Artificial intelligence is making a lot of waves in the digital marketing space.
pickup — stock.adobe.com
Opinions expressed are solely those of the author and do not reflect the views of Rolling Stone editors or publishers.
Artificial intelligence (AI) is making a lot of waves in the digital marketing space. Companies testing ads with AI are seeing significant increases in conversions along with equally significant reductions in costs as the cost of running successful ads also drives down the cost per acquisition.
Until recently, martech AI was out of reach for most companies and, like any new technology, it’s been the subject of a lot of conjecture. While AI is smart and fast, it won’t be putting anyone out of a job in the foreseeable future. But it can make your work a lot more efficient and could make your efforts more profitable.
In a nutshell, AI consists of two main components: algorithms and data. It’s similar to how a spreadsheet can perform complex calculations for thousands of numbers, but with three major differences: The algorithms are more complex, the data is more massive and the AI can teach itself how to improve.
To improve itself, AI relies on three cognitive skills: learning, reasoning and self-correction.
.defineSlot( ‘medrec’, ‘gpt-dsk-tab-article-inbody1-uid0’ )
.setTargeting( ‘pos’, [“mid-article”,”mid”,”in-article1″,”mid-article1″] )
• Learning: With each use, it learns how to better sort data and create new rules that turn data into useful information.
• Reasoning: AI technologies use reasoning to compare algorithms and determine which ones it should use to get the most useful outcomes.
• Self-correction: AI self-corrects itself by adjusting its algorithms as needed to improve the accuracy of its results.
AI does all of this at lightning speeds, which makes it ideal for detail-oriented work that requires a lot of data — like sorting through the preferences of thousands of customers to determine what the best ad layout will be. This is something you could do yourself on a spreadsheet, but it would take you months to get the results AI can get in seconds.
AI is a powerful tool for testing ads, but it doesn’t work well unsupervised. Because it works so fast, the smallest error could quickly lead to catastrophic results. Until recently, you couldn’t use AI without having access to a small team of data scientists who could monitor its progress and tweak the model when needed.
You can’t fully automate your marketing with AI, but you can now integrate it into your current marketing efforts, regardless of what your budget may be.
A common misconception about AI is that it works like any other piece of software. There are a lot of apps that mimic AI, like chatbots that are programmed to give canned responses to specific keyword requests. These can be installed, configured and then left to run on their own with little to no problems.
True AI is a completely different technology. It learns and adapts to new data, essentially reprogramming itself, or re-teaching itself, over time. Consequently, it can learn unwanted biases and drift away from its original design, giving you incorrect results. This means that it needs to be monitored. And you’ll likely need someone qualified on hand to bring it back into line when it does start to stray.
The Rolling Stone Culture Council is an invitation-only community for Influencers, Innovators and Creatives. Do I qualify?
Anyone who does A/B ad testing knows that one small change can dramatically affect an ad’s performance. But testing each variation takes time and money, while the number of variations is practically endless.
.defineSlot( ‘medrec’, ‘gpt-dsk-tab-article-inbody2-uid1’ )
.setTargeting( ‘pos’, [“mid-article2″,”mid”,”in-article2″,”mid-article”] )
Using an AI audience model, you can do all of that testing off-stage at digital speeds. Data scientists feed the AI model with existing data on audience reactions. Once the AI has learned from that data, you can then use that information to predict how future audiences will respond to a new ad. Instead of waiting a day running small tests on platforms like Google or Facebook, AI can complete numerous tests in a matter of seconds.
This doesn’t mean you have to stop A/B testing, but any additional testing you do is more about fine-tuning a great ad. All of the major sculpting and heavy lifting has already been done by the AI.
No marketing tool should be implemented without first establishing its position in your marketing strategy. Of the companies that have embraced AI so far, there is a key distinction between those who have fully defined their AI strategies and those who have not. About 70 percent of those who have developed a strategy are excelling in their rollouts, whereas only 35 percent are doing well who have not defined their strategies.
Talk to your marketing team about their pain points. In the case of A/B tests:
• How much time does it require from your team?
• How many days or weeks does the testing run?
• What is the ad cost of this testing?
• What is the net gain in conversions between a first ad and a fully optimized ad?
Once you have these numbers, you will have a basis for determining whether or not AI can bring value to your efforts.
Most digital marketing teams love innovation. However, this doesn’t mean that AI will be automatically embraced, particularly if your employees worry that it will drastically change operations or result in job cuts.
Instead of presenting AI as a way to reduce costs, present it for what it is: an opportunity to become more effective and to either gain or retain a competitive advantage. Integrating AI and a digital marketing consultancy into your team’s environment can free their time and open them to new opportunities.
Until recently, AI was out of reach for most companies. It required significant investment to get access to the models and substantial monthly fees to monitor and update them. Maximizing AI technology can allow you to get the most out of your efforts.
.defineSlot( ‘medrec’, ‘gpt-dsk-tab-inbodyX-uid2’ )
.setTargeting( ‘pos’, [“mid”,”mid-articleX”,”in-articleX”,”mid-article”] )
Want more Rolling Stone? Sign up for our newsletter.
Have a Tip?
Alerts & Newsletters
© 2021 Penske Media Corporation