LensAI: Associative Advertising as An Inevitable Evil | Hacker Noon

@pavel-shkliaevPavel Shkliaev

I grew up on such books as «Roadside Picnic» by the Strugatsky Brothers and «The Inhabited Island»

“Association in psychology refers to a mental connection between concepts, events, or mental states that usually stems from specific experiences.” [1] Once the associative link between events A and B has been built, the appearance of event A naturally entails the appearance of event B. [2]

Cool, huh? Now, imagine a scenario in which a machine (AI) understands what is happening on the user’s screen (while he/she is surfing through online content) and then uses that information to show brand awareness ads designed to implant an apparent associative connection into the user’s brain.

Hey everyone, I’m Pavel, the CEO & Founder of LensAI. We build AI-powered advertising technology that performs in-depth aggregate analysis of visual, textual, and audio contents to enable our unique algorithms to select and then embed relevant ads into visual content automatically. Just a quick but crucial note, there is no-preprocessing needed, and everything is happening in under a sec once the user comes across the visual content.

Let’s start with the questions that bother me the most:

  • How vital is non-stop complex content analysis for Advertisers?
  • How valuable is an opportunity for ads to promote brand awareness while simultaneously being converted into real $ by facilitating impulse purchasing?
  • How necessary to provide Advertisers with a tool/service that allows careful curation and filtration of circumstances under which to show ads for each brand/product?

I am a tech guy, so maybe it will sound not very clear to you, but I will try my best to explain. Right now, our technology extracts separate frames from the video to analyze. For example, while the user is watching a reality show, the technology counts specific objects, environments, events, plus other important elements, and then it tracks the duration of their presence in the video along with the appearance of particular classes or IAB categorization.

The data we get helps us design formulas that bring value to any brand as they allow the machine to get the highest level of understanding of video content and how it satisfies the needs of, let’s say, the Veet brand (for those who are not familiar with a brand, it is hair removal products).

We have created a technology that does a superior valuation of content based on made-to-order AI formulas: an exclusive way of setting circumstances (e.g., type of entities, environments, words pronounced, and other various parameters) under which to show ads for a specific brand.

We ask our Advertisers to express their “wants” for each brand or even individual product. Then, we put them into action by enabling our algorithm to monitor particular objects, actions, events, and every critical piece of information. LensAI will never stop to enhance the machine’s understanding of how different video content fits each brand/product.

See the results of implementing an exclusive formula for the Enfamil brand below. From now on, LensAI can build a valuable connection between any brand/product and the relevant content. Again, no manual work is involved as a machine does everything within milliseconds once it sees that the user has landed on the page of our Publishers. The speed of the whole process is mind-blowing and one of the gems of LensAI.

The formula incorporates multiple pieces of data that have to be present to initiate advertising.

Let’s take a look at a few examples that demonstrate how LensAI works from an end-user perspective. We used visual content currently available online for demo purposes:

Due to the short length of the videos, we set the Ad Display Setting to “Highest Frequency” to demonstrate technology possibilities.

Now let’s watch the same video to observe how the machine inserted completely different ads into it. Why is that? What made it change its mind? That’s what happens when we make our machine respect each viewer’s individual characteristics based on the IAB Audience.

These two examples prove our machine’s ability to select various objects and use them for building associative connections with different brands and their products. Demography, interests, previous purchases, and other additional data about users influence the machine’s decisions about selecting the most relevant ad, its placement, and advertising purposes it should serve.

I have mentioned above that LensAI can build a valuable connection between any brand/product and the relevant content while the user is surfing on the page or when he/she hits the play button. We didn’t stop there and added another precious piece of information – specific data about each user – to allow AI to enhance targeted advertising.

Your Questions

Here are the top three fair comments/questions/concerns that our followers expressed:

“I saw a total of five ads for Lysol in your videos. The ads were shown in places where no Lysol disinfectant spray was present. That’s exactly what other ad networks do: show ads of anything that has to be sold everywhere and anywhere. So, what differentiates you from them? If a beautiful cover does not lead directly to sales, it has close to zero value. So, why should I choose your product? Why should I use your software?”

“I was hoping you could explain the benefits for me as a potential client rather than a consumer who is the last link in the advertising chain. Make it simple. Suppose I have pictures and videos on the site and participate in the affiliate programs (ranging from marketplaces like Amazon to affiliate marketing service providers like Rakuten Marketing). Why do I switch to you? What are the distinctive advantages you have over your competitors?”

“I have a question from the end-user perspective. What if I see some item of interest in the video, but the Publisher who partnered with you does not market that product (let’s say, it’s a dining table). Suppose I keep seeing Lysol ads embedded into that item and, even worse, other ads embedded into other objects according to the same scenario. In that case, I will stop paying attention to your ads due to seeing “wrong” ads. If users like me were able to build an immune system for banner ads, we could build one for LensAI ads too.”

My Answers

I love questions like these and take any criticism with appreciation. So, here are my answers to those who reached out to me personally and all of you who might have similar concerns.

Everything you currently see is a basic framework. We currently limit inventory that we use for analysis to cut computational resources expenses and choose short videos for demos for the same reasons. The high density of ads that you witness now is not pleasant even for me. The final product will enable Publishers to control it and adjust what is needed based on Analytical Data we make available to them (e.g., % of users who watched the video to the end).

Let me speak the truth. Yes, we work hellishly hard to ensure the most accurate matching of identified objects to products available to advertise. We do want users to enjoy the interactive advertising experience with the eCommerce touch we offer. Even though we consider all stakeholders’ interests, the software is being developed to precisely satisfy the needs of Advertisers and Publishers.

Regarding the question about our competitors, please study the slide below.

Participation in various affiliate programs does not allow Publishers to market products through image and video content available on their sites. No one but LensAI can enable them to do so.

None of our competitors is technologically capable of selecting the fittest product to advertise from numerous different product lines and categories among several marketplaces. That’s why our competitors by technology cannot sufficiently monetize traffic; they cannot cover a large volume of goods and, therefore, cannot find the ideal goods from advertising inventory to display (feel free to read about it here).

Impressions equal money; that’s clearly why they are important to Publishers. Each of our technological competitors takes away an opportunity for Publishers to dramatically increase their revenue by their inability to match advertising inventory to the graphic content that Publishers posted, post, and will post.

Suppose our machine cannot find a perfect product to advertise among products available on the marketplaces we affiliate with. In that case, we will promote the relevant item that satisfies either cross-sell or up-sell requirements (we wrote about it here). Meaning, Publishers will never lose an opportunity to make $ while users will never question why this particular ad pops up. Relevancy and similarity are the keys to success. Don’t forget about our brand associations formulas we do for brand awareness ads.

Here’s quite an upsetting message for the users. The dot format is simply one of the numerous designs and UX options. We take ad blindness seriously, so we made plans A, B, and Z to fight any derivative form of ad blindness syndrome. Maybe I should write another article about the advertising standard that we developed as it allows superimposing any UI/UX as easy as one-two-three. Imagination is the only thing that can stop us, and we all know it is limitless.

Long story short, no technologies and products that are known today can operate according to this scheme: the installation of the script on the website/app within minutes-> the extraction of the content in real-time (once the user lands on the page) -> the detection of objects and matching them to any possible product available for sale ->  embedding the fittest ad to re-sell the product for a commission -> if no similar product is found, either a relevant brand ad or a relevant product ad pops up. Easy-peasy!

Nothing is more upsetting than seeing someone do something you thought was impossible!

Create brand formulas with us! Invest in us! We are open to any collaborations, so let’s do something novel together!

For those who want to dive in, read the previous articles I wrote. The truth is somewhere near.


Join Hacker Noon

Create your free account to unlock your custom reading experience.

read original article here