AI in Marketing — Taking AIM on 2019

AI was already a hot topic in marketing for 2018 — What trends can we expect to see in 2019?

Originally published in MarTech Advisor

What does the data think?

In 2018, articles on AI needn’t be written just by a human! So I made a simple AI to forecast trends for 2019. Taking Google’s estimates of web-page-volume for AI marketing topics from the past year, we used that data to train a machine learning model (i.e. an AI). The AI predicts:

  • Big growth in AI data and tools.
  • Rising trends are chat and creative AI.
  • AI targeting and automation continue a slow steady growth.
  • AI intent and voice marketing are predicted to do little.
  • As for AI and blockchain — that’s declining, and we have passed peak blockchain.

These predictions sound sensible and may be right, but they are about as reliable as a horoscope! An AI is only as good as the data it’s fed on. The data here was like gruel — thin, low quality, and with some odd lumps in it — which makes for an unreliable model. Plus by their nature, AIs can only learn patterns that are already showing through — they can’t predict breaking changes. For example, this AI can’t yet see how Alexa and Google Home are changing how people use the internet.

1. As Daft Punk predicted: Bigger, faster, stronger.

But I trust those predictions are right when they say that AI tools for marketing are set to grow. Expect to see more automated optimization of campaigns. Also more ways to express pricing (e.g. CPCV, CPE), with smart systems converting those into CPM. AI in 2019 will be, in the words of Daft Punk: Bigger, faster, stronger.

Daft Punk are famously musicians dressing like machines. Until now, some AIM has been like that: companies have talked about AI, but behind the scenes there can be more human labour than is let on. Ultimately performance will sift out the quality from the hype and hope. As the benefits of AI increase, those vendors who can really deliver will shine through.

2 The Growth of Chat Commerce

E-commerce is evolving: two AI technologies are coming together to shift this field from visual to verbal. The first is voice, with Alexa and Google Home bringing voice-driven computing to the mainstream in the US — nearly one in five adults in the US have access to a smart speaker.[1] The second is chat, where chat-bots — AIs that can hold a simple conversation — have gone from being the preserve of spammers, to a reputable tool for fast customer service. SkyScanner took the lead with their chatbot interface for flight enquiries, surpassing 1 million traveller interactions in February 2018.[2] Filip Filipov, VP Product Management at Skyscanner, thinks “Voice is a natural destination… over time, as technology improves and the flows become more natural, voice will return.”[3]

The knock on effects of conversational commerce will shake up marketing. Web browsing, like real-world shopping, lets you survey many products and is primarily visual. Conversation is a much more constrained format. What will this mean? The jingle is coming back, and with fewer options being heard — often just the first one — SEO will be vital. We can also expect a new type of advert to emerge: the helpful AI assistant ad.

3 Intent driven advertising

Intent advertising is where you recognize what the shopper is trying to achieve. It’s already a powerful technique. But currently it focuses on search keywords, and relies on marketers to manually link their products with search phrases that show an intent. Expect that to change, with machine learning identifying intent automatically, including from other sources such as social media and email, and automatically matching intent to offerings.

Access to data is key here. Google, Facebook and Amazon hold monopoly positions, which will limit innovation to their pace and goals. Nevertheless, 2019 should see some breakthrough intent-based services in niches e.g. travel. For example, recognizing the journey of: Alice is day-dreaming about a holiday, Alice is actively comparing holidays, Alice needs insurance, Alice is back home. At the top end, intent driven ads will blur the lines between adverts and advisers. Users will actively choose to engage with systems that do a good job of understanding their needs. This will in turn lead to the demise of the annoying out-of-date advert — those ads (like travel ads when you’re already abroad) that follow you round the web, chasing last week’s searches.

4 AI creatives

Surely a machine can’t be creative? Tell that to Deep Dream, the introspective image analysis AI from Google which produces hallucinatory and undeniably original images. A closely related technique called deep style can intelligently combine artworks. For example, here is the excellent Dino-Flowers by AI + Chris Rodley, and the surreal drapery/Napoleon by Reddit user Vic8760.

There are also AI musicians, for example JukeDeck can automatically create music to go with a video.

Impressive as these systems are, they don’t really know what they’re doing. They are mashing styles together, rather than doing purposeful creative work. So AI is not going to replace the creative team for a while. We talk about “deep-learning”, but today’s AI is deep in the way that a pond is to a puddle — whilst the depths of the human mind remain an unknown ocean. However AI will become a valued assistant to the top creative teams. Machines are good at exploring many options: such as trying out different wording, colors, and stock image choices. These choices can be linked to performance metrics, providing feedback to automatically pick the best options — and to train the AI. Companies like Phrasee are leading the way here. Phrasee will auto-generate headlines and tagline copy for articles and mailing-list posts, and estimate which will perform best. In this way, AI assistants can let us better explore and validate ideas.

5 AI Entrepreneurs

2019 may see the first artificial entrepreneurs — AI systems that spot a market opportunity, and create an e-commerce business to satisfy it. This will be the subtle start of a seismic change: machines employing humans, rather than vice-versa. But for now, it will be limited to quick-fire e-commerce following a dropship model, where a niche online store acts as a bridge between consumers and manufacturers.

More mainstream, we’ll see increasing use of AI assistants to draft and optimise media spend plans. People will still make the strategic decisions, with these software assistants will increasingly automating and fine-tuning the setup.

For example, the Edinburgh-based TV analytics company TV Squared, have an AI media planner named “Predict”. It analyses the historical performance of your adverts, and then suggests a TV spending plan, tailored to give maximum bang for your buck. Here at Good-Loop we’re working on a similar tool for buying digital media space. Meanwhile, the big exchanges like AppNexus, are making AI optimized buying part of the fabric of the market, with tools like CPV pricing linking to smart CPM bidding.

Conclusion

AI is becoming increasingly important. It has moved from niche, to buzzword, to mainstream, and marketers cannot ignore it. For 2019, every major agency should have an AI strategy.

Don’t worry though that 2019 will be the year when the robot overlords take control. The growth of AI is in support systems, automation, and optimization. Strategy and creativity remain in human hands — at least for now!

References

[1] TechCrunch https://techcrunch.com/2018/03/07/47-3-million-u-s-adults-have-access-to-a-smart-speaker-report-says/

[2] SkyScanner chatbot https://econsultancy.com/skyscanner-chatbots-pass-one-million-unique-traveller-interactions/

[3] Google Home vs Amazon Alexa https://www.forbes.com/sites/johnkoetsier/2018/05/25/massive-reversal-google-home-sales-explode-483-to-beat-amazon-for-smart-speaker-crown/

Picture credits

Daft Punk: (cc) James Whatley. Photo can be used with attribution. https://en.wikipedia.org/wiki/Daft_Punk#/media/File:Daftpunklapremiere2010.jpg

Deep Dinosaur: Chris Rodley, contact for reproduction rights. https://chrisrodley.com/2017/06/19/dinosaur-flowers/

Deep Dream example — (cc) Kevin Dooley https://www.flickr.com/photos/pagedooley/19529912369. Can be used with attribution.

drapery/Napoleon: Vic8760, contact for reproduction rights. High quality version: https://www.deviantart.com/vic8760/art/Napoleon-Bonaparte-A2-664952584