which can be used check prospective customers compatibility.
My plan with ML is to be more effective with my PPC campaigns by having a better understanding of my customers. For example, I know my future competitor started targeting women between the ages of 25 and 44 for tutoring service; thereby, reducing their Cost-Per-Acquisition (CPA) by 25%. Was that a hunch to start targeting this demographics? Are there other parameters that should be considered? This is where I want to use ML to get a better understanding of my customers.

With better customer insight, we can
  1. Narrow PPC campaign based on customer segment and increase the bid for high conversion profiles.
  2. Widen PPC to demographics that you would normally exclude
  3. Customer insights can make print campaign mailing a cost-effective strategy.
  4. I haven't considered email campaign; but, it depends on its cost (lead acquisition or email broadcast)

I think a really neat use case is going beyond our own customer data...
Lets assume that using ML we have identified high conversion customer profile for a tutoring company.
A user searches google for the keyword "K-12 Tutoring", which we would normally want to bid-for it. But, what if I learn that the visitor IP is originating from a school, or that the user household income is below my customer profiles; obviously, the likelihood of customer conversion in such cases is low.
But lets assume that the user matches our desired customer profile and is originating from non-academic institution. However, the user recent activities related to "tutoring business plan". Again, if someone is reading about tutoring business plan, then they are unlikely to become our customer for a tutoring company. Third-party vendors like Demandbase have data about visitor income and institution; Data Management Platform (DMP) vendors like Liveramp can provide visitor internet activity information.
Admittedly, these vendors cannot be inclusive of all internet users and their activities; none-the-less, the data that they do have for visitors provides another decision point for a smart system. This is nothing new; the novel idea is to use AI to determine best converting customers and adjust bids accordingly.
Last edited:
Top Bottom