Year-end pieces tend to be a bit hyperbolic, complete with splashy predictions for the next big things that will take over the market.
In reality, tech tends to advance at a more measured pace, with improvements stacking on top of each other until a technology delivers on its initial promise.
The potential to decode social media data with AI has intrigued marketing departments for years, just as its potential pitfalls have kept us awake at night. The problem, of course, is there’s just so much of it, and it’s so unstructured.
The analogy I like to use for structured data is Lego — organized, consistent, fits well together — whereas social media’s unstructured data is like Play-Doh — a complex blob of information that’s variable, inconsistent and unpredictable.
For our 2018 predictions, we wanted to focus on this confluence and deliver marketers some positive news: While this may not be the year that the “Facebook of AI” is invented, the progress will be very real and, more importantly, materially helpful to marketers.
These predictions will highlight four situations that continue to keep marketers up at night but will be put to rest by AI in 2018.
Identifying relevant posts and consumer insights
Situation: You’re overwhelmed by social media posts about your brand, as well as posts with relevant keywords that actually have nothing to do with your brand or message.
As a result, you end up with an ineffective social media presence.
AI has allowed us to comb through tweets and other posts like never before, but that can come with a cost.
If you’re a marketer for the University of Alabama Crimson Tide, for example, trying to understand what people are saying about your brand on social media could uncover posts about Crimson Tide the movie, or even red tide, the algal bloom.
Sorting through all these posts takes time and can result in lost opportunities if a good piece of content is missed in the flood.
It’s simply impossible to sort through every potentially relevant social media post by hand, though, no matter how many people you have on the task. AI has been able to help for some time, but in 2018, the technology will make it possible to make informed calculations and present you with only the most relevant posts to interact with.
In other words, it will only pull Crimson Tide references about football, not shellfish beds being closed.
Dealing with images
Situation: A user posts a great shot of your product being used in a new way you didn’t expect, but you miss it.
An increasingly important part of this mashed up Play-Doh ball of social media data is images. Instagram reports its users alone post 95 million videos and photos every day.
Even with tagging and other sorting mechanisms, there’s simply no way for a human to go through and learn from all of the user-generated content for a brand.
A classic example is the hula hoop. These toys were traditionally geared toward children, but with image analysis of photos shared on social media, marketers noticed pictures of people using them at music festivals.
This opened up a new audience to target. With the sheer amount of images being posted every minute, it would be very difficult for a person to notice an opportunity like this today on social media without the help of AI.
Recognizing real moments of consumption in real time is an area where AI will make progress in 2018. More than that, however, AI will become intelligent enough to recognize new cross-promotional opportunities and potential sales opportunities and present consumer insights to you accurately.
With that information, your marketing department can then do what it does best: capture these potential opportunities with targeted content.
Making the most of influencers
Situation: You pay someone on social media to promote your product, but this influencer just doesn’t resonate with your audience.
Social media influencers have become increasingly important to marketers. However, that doesn’t change the fact that trusting your brand in the hands of someone you only know through social media — and maybe emails or a phone conversation — is scary.
Just as influencers can help you gain visibility, they can hurt your brand or fail to move the needle at all.
Matching your brand to an influencer has long been a manual process. Maybe word-of-mouth leads you to discover someone. You then do as much due diligence as possible by reviewing their social media posts and conversing with them as much as you can.
In the end, there is a significant amount of time and effort involved and too much blind trust for a marketer to be comfortable.
As AI improves, finding the perfect influencers to represent your brand will become a much more accurate, less time-consuming process. Algorithms are already using these influencers’ posts, followers, interactions and other critical data and comparing it to business needs. This allows — and will continue to allow — marketers to replace blind trust with well-informed choices.
Posting on time
Situation: You have a great post lined up, but when you send it out, it doesn’t get the engagement you expected.
To close, let’s look at what seems like a simple problem: when and how to distribute your message, on social media, via email and through other channels.
As the number and diversity of channels has increased, so has the level of difficulty in choosing where and when to deliver your message.
With the sheer number of social media posts, it’s only natural that most get lost in the shuffle. If you post your message at the wrong time, its lifecycle could literally be seconds — just enough time to be pushed off the bottom of a user’s screen. If you pick the right time, though, your chances of going viral go up considerably.
This has long been an area where AI’s potential benefits have been clear. It can go through data on the reach of posts and when and where they were posted to come up with a pretty optimal strategy.
In 2018, that advance continues, integrating channels, media and interactivity more seamlessly than ever.
With all these examples, a human decision still needs to be made as a final step. In fact, the purpose of AI in marketing has never been to eliminate the human element altogether.
The goal is to parse the overwhelming amount of data available and to present marketers with the right information to make the best decision. These steps will bring marketers closer to the ultimate goal of harnessing consumer insights through AI.
In 2018, this probably won’t mean a bolt-of-lightning breakthrough that changes the face of marketing forever, but it will mean many smaller improvements that together make social media much more valuable for marketers everywhere.