There is a fundamental shift occurring in the world of digital marketing right now. Since the beginning of advertising, marketers have tried to understand who their target audiences are and what they respond to. First in person focus groups provided insight to consumer sentiment. Next digital came about with individual user data and A/B tests dominated the field. The next iteration of ad technology going on today is “hypertargeting” or targeting the “audience of one.” The five social media marketing trends listed below show how machine learning and artificial intelligence tools allow marketers more ability to target and create messaging to individuals.
1. Machine written copy
Why it’s important? Saves time. Better targets. Uses individual voice.
Just as every brand has their own voice, every viewer has their own voice too. In 2017 there were 3.77 billion internet users. That means there are 3.77 billion people online that all talk a little bit differently. Back in the days of A/B testing a marketer would try two different messages targeting the same audience and see which performed better. The better performing copy would remain, and the worse performing would be change based upon a well-researched hypothesis of what might work better.
Persado is the next generation in copywrite testing — particularly for social media. Using natural language processing Persado AI is able to learn the voice of a brand and learn the voice of each target in an audience and automatically write copy to that individual. This is much more time and cost efficient than the alternative. Hiring hundreds of human copywriters to research each possible audience member and write to each of them individually. An AI can do this in seconds.
According to their website Persado’s copy outperforms human created copy for audience engagement 100% of the time. They claim that marketers using their tool see a 49.5% lift in conversion across web, email, mobile, display, and FB.
It makes one wonder whether the role of copywriter might be soon replaced by AI. It also makes one wonder about how magical web experiences might be if every interaction was custom tailored to the individual.
2. Photo recognition and tagging
Why it’s important? Quickly select photos. Saves time.
The idea of computer vision began as a summer class assignment at MIT in 1966. The assignment was to make it so a computer could use visual references to make decisions. One of the first large scale applications was the bar code reader. Recently facial recognition has become a global standard — Facebook can tag faces and snapchat can add augmented rabbit ears.
If marketers are to customize their messaging to the audience of one, a greater range of pictures will have to be made available. Our brains are primordial organs and we are all subject to “in-group bias.” We all strike a connection faster with something our brains consider to part of our “in-group.” Persado does this with written communication Clarifai might be the key to doing this with photos.
Clarifai has developed an advanced photo and video tagging program. Their demo is quite breathtaking. Upload a photo or video and Clarifai tags the asset with demographics, emotions, and suggestions for creative campaigns. Soon it is likely that marketers will use a tool like Clarifai in conjunction with a tool like Persado to allow machines to create hypertargeted text/photo/video ads to the individual level.
3. Influencer to brand matching
Why it’s important? WOM is best sales. Access to loyal audiences.
Influencer marketing is one of the most talked about strategies of 2016 and 2017. It is particularly effective on social channels. Unfortunately one of the big downsides of influencer marketing is the risk that the audience might not see the influencer’s promotion as genuine. Long-tail bloggers and influencers are seen as more genuine, because their audiences, while smaller, are stronger.
But how does a busy marketer go about finding lesser known influencers that would match their brand? Soulmates is working on solving this problem. One of the founders is from eHarmony, and in a very eHarmony way Soulmates matches influencers to brands. Soulmates is able to analyze the conversations influencers are having and match those combinations to brands looking to take part in such conversations.
If companies like Soulmates are successful this could lead to more smaller and “startup” bloggers being matched to brands that are genuinely on point with their lifestyle. This would mean that more people could quit their day jobs and pursue their passions thanks to the sponsorship of brands who want to be part of that passion.
4. Social media chatbots
Why it’s important? Saves time. Runs leads through funnel.
As more traffic is generated on social media channels, both brands and influencers are going to have to rely on chatbots to answer simple questions and keep audiences engaged. It will not be possible for human social media managers to personally reply to each engagement point; however, marketers know that whenever a person engages and receive a reply their brand loyalty is strengthened.
The good news is that millennials are comfortable interacting with chatbots. 70% report a positive experience with chatbots and 67% said they would be willing to make a purchase via a chatbot. Your audience wants to connect with your brand in as easy a way as possible, for most this is social media. You can program a chatbot to answer questions about inventory in-stock, pricing, related products, all shown to the customer on social media and linking to a specific landing page.
Chatfuel is a chatbot for Facebook messenger. They claim you can build an AI chatbot in as little as 7 minutes. You set up conversational rules and the bot recognizes similar phrases a user might say instead. They are lead through the conversation path you program the bot to have with them.
5. Personal AI for filtering
Why it’s important? Better matches user to products.
If you have been online in the last five years, you are aware that there is an onslaught of information. Both Google and Facebook have done an excellent job of using machine learning to understand individuals’ behaviors and interests. This data is used to determine whether you are a match for a potential ad campaign, because an advertiser doesn’t want to waste money showing you and ad if the likelihood that you’ll buy is low.
The next step for Google and Facebook is to create personal AI bots that learn your behavior and communicate with advertiser bots about what you like and are looking for. This would maximize efficiencies in ad exchange networks. Facebook has already released M for its messenger, which gives users suggestions based on context. And Google released Allo, which allows you to respond to a conversation with just a tap as it gets to know you and your behavior.
It is very cool to think that in the near future, each individual user of Facebook and Google will have a digital “twin” that represents them online by finding the resources you need when you need them. And as you use it more, the more it is accurate in its predictions.
6. Augmented Influencers
Why it’s important? Only time will tell.
Augmented influencers is the sixth category in a five category post because the stuff talked about here is so wild it’s more sci-fi for the time being. Unlike the five trends listed above, this one will likely go through a few legal battles before they can start being used.
Could you imagine controlling Donald Trump’s face and voice? How about Hillary’s? How about Rihanna’s? The technology now exists to do that. Not their face in real life, but their media persona. Computer scientists at Stanford have made a system where a webcam watches how you move your face and moves the on screen target’s face the same way. Another company, Lyrebird.ai, analyzes the voice patterns of a target and allows then creates speech using their voice. While Lyrebird still has a ways to go to perfect the speech, the concept is amazing.
Can you imagine how an influencer could allow a brand to literally put words in their mouth? While the public would support using chatbots, how would they feel about always questioning whether a video is the fact or a creation?
The internet is getting much more interesting.
*Articles on this blog contain affiliate links that GMP receives a small commission from at no cost to you. We only recommend tools we trust. And never ones we don't trust.