A.I.mpact, Part 2: How A.I. Will Affect Marketing In 2017 & Beyond
Before we dig into the developments you’ll likely see in 2017 as artificial intelligence continues to transform marketing and advertising, let’s cast a glance backward.
For old-school ad execs and creative teams, here’s the moment where one of their most hated enemies first showed its blinking, beeping, buzzing face, as beautifully depicted on Mad Men:
Don Draper may have had his qualms about the IBM 360. But this was the Space Age. And the arrival of the computer and data-driven analytics rocketed agencies and brands out of antiquated approaches. Much to the consternation of those clinging to a more seat-of-the-(grey flannel)-pants philosophy.
In any revolution, there are leaders and losers. Right now, there’s another epochal transformation happening, one that would really muss Don’s Brylcreem.
Artificial intelligence is already dramatically impacting most present-day marketing practices, and that evolution will turn into a tidal wave in 2017, overwhelming anyone who’s not ready.
Follow the ‘smart’ money
When investor cash blows in a certain direction, it’s a good sign which way the winds of change are going. It doesn’t take much native intelligence to realize an increase of 300% in AI investment is a pretty big bump, a predictor we’ve cited before.
Forrester Research isn’t reaching when they project how artificial intelligence will be a hot investment category in 2017. In 2015, this Bloomberg article went into great detail about how investors were already pressing their Gucci-shod feet to the floor in accelerating the funding of AI startups.
But really? A jump of 300% in a single year? It shows how AI has been anointed the Next Big Thing in tech investment. Where’s that money going, specifically?
This heatmap of corporate investment in A.I. over the past five years tells the story. Healthcare is leading the race, but check out the runner-up.
Why do marketers need AI ASAP?
The benefits of integrating AI and machine learning into a marketing program are legion. But the best macro argument for adopting AI might be summed up in a single graph from Forrester.
As they see it, businesses that incorporate AI to uncover actionable insights “will steal $1.2 trillion per annum from their less informed peers by 2020.”
If that’s not competitive pressure, what is? Laggards will find themselves shedding business to AI-equipped rivals who are able to give consumers and customers (in B2C and B2B alike) what they crave the most: insights-driven personalization at every phase of engagement.
This eventuality is what’s compelling corporations and investors to suddenly pivot toward investing in AI today, so they can hold onto market share tomorrow.
But where does AI already have traction in marketing and advertising today, and what’s it projected to do tomorrow?
“Cognitive technology is there to extend and amplify human expertise, not replace it.” Rob High, Chief Technology Officer, IBM Watson
In 2017, AI accelerates
Marketing has always owned a unique advantage over other parts of an enterprise, such as HR and logistics, when it comes to adopting AI: it’s in the business of constantly generating large volumes of data, whether in the form of clicks, views, email responses, or information from other touch points. So it simply makes sense to put AI to work there first.
AI has taken on the mantle that once belonged to the “expert systems” developed in the 1980s, designed to emulate the decision-making expertise of human experts. AI can now utilize data to support brand managers and marketing departments in analyzing their brand’s attributes, channels, competitors, targets, pricing and budgets. But beyond the C-suite, AI has also gotten woven into tactical execution at every level.
In 2017, AI will get smarter and more cost-effective, accessible via SaaS, mobile apps or IoT embeds. So it’ll not only become more sophisticated but more pervasive, available to marketers at every tier in every sector who’ll use AI to tackle a huge array of marketing needs.
Analytics, insights & forecasting
We’re at the point where there’s an enormous amount of data available to practically every marketer who wants it, whether they’ve culled it themselves or secured it from external sources. The hurdle, of course, is the enormity of these data pools, and the limits of human beings when it comes to making sense of them.
Mining Big Data for its potential value mandates using artificial intelligence to extract its riches. Machine learning can pull insights out of sprawling agglomerations of information, moving “business intelligence” into a whole new sphere.
With over 2,000 marketing technology vendors out there already, most of them claiming some focus on analytics, an ability to deliver or integrate with AI systems is going to separate the wheat from the chaff in the eyes of enterprise tech buyers in 2017 and beyond.
Here are only a few examples of where AI has already been a player in predictive analytics, and where it’s headed in the next few years:
- In the airline industry, neural networks such as the Airline Marketing Assistant/Tactician have been used for decades to forecast passenger demand, guiding seat allocation and ticket pricing.
- Spiegel used pattern classification via a neural network to fine-tune its marketing decisions. By analyzing lists of people who had made just one purchase from a Spiegel catalog, using demographic information on hand about them, it could recognize patterns that identified the people most likely to be repeat purchasers.
- In a notoriously fickle category like fashion marketing, companies like TechStyle are now using predictive analytics engines to minimize guesswork about style trends and deliver increasingly personalized products to consumers.
- Salesforces’s Einstein platform is helping marketing and sales teams predict which email leads are more likely to convert, then recommends the best followup actions. It also advises on campaign spends and timing; when ShopAtHome.com used Einstein, it saw a 23% lift in email clicks and a 30% increase in opens.
