AI-Driven Digital’s Human-First Strategy Targets Consumer Behavior vs. Bias


October  31,  2018 by Alyson O’Mahoney , Managing Director in Artificial Intelligence (AI) digital advertising

I’m a sucker for Instagram ads for shoes and home décor, in particular. You can expose me to these products or brands within display, YouTube or Facebook ads, and I won’t typically click or buy, but for some reason, when I see a product in its Insta-glory, I’m whipping out the credit card online or looking for it at brick and mortar within a week.  This isn’t because these brands don’t have digital impact on me in other platforms – I may be accumulating the number of impressions I personally need to be pushed to sale (it’s estimated that you need around 23 impressions before you are driven to actually act!) – but I spend more thoughtful (free) time perusing Instagram in shopping mode, so this is where I ultimately make my purchase decision. For you, it may be different altogether.

That’s why I mention this.  Not because I think every brand needs an Instagram strategy, but because EVERY brand needs a consumer strategy that eliminates bias.  It’s no longer enough to consider a demographic, a geo-target, a retail shopper or a keyword.  It’s not enough to pick one or two digital channels, and put all your eggs in those baskets. Today’s marketing needs to not only address the needs of the brand (sales), but the preferred action of the potential shopper in a channel which can vary widely even among those with similar demographic data. Enter AI-driven digital delivery.

Artificial intelligence (AI) is one of those current buzzwords that conjure up those “Oz behind the curtain” or “computers taking over the world” comparisons.  Is it purely smoke and mirrors? Frankly, we thought it might be when it came to executing digital delivery.  However, after several years of watching programmatic digital drive lots of website traffic or social ads getting great reach, both with underwhelming conversion, we honestly felt there was nothing to lose to test out the AI hypothesis.

While we entered into it as cynics, the results quickly changed our minds. The biggest benefit was not that we could market across all major digital channels (display, search, social and video) simultaneously, or that the speed of computer media buying decisions was hundreds of thousands of times faster than human optimization, or even that we were able to A/B/C/D (and so on) test creative in all channels, including search, in real time – yes, we had all those things – but it also removed the personal bias marketers place on their campaigns and ACTUALLY SOLD PRODUCT.  Not like, maybe sold some product, it actually sold product in a significant way that we could measure against client sales data at retail.  Sales increased when it was on (often times by double digits even in lagging categories) and didn’t grow when it was turned off for a bit (now most of our clients are not turning it off for long, if at all). These increases were significant against the brand sales, the category and the budget the brand was spending. The return on ad spend (ROAS) was mea-sur-able!

By tapping the algorithms available to us through our AI digital platform, we focus on brand message delivery based on a variety of factors, including speed, targeting accuracy, and, most importantly, human behavior in the digital environment (I convert on Instagram, you convert on display). This allows us to be more efficient with the targeting and the spend (so much less waste).  The best ad concepts and calls to action rise to the top based on consumer vote (clicks/conversions), not based on the one concept we all voted on in the board room.

So now, when someone asks us for a digital campaign, we say, you need a “people” campaign.