Jan 27

The Spy Who Tracked Me: Online Privacy and the Algorithm Compromise

If you had to be watched wherever you go, would you rather be followed by a person or an algorithm?

Personally, I’d choose the math equation. Machine learning is the new chivalry.

Let me tell you why.

Stranger Danger

Our lives have moved online, whether we like it or not. In a connected world, people have become… shall we say… paranoid about their data.

Marketing surveys show that Americans are increasingly voicing distrust in commercial (and governmental) uses of private information.

People have major reservations with the platforms they use daily.

In response, online consumers want alternatives. A private oasis where they can frolic hither and thither, discretely shielded from the roaming eye of marketers like you or me.

DuckDuckGo provides one such alternative to Google and Bing. This search engine touts a rapidly growing space free from tracking, and people are exponentially flocking to this concept. DuckDuckGo still serves ads as an integral part of their business model, but these ads are simply contextually relevant. This works, to some extent.

Digital Matchmaker, Make Me a Match

Do you think online advertising is better or worse than 10 years ago?

Before you answer this, think back to those random pop ups that pervaded your browsers. Think back to accidentally landing on terrible websites which were completely irrelevant.

Now go ahead and answer that question, In your mind, of course.

Yes the internet is far from perfect, but online marketing continues to adjust and improve.

Relevancy is important to people, and data is the key to be relevant.

More marketing surveys show that people tend to appreciate personalized experiences.

Not everyone wants personalization though, and that’s okay. That is why opt in and opt out choices should always be presented to the consumer. 

CCPA (California Consumer Privacy Act) and GDPR (General Data Protection Regulation) laws were recently created to enforce these options and protect data security. While the digital marketing world is still reeling from these changes, these steps are actually in the right direction. Just like the internet, digital marketers must adapt. 

Love at First Site

Here’s where it gets tricky. How can we as marketers use these tools with the highest ethical standards and not be “creepy,” while providing relevant content to potential customers? At Hey, we view algorithms as the solution.

When we say algorithms, we mean marketing bid strategies that ingest tons of data in order to serve ads in the right places to the right users at the right time. This could be across Google Search, programmatic display on web content, Facebook newsfeed, or a multitude of other online placements. 

To be clear, we do not have a master list of individuals we are targeting. Instead, we feed these machine learning tools suggestions as to who our optimal user would be, and then it goes off and finds them. 

Let’s say we are selling flat brim Boston Red Sox hats. We don’t want to show ads to the whole world. We want young adults age 18-35 in the Boston area to see these ads. Preferably they are interested in sports and have even shopped on our website before. All of these online cues (browsing history, site interactions, page likes) can help us find our ideal customers. And help them find us.

Proof in the Pudding

Don’t believe us, believe the numbers.

We ran a bid strategy for our client in the internet services industry across our paid search and display campaigns. There was a clear goal: drive as many leads as possible (form fills & calls) on the smallest budget possible. Pretty straightforward goal. But behind the campaigns were an infinite amount of variables that could be toggled at will.

Fortunately, an algorithm has way more eyes, hands and time than I’d ever profess to desire. It could see things like which zip codes were driving the best results. What times of day people were most likely considering internet providers. What web pages or searches were most related to our client.

After an initial three months of learning and five months of optimizing, the results were evident.
The cost per acquisition (CPA) was down 65%!

Proof in the Pudding

Don’t believe us, believe the numbers.

We ran a bid strategy for our client in the internet services industry across our paid search and display campaigns. There was a clear goal: drive as many leads as possible (form fills & calls) on the smallest budget possible. Pretty straightforward goal. But behind the campaigns were an infinite amount of variables that could be toggled at will.

Fortunately, an algorithm has way more eyes, hands and time than I’d ever profess to desire. It could see things like which zip codes were driving the best results. What times of day people were most likely considering internet providers. What web pages or searches were most related to our client.

After an initial three months of learning and five months of optimizing, the results were evident.
The cost per acquisition (CPA) was down 65%!

And this is just one example of many. 

The algorithm saved us boatloads of budget while delivering better marketing outcomes. This frees up funds for other business gains. It also helped connect thousands of internet service customers with a company they may have never found otherwise.

A Calculated Marriage

Advertising, and the internet in general, is rife with moral conundrums AND tangible possibilities. It is our job as marketers to navigate the marriage of advertising and data without crossing too many boundaries. 

Algorithmic marketing strategies give us the ability to deliver better results while not disrespecting individual privacy too much. 

It’s not perfect, and digital marketing will definitely be different 10 years from now. But for the time being, math may be the least “creepy” compromise.