AI innovators have been finding ways to apply machine learning against key marketing challenges, and that’ll gather steam in 2017. One of those challenges? Making account-based marketing (ABM) a reality.
There are numerous shortcomings with present-day lead generation and marketing automation models that prevent ABM actualization, starting with the problem of identifying quality leads. Yet the B2B prospecting adage that 95% of leads are misses needn’t hold true anymore, thanks to AI.
A Deep Learning platform (like the products developed by MarianaIQ) can analyze thousands of inputs about demography, psychography, social media behaviors and a raft of other data to create hyper-accurate personas, constantly updated in real time. That enables marketers to find and engage only the best prospects within the accounts they want to target, achieving 1:1 personalization with each of them.
Since AI cost-effectively automates much of the leadgen and engagement process, tomorrow’s ABM programs will be implemented at scale, across all of a marketer’s accounts, by more and more companies.
Bots and AI marketing assistants
AI-powered bots and chatbots will grow in application and diversity for sales and marketing. Facebook’s new Send/Receive API for Messenger allows bots to respond to customer queries with structured messages that include images, links and call-to-action buttons.
So users could potentially use Facebook’s bot to book a flight, reserve a dinner spot, fulfill an e-commerce order or pay for purchases, or any other task addressed by third-party developers. Facial recognition will soon allow these AIs to recognize us on sight, and adjust interactions accordingly.
Chatbots are one thing, but putting sophisticated AI “on the payroll”, so to speak, is another sign of things to come. An offering like Shopify’s Kit, which provides an AI marketing assistant to help Shopify customers market and promote their online stores, foreshadows how AI will take on other marketing roles in the years ahead.
The tasks Kit handles include creating Facebook and Instagram ads, posting Facebook updates to drive customer engagement, sending personalized “thank you” messages to generate repeat purchases, promoting new or back-in-stock items, and outputting sales performance reports.
Some marketing departments can’t get that much work out of a human employee, frankly.
By 2020, smart agents will manage 40% of mobile interactions. Source: Gartner
Marketing automation was ripe for the arrival of AI, and content creation and management tools such as Marketing.AI and ClearVoice, among others, are already making content publication more efficient. That’ll expand in 2017 and beyond, as AI fine-tunes inbound marketing programs in real time to capitalize on audience hot buttons, identify key influencers and promote content more efficiently.
AI will also keep evolving its ability to actually author content, creating material that reads naturally, not like someone’s been taking dictation from Robby the Robot.
News media such as the LA Times, Forbes and the Associated Press have already been using AI tools to produce fact-based articles, summaries and reports on a basic level. However, companies like like Articoolo and Quill have developed AI platforms for writing content on a more sophisticated level. As Quill’s website puts it, organizations are increasingly using“advanced natural language generation (Advanced NLG) to transform their data into narratives.”
Content writers may dread this development, but the upside is that whatever a machine writes can’t be any worse than most of the crappy content already glutting the web. You can’t pin that on AI.
AI will replace 16% of American jobs by the end of the decade. Source: Forrester
Customer journey mapping
Delivering a seamless customer experience is job #1 for digital marketers. And that seamlessness relies on predicting a customer’s behavior, so every step along their path can be mapped out and made personalized and consistent.
Up till now, this depended on assigning a static persona to each customer. Those were tied
to pre-mapped journeys, where offers and messages were designed to match up against the specified persona.
Does it work? Meh. People have a knack for going off in unexpected directions when their path offers options. A customer labeled as Persona A may branch off in a direction intended for Persona B, for instance.
By being able to continually process huge amounts of customer data, Deep Learning systems can create way more granular, more accurately-predictive personas, allowing real-time adjustments to each customer’s journey. Every interaction and new set of options can be customized to the actual individual’s behaviors, context and timing, heightening the quality of their brand experience.
In the next three years, 60% of digital commerce analytics investments will be spent on customer journey analytics. Source: Gartner
Predictive customer service
Being able to predict when and why a customer is contacting you is insanely valuable to any organization that relies on delivering effective and cost-efficient customer service. AI supplies that edge, by allowing them to accurately deploy CS resources while also delivering exceptional personalization for the customer.
- Take banking, for instance; USAA has been utilizing AI to improve its “guess rate” about when its customers will next contact the bank, and for what products. By analyzing thousands of factors, USAA is able to match broader patterns of customer behavior to individual users. The result? Guess rate accuracy rose from 50% to 88%.
- Expedia plans to use AI for customer service, not just to help travelers search up flight and hotel options. One startup, Lola, even plans to combine AI with messaging and mobile to create a hybrid platform where human travel agents can provide users with AI-enabled personalization.
- Chatbots with AI smarts may take over customer service, which is why Microsoft is investing heavily in “conversation as a service” and Facebook has said chatbots are how they’ll monetize Messenger: “We’re going to build AI to help automate responses (for businesses).”
