From layout and language to pictures and product selection, myriad elements impact a visitor's experience with a brand's owned digital assets (e.g., landing pages, app screens) - with the end-user woefully unaware of the decision-making that happens to deliver such items.
In many organizations there are numerous people who have a say in how pages are presented with some opinions being backed by metrics like conversion rates of similar collateral, some opinions backed by data like through A/B testing, and others backed by the highest paid person's opinion (HiPPO) or someone's "hunch."
The naturally occurring give and take that happens when design and marketing teams work together to produce creative could soon, however, meet its fate. Lately, both design and marketing can, if they are willing to surrender some control, rely on software to be the unbiased third-party opinion in the room answering questions about what should go where, who should see what, how long it should run and why certain variations will outperform others. Yes, machines will make decisions, learn from themselves and make even better decisions next time.
Where it may have once been impossible to create a perfect experience for every visitor, personalization technologies are increasingly providing opportunities to do just that. In a recent interview with Website Magazine, Kevin Lindsay, head of product marketing for Adobe Target (the "personalization engine" that powers Adobe Marketing Cloud), suggests that the system he and his team work on is now capable of bringing out the best experience for each individual user (i.e., the perfect page).
A new "Auto-Target" feature uses machine learning powered by Adobe Sensei (Adobe's artificial intelligence; think SAP Clea, Salesforce Einstein, IBM Watson) that, with one click, makes complicated decisions about who should see what experience and who should be exposed to different variations across channels (e.g., Web, app, single-page apps, React, IoT, game consoles).
By feeding the content and "turning on" this feature (which will be an overnight release, tonight), marketers can introduce as many variations as they choose in order to personalize experiences across their digital properties (Lindsay writes) with Auto-Target automatically determining the best content for each consumer and continually optimizing those experiences over time as the consumer takes additional actions (think of it as automated and scalable A-Z page-level testing).
As for who sees what, there can be as many variations as creativity and traffic allows, says Lindsay, with Auto-Target adapting to what does and does not resonate with customers. There is a caveat, however, in that companies will need the "right" amount of traffic so the engine can have all the statistical reliability that is expected and marketers can be confident in the results.
As marketers add variations, they will either need more time or more traffic in order to accommodate those variations in customers' experiences. Adobe Target thought of this, offering a calculator within Auto-Target, so marketers can understand how much traffic is needed. Traffic, however, will be a non-issue for most Adobe Target customers as its customer base is mainly large banks, retailers, travel and consumer goods brands.
Those worried about taking a more hands-off approach should know that Auto-Target was developed with a backup policy. None of the variations served through Auto-Target will ever perform worse than the best control (the content, images, layout) that everyone agreed would be a baseline experience to offer visitors (in a similar process as mentioned earlier in this article).
Designers and marketers will still play a critical role in Auto-Target, as personalization, says Lindsay, is hungry for content and data. More variations will need more creative, which will allow teams to create all the different ideas they have - a spot on a homepage, for instance, is not restricted to one idea but can essentially be changed tens or hundreds of times. As Auto-Target learns what works best, creative teams will too.
(Countless combinations of content could be automatically served to customers based on the context of their visit.)
To what effect will Auto-Target have on an enterprise's bottom line? Simply, consumers will have a "great" experience because it's contextual to who they are, where they have been and what they want - staying on the site (versus bouncing) with marketing assets nudging them in the right direction to complete the marketer's goals.
Marketing teams will have to be enterprising with how they can use these takeaways in other marketing channels to get more clicks in emails, more reach on social or more views on ads. In the future, it is likely that Adobe Sensei will play a role in this regard - acting as an intelligent marketing assistant (we're waiting on a mascot of some sort for the product) that makes recommendations across all Adobe products based on what it has learned (pure prediction) or deploys these findings automatically across the Adobe line.
For now, let's back up and think about how a banking site could use Auto-Target.
If a person goes to their bank's Web page and the key offer spot is showing content about new homeowner incentives - but they already own a home and, more frustrating, have their mortgage through the bank - they may continue on to complete their goal (like signing in), but the financial institution has lost an upsell opportunity. Instead, that person should be presented with an offer like a 0% credit card with travel rewards if they are known travelers, do not yet have a credit card with their bank or a number of other variables that will be taken into account automatically. These types of automated offers have been available through Adobe Target for a while. With Auto-Target, however, there can be hundreds of different offers that take into account all the variables and traits about that visitor and optimizes those offers based on further actions they take over time.
Now, think of a company that offers a monthly subscription box and is reliant on recurