Demystifying AI for Personalisation

Demystifying AI for Personalisation

I originally wrote this article for MediaPost’s Marketing Insider section

What comes to mind when you think of Artificial Intelligence (AI) and Machine Learning (ML)? If you’re picturing HAL 9000 from sci-fi classic 2001: A Space Odyssey, you might be getting a bit ahead of yourself. AI and ML are hot topics these days, with some of the discourse centering around the potential negative consequences of unregulated AI advancements – computers gaining sentience and taking over the world While this is a scary thought, the reality of how this technology is being used in the real world is a little more prosaic.

The Reality of AI: Personalisation

While advanced use cases like image recognition and autonomous driving are often praised, the most common use cases for AI and machine learning from a commerce point of view revolve around personalisation. In this context, like many other types of personalisation, it basically comes down to taking signals from a particular data set or some past behaviour, and using that to inform a future action. For example, if a system finds out over time that the users that interact with content about car maintenance tend to end up buying more car insurance, then it can prioritize showing more car insurance product suggestions to users that interact with car maintenance content, thus streamlining the journey.

This action can also be done manually of course, but with the help of AI / ML it can be done without a human having to trawl through a sea of data to find the insights and action them. AI / ML can uncover patterns that humans may not see, and can be set-up to automatically action them without explicit go ahead. Look at it like a helping hand in implementing personalisation, letting you free up your marketing team to concentrate on other creative tasks like creating campaigns or copy.

Identifying Consumer Segments

Another particularly helpful role AI / ML can play is to detect customer segments and help create personas. This technology can be very effective at finding ways to group customers together that might not be overly apparent to the human eye. Creating personas can be a difficult thing for brands to get right. Much of the time, true insights can be hidden behind the data and marketers can end-up relying on basic demographic-based characteristics like age, gender, or geography when creating their personas. AI / ML can help discover nuanced segments that human analysts might have missed.

AI won’t work without data

When done well, personalisation can have an outsized benefit for retailers, increasing conversion, cross-selling and brand affinity, and AI / ML is making it easier for companies to get it right. All this comes with a caveat though. To really utilize AI / ML you need data. The more, the better. It’s never been more important for brands to take their first-party data strategy seriously and it’s not just to take advantage of AI / ML. Bolstering your first-party data can have benefits across your whole business – strengthening your ability to sell direct to your customers, increasing margins and lifetime customer value. In a post-third-party cookie world, brands that rely on middlemen to reach their customers will be increasingly at a disadvantage while brands that take their first-party data seriously and use technology like AI / ML to utilize it will be well placed for tomorrow.

Posted by Rob in Tech

My talk at the 7th Sitecore Strategy MVP webinar

I was delighted to present a Horizontal case study from a mobile app personalisation project that we worked on for one of our financial clients at the 7th Sitecore Strategy MVP Advisory Council (SMAC) webinar on 7th September.

Watch it below from 1:35 – 15:50 mins.

Posted by Rob in Speaking, Tech

How Sitecore’s Automation Tools Can Make Personalisation Easier

15th August 2021

Personalisation is hard. But as American educator, author and politician, Hamilton Holt once said, “Nothing worthwhile comes easily”. Most companies today are aware of the value behind understanding their customers better and showing them relevant content. According to Forrester, 89 percent of digital businesses are investing in personalization. But a study they conducted found that only one in five organizations are effective at personalizing content at-scale, despite personalisation being the top success factor for customer and prospect engagement.

So it’s clear that companies need a hand with personalisation. Good thing Sitecore is there to help. At a recent webinar on ‘Leveraging Automation and AI for 1:1 Engagement’ representatives from Horizontal Digital and Sitecore laid out how Personalisation can be made easier by Sitecore’s new Automation and Machine Learning solutions.

What Problem Are You Trying To Solve?

While there is seemingly constant innovation in the field of Artificial Intelligence (AI), Machine Learning (ML) and Automation, it’s important not to get carried away with the hype, and really delve into how exactly these technologies might help your business rather than simply engage in a box-ticking exercise. Whatever technology you are looking to utilize to improve your business, ultimately it’s a means to an end. Artificial Intelligence, Machine Learning and Automation is no different. Horizontal’s George Smith, Regional Managing Director MEIA elaborates, “AI is not a solution in itself, it is a way of solving particular problems. You’ve got to acknowledge what the business challenge that you’re trying to solve is. There is always some deeper problem that you are attempting to solve with AI”.

This is where Sitecore comes in, with a range of new tools that utilize AI, ML and Automation that can make Personalisation a whole lot more accessible and effective. Smith expands, “Sitecore AI’s objective is continuing to fulfill the promise that Sitecore has always had of real-time personalisation – that ability to orchestrate customer experiences around not just a generic user, but an individual in real-time, at that moment showing them the content and functionality and messaging that will convince them to purchase. Sitecore AI makes that process easier for any size of marketing department in any organization”.

The History of Personalisation in Marketing

If we think of marketing as having three stages – Strategy, Operations, and Learning – we can look at how the opportunities for Personalisation for each has really changed over the last few years:

  • The past – Traditionally speaking, every type of Personalisation was manual. Whether that be across Strategy, Operations or Learning, only brands with significant resources could afford to justify treating each customer as an individual.
  • Today – Many Operations of the typical marketing department can be automated today – a lot of the content management, online sales and marketing operations can be controlled and delivered in a Personalised way by Sitecore for example. But Strategy and Learning are still manual processes. From a Learning point of view, whether that’s looking at the analytics, looking at segmentation, experimenting and testing, basically anything that relates to constantly understanding and refining your knowledge falls into this category. Because over time, trends change and consumer segments change etc. no organization can afford to stand still and must continuously be Learning.
  • Tomorrow – We’re at the stage now where AI has begun to take over the Learning aspect of Marketing Personalisation. Things like auto-segmentation, auto-testing, and auto-personalisation let brands focus on Marketing Strategy, with campaign concepts and creative work that people do much better than machines. Sitecore can finally handle the Operations and Learning process and let the Marketing Department focus on Strategy

Smith continues, “This is not about getting a robot to run your organisation. It’s about the division of labour – let machines do what machines do best, and let people do what people do best. You’re always going to need that human touch. Lets stop getting creative and innovative marketers to do CMS administration and figure out how complicated workflows in the back-end of a system work. Lets allow the system to do that. Lets allow machines to recognise patterns that humans can’t even see in terms of segmenting customers, and lets bring all that insight together so it can be used by humans in creative pursuits, in campaign creative and marketing strategy”.

Sitecore’s New Automation Tools

It’s an exciting time for brands that realise the importance of the Customer Experience on digital channels. Platforms like Amazon and Uber have raised the bar when it comes to providing customers with fast and tailored experiences and users now expect this type of treatment from any brand they interact with in a digital setting. As Smith outlines, “Fundamentally, these tools are for organisations that understand that their future competitive advantage, the thing that is going to differentiate them from their competitors, is no longer to be found in the way that they produce their products, but in the way that they can orchestrate consumption experiences. The moment you realise that your Customer Experience is your Competitive Advantage in a digital economy is the moment you should realise that you need Sitecore. It is the best digital experience platform in the world, from content to commerce”. These new tools from Sitecore can help your organisation make this a reality:

  • Sitecore Cortex – This tool combines advanced machine learning algorithms and an innovative machine learning engine to deliver real-time insights across both native and third-party customer data. It does everything from automated segment and audience discovery, attribution analysis and automated content tagging. If you are an organization that has a team of data scientists, and you can use the raw data, Cortex is a tool that lets you take all that raw data and put it to work.
  • Auto-Personalisation – If you don’t have a team of data scientists, don’t worry, Cortex’s Auto-Personalisation functionality will essentially do that work for you. Sitecore automates analysis and segmentation, and finds the right combinations of content to show to each user based on its past learnings. Importantly, there are no minimum data requirements for this to operate so you don’t need to have huge volumes of data to make this work. This feature is included in the most recent v10 version of Sitecore. 
  • Data Exchange Framework – This feature allows you to enrich the insights that the system is using by bringing in other data sets to compliment its own data. Cortex really lets you truly utilize the data that you already have.

Biren Balakrishnan, Sitecore’s Sales Engineer Manager for Asia explains how to think about using Cortex in your organisation, “Think of the relationship between the left side of the brain, which is more analytical, and the right side of the brain, which is more creative. You need both sides to effectively engage with your customer so there is a balance needed. From an organisational point-of-view, the ‘left side’ is the Data Scientist, Analysts and Developers, and the ‘right side’ is the Marketing Department”. Sitecore Cortex can compliment both:

  • Data Science – They want to train models, run algorithms, project data and apply scoring. They need the ability to bring their own skills to the table to solve those problems. Sitecore Cortex lets them do this by providing them with all the data that they need to make these decisions using state-of-the-art tools.
  • Marketer – They’re less worried about the behind-the-scenes processes and more focused on delivering great experiences for the customers. Sitecore Cortex lets them do this by taking the analytical workload away from them and letting them focus on strategic, conceptual and creative endeavors that can improve the customer experience.

According to Balakrishnan, when it comes to the challenges of Personalisation, it typically boils down to resources, knowledge, and having the technical set-up to do this. This is where Sitecore’s AI tools can help, “What we’re looking to use AI for right now is to remove the heavy lifting and make it as easy as possible for organizations and individuals in the marketing team to get their job done, and to realize the potential of things like personalisation and excellent digital experiences. AI can be a tool to speed up your Go-To-Market strategy, lower the barriers to entry in terms of resources, structure and knowledge, and get you where you need to get faster”.

The Power of Auto-Personalisation

If you had to focus on the one thing Sitecore does better than any other system in the world, it is probably the real-time personalisation of content shown to the user. Traditionally this has been driven by a set of rules manually pre-configured into Sitecore, however this has now evolved to be possible using such Machine Learning and Artificial Intelligence. The potential of Auto-Personalisation is so profound as it essentially democratises Personalisation, making it available even to organisations that do not have a data science division or the traditional resources to make this a reality. According to Balakrishnan, “Auto-Personalisation is about making Personalisation as easy as possible. The marketing team only needs to worry about the content variations and they can let Sitecore AI do the rest without having to do the heavy lifting themselves”.

With Sitecore Auto-Personalisation, reducing the workload for the Marketing Team is as easy as as one, two, three:

  1. Marketers create all the content variations that they want to use (e.g. header banners, product feature components, image galleries, offer components).
  2. Select which components they want Auto-Personalisation enabled on.
  3. By pointing the component at a folder of options, the Auto-Personalisation engine can then choose the correct content variant from that collection to show. There’s nothing to set up in terms of segments or anything else.

These features will save your Marketing Department a lot of time and effort on your Personalisation journey. As George Smith outlines, “Understanding the customer, being able to serve that customer what they want, when they want – that is the future. And what Sitecore will give you is a tool that will do the Operations and the Learning for you so you can just focus on the Strategy behind that”.

Find out more about how you can leverage Sitecore’s new Automation tools to enable Personalisation by watching the webinar video here.

Posted by Rob in Sitecore
WTF is an NFT?

WTF is an NFT?

I originally wrote this article for the April 25th 2021 issue of Campaign Middle East magazine

Unless you’ve been living under a rock for the last few weeks you’ve probably heard people talking about NFTs, or non-fungible tokens – digital files such as images, videos, audio or text files that have recently been changing hands for millions of dollars.

If you have no idea what ‘fungible’ means, you’re not alone.

Essentially, an NFT is a one-of-a-kind digital file that uses blockchain technology to register a unique version of itself – generally considered an ‘original’ or definitive version of any file that, in theory, can be copied an infinite amount of times.

With me so far?

It’s basically a way of manufacturing scarcity using a digital stamp of authenticity. And where there’s scarcity and exclusivity, there’s usually a cohort of people lining up to hand over their cash.

Over the last few weeks a piece of digital art, Beeple’s ‘Everydays: The First 5000 Days’ (pictured above), was sold for $69 million at Christie’s, a cat GIF sold for $560,000, and, in a somewhat meta twist, a New York Times article about NFTs was itself turned into an NFT and auctioned for $560,000 with the proceeds going to charity. In many ways, the buyers of these NFTs are just buying bragging rights, but they also believe that it’s an asset they may be able to resell later.

It might sound like pure alchemy but in tech circles it’s seen as a way to address a problem that has emerged over the last 20 or so years of how to register value for once-tangible assets like art, books, music etc. that now primarily exist in a digital form. NFTs give creators a way of engineering value for a digital asset and making digital collectibles possible. Some creators are already starting to take advantage of the buzz.

Tennessee rock band Kings of Leon was the first band to release an album as an NFT this March with their latest album, ‘When You See Yourself’. The band released three types of NFT of the album, all including exclusive digital artwork, one of which offered a limited-edition vinyl and another that included live show perks like front-row seats for life. A company called YellowHeart developed the smart contracts and intelligence within the tokens and claim to want to “use blockchain technology to bring value back to music and better direct-to-fan relationships”.

Outside music, there has been an explosion in sales of digital-only products recently especially in gaming, from exclusive outfits in Fortnite to various in-game items in Minecraft, Roblox, or any number of other games. Fortnite is a free-to-play game yet generated $1.8 billion in revenue in 2019, most of which came from selling in-game items.

It’s not difficult to see how NFTs might be used to enhance the value of such digital goods – letting platforms effectively create limited editions of items that can be resold and traded, not unlike how an artist might release a limited run of prints of an original artwork.

Brands that sell tangible goods face a steeper challenge in taking advantage of these digital trends, but even some of these guys are getting in on the action too. In March, luxury brand Gucci launched a ‘virtual sneaker’ that can only be worn in digital environments. The neon-coloured digital shoes can be purchased on Gucci’s mobile app from $9 and users can try them on using augmented reality and “wear” them in photographs on social media. Similarly, Nike has patented NFT versions of shoes called CryptoKicks, which allows users to create custom sneakers that may then be manufactured in the real world. In this way NFTs can blur the line between physical and virtual goods, while capitalizing on monetization opportunities in both.

There are a bunch of other ways that brands can get creative like this when it comes to taking advantage of NFTs, from exclusive and limited edition digital content around new products or a physical experience like a concert or sporting event, to even crowdfunding new ideas and products with early customers being able to sell their initial purchase if the concept takes off. While the hype and hyperbole behind emerging concepts like NFTs can make it hard to distinguish the opportunity from the noise, behind most novel ideas is often the possibility of something useful. For NFTs we’ll just have to wait and see whether the emperor has any clothes, digital or not.

Posted by Rob in Campaign Magazine