The Path to CX Individualization in the Age of AI
by Sarah O'Neill, on 12 July, 2021
Beep Boop Beep Boop, the robots have taken over. The martech industry is just like that one scene from I, Robot, where Will Smith is just walking around that warehouse of androids. But instead of Will Smith, picture our very talented speaker Joey Moore.
There we go.
Optimizely is all about data-driven decisions. But, as Joey points out in his latest webinar for #MarTechFest Dial Up (Mini), 1/3rd of the companies employees are all there to work directly with customers. So, how do you use the data-driven techniques, in a way that benefits such a large chunk of the business?
91% of consumers report that they are more likely to shop with brands who recognize. remember, and provide offers and recommendations that are relevant to them.
80% of consumers are more likely to make a purchase from a brand that provides personalised experiences.
So, by building advocates out of your customers, they are more likely to make purchases. And a lot of this comes from detailed, personalised, and streamlined experiences. This is less easy than it seems. That's where the robots come in.
Only 32% of marketers agree that they are getting personalization right, and a mere 18% are "very" or "extremely" confident that they have a successful strategy for personalisation. So, only 1/5th of companies are confident they can deliver these vital experiences, despite most knowing that customers demand modern experiences.
But first, Joey walks us through how these experiences have evolved.
From the mailing list, to 1-2-1 experiences.
Today, customers are connected, aware, and demand personalised experiences. If you don't have what your competitors have - customer-centric focus - then they'll see the difference.
No longer is the one-size-fits-all mailing list sufficient. Not even the 2010's segmented digital campaigns are enough. Today, in order to provide the best experience possible, brands are turning towards 1-2-1 machine learning in order to gather the best data, and make the best use out of that data.
The expectation of the customer has grown exponentially. Especially if they've interacted with a brand before. Consumers know that, if they've spoken to, or interacted with, the company before, that brand has their data. So, in terms of providing a consistent, cross-channel experience, the brand needs to avoid making the client feel like they're starting from square one every time.
As a result, everyone is trying to achieve these 1:1 journeys, with content that differs by the individual, recommendations, and a system managed by AI. But this, of course, is a challenge for companies.
The Challenges of Personalisation
With an email list, companies know what's expected and required from a marketer. Whereas this is new ground.
Segmentation came with it's own challenges. It was difficult for the teams who had to grow the segments out, with the 'segmentation ceiling', as Joey puts it, becoming a hurdle. How many segments are right for your business? How do you scale? How many people do you need to hire?
In the end, no matter how many people were added, some brands are never able to reach that 1-2-1 individual experience for visitors. This is a case of having the ambitions, but not the resources.
So, it's important to take out the theorising, hypothesising, approximations and guesstimates.
Taking out the guess work.
So, what content do you need to be recommending to your customers? Using what data?
Instead of taking a shot in the dark, it's important to experiment, and optimise. Conduct ongoing tests, to be sure the changes in the customer journey are effective, and drive results. And always remain cognisant whether or not you can support these changes.
Optimisation technologies are part of the digital intelligence architecture. From customer context and behaviour, you'll move through digital touchpoints, and reach the optimisation stage. This involves three points:
- Behavioural targeting
- Online testing
And to succeed, you'll need to use a blend of all three. And as a result, you'll be putting the right content or product in front of a customer, based on proven and solid data points.
So, how do you utilise the data, in a way that is clean, and accessible?
A Blueprint for your CX individualisation.
Basically, you want a robust customer data platform, or CDP, to rely on. All the injest data, profile data, ingest customer or visitor data from other solutions within your tech stack are all siloed in the other parts of your business. A data platform harmonises this.
"Having that data platform so you can begin to collate, and harmonise, all of that data in a single location is absolutely key," says Joey.
Data needs to be accessible, and it needs to be usable. With all the different channels, and different tools, you need to ensure you have the necessary data to improve these experiences as well.
AI and ML can also be used to make content recommendations to users as well.
Then, you have to consider how you segment, advance the data, and predict behaviour. This might be identifying churn threats, or even customers who are just about to make a purchase.
Product and content recommendations.
It's hard to know to present to an individual. Plus it's a challenge to do so at any kind of scale, without the assistance of AI or ML.
In terms of product recommendations, you might consider collaborative filtering, creating models made up of data taken from scanning hundreds of thousands of visitors, content, or products. With this model, you can create a individualising experience for every single user.
Make sure all these efforts are moving towards a predefined goal. In a retail scenario, this might be revenue, or the amount of users moving through the entire process. For a content or lead gen scenario, it might be putting the right content in front of the right people, in order to encourage them to download a white paper.
Solve for 'When'.
Some event-based triggers might include:
- Segment-based triggers
- Event-based triggers
- Activity-based triggers
- Ecommerce triggers
- Conditional activation
To capture high intent point for users, e.g. a purchase scenario, using triggers to put a promotion in the eyeline of a client, reiteration of your USP, or even an abandoned basket email, you have to use data to inform these actions.
This might be as simple as deciding how long it'll be between the customer exiting the page, and the abandoned basket email. This is all about maximising the opportunities you get from visitors.
Ask yourself: "where is the most interaction coming from?"
'Where' is one of the most important questions to ask in this era of multi-channel, cross-platform interaction and communication.
A customer journey might involve interactions with your brands that span across multiple different platforms; desktop, mobile, SMS, and beyond. You should use the data to see where you need to be pushing and managing specific messages. Experimentation can be used to see what combination is the most effective in delivering this message.
Then, all this data needs to be consumed back into the customer data platform.
You can then see whether customers are reading the emails, whether they're opening the SMS, whether the desktop site is generating the outcomes you've invested in.
So...how do you do it?
In terms of how to conduct this kind of optimisation, it's about having the range of capabilities to be able to test, be it a simple front end test, AV test or multi-variate test. With multi-variate, you maximise exposure through having multiple versions of your messaging. In this way, you get to results quickly, and see what's effective, plus you're testing things you wouldn't've tried otherwise, in a safe, data-driven environment. Nice.
You might also run a stats engine in the background, to really make sure the results you're getting are statistically significant. It's important that the same, or comparable, stats are coming back each time. With good, clean data comes the ability to support internal business cases, plus the confidence to invest further in the areas of the business that heighten CX. It's all about having confidence in your data.
These analytics mean you can look at an individual's journey throughout the whole experience, what are they looking at, that are they clicking on, and what was the most exciting piece of content they clicked on before downloading the white paper. With this, you can understand what the right content is for your users, and what content is driving them towards your goals.
Recognise that you need to be able to activate your data, and not just hoard it. this might involve creating an 'interest profile', to target visitors with the right, compelling content.
Want to hear more from the very talented, very charismatic Joey Moore, Senior Product Director for Optimizely?
Well, come watch the full websesh on-demand, right 👉here👈. That's right, there👈. Click it. Click it. ☝️