By the end of 2018, “customer digital assistants” will recognize customers by face and voice across channels and partners. Source: Gartner
Ten years ago, typing in a search term like “men’s leather jackets” at an eCommerce site wouldn’t have yielded the best results, because the term wasn’t exact enough for the search tool’s crude capabilities. Today, search improvement has grown by leaps and bounds, largely thanks to AI, and the pace of change won’t slacken in 2017:
- Google’s use of deep neural networks has grabbed headlines of late, for good reason. Its newest trick? The use of “sentence compression algorithms” giving its search engine the ability to effectively understand and respond to human speech.
- Advances in voice search from companies like Google and Baidu will become more and more important for burgeoning markets like Korea or China, where it’s difficult to type out searches on mobile keyboards.
- Content relevance and value will be a much bigger factor in search, as AIs like Google’s RankBrain grow to understand the context of content on websites, making it necessary for marketers to serve the searcher’s intent by focusing on content quality, not just keywords.
- Visual and audio content optimization for SEO will become more important, as search engines gain greater comprehension of the meaning of images, videos and audio clips.
- For any size eCommerce marketer, even at the smallish end of the spectrum, platforms like Elasticsearch will make sophisticated search increasingly accessible for its customers.
- Indix and other DaaS providers make it easy to extract and analyze search and product data from massive external sources, improving a marketer’s own onsite search tools without having to build them from the ground up.
Isn’t it amazing how Netflix knows exactly the titles that’ll slake your thirst for, say, documentaries about Icelandic sheep shearing? Well, maybe it’s not so amazing anymore: It’s already become ho-hum to acknowledge how a Netflix, or Spotify, or Amazon can personalize recommendations to a customer’s tastes.
Recommendation engines are indispensable, though, to brands with huge inventories of digital or physical products who need to connect with user preferences as fast as possible. Netflix estimates their recommender system saves them $1 billion a year by reducing customer churn and maximizing the value of their catalog. That’s a lot of wool.
Today’s AI recommendation engines far outstrip the old-fashioned way of doing things, which involved human-generated guidelines and historical rankings based on prior user preferences. The use of AI-driven recommendations will keep growing in 2017, part of the broader onrush of personalization across digital channels.
They’ll only get smarter and more nuanced as they integrate Big Data and behavior tracking, at the same time they’re becoming more available to every size marketer. Amazon has open-sourced its AI recommendation software, DSSTNE, in order to “promote innovation” from researchers and other third parties, though it also helps them show up competitors as they try to pull off another Android and make their platform an industry standard.
75% of what people watch on Netflix is from an algorithm-generated recommendation. Source: The Netflix Tech Blog
Using automated processes and exchanges to connect advertisers with publishers, all in search of the right eyeballs for their messages, is another area where AI will keep driving seismic shifts.
Being able to bid in real time to buy ad inventory across websites, social media, mobile, and now even TV would be impossible without artificial intelligence.
The classic examples? Demand-side SEM platforms like AdWords, Facebook and Twitter, sucking down ever-larger shares of advertiser media budgets.
In 2016, programmatic digital display ad spending was expected to surpass $22 billion, nearly a 40% jumpover 2015 and comprising 67% of total U.S. digital display ad spending. For 2017, that’ll grow to over $27 billion, as AI makes programmatic advertising more targeted and relevant than ever to each audience member.
Dynamic creative optimization
It’s one thing to target people at the right time and place, but putting the right messaging in front of those eyeballs is another place where AI will take even greater hold. In the next few years, more advertisers will turn to dynamic creative optimization (DCO) to create and test copy and design options in real time to see what scores best.
A 2015 eMarketer survey found that nearly two-thirds of ad executives were already planning on using data to drive customization or optimization of ad creative.
- When Weight Watchers sought to draw more attendance to its local meetings, it employed a DCO program and saw a 56% increase in acquisitions and a 50% bump in clickthroughs. The top-performing ad banner the DCO created drove a click-to-conversion rate 190% higher than their best-ever non-DCO banner.
- Ben Kartzman, co-founder and CEO of DCO platform Spongecell, was quoted in 2016 as seeing sees DCO as being more central to the creative process in the year ahead: “Over the next 12 to 18 months, I think executive creative directors will adopt that mantra and philosophy — and build briefs with programmatic creative in mind, rather than simply adding it on.”
- But AI may soon be producing passable creative work all on its own. Last year, MC Saatchi tested an AI which generated bus stop ads for a fictional coffee brand — and evolved the creative over time in response to people’s facial responses.
- McCann Erickson in Japan recently rolled out the world’s first AI Creative Director, AI-CD β, whose first commercial was aired for an actual client, Clorets Mint Tab.
We can probably guess what Don Draper would have thought of that.
Who’s on the AI marketing team of tomorrow?
The list we’ve run through doesn’t touch on all the ways machine learning will impact marketing in years to come, because it’s impossible to predict every innovation.
One topic that brings a wrinkle to nearly every marketer’s brow, though, is how AI is going to affect the human capital involved. Will AI lop off marketing jobs? Will it create new ones? What will the marketing department of tomorrow look like?
In our next A.I.mpact post, we’ll take our best guess at the skills, structure and roles your marketing team may need in order to succeed in an AI-empowered future.
Got thoughts about the impact of AI? Drop us a line!
Originally posted on www.marianaiq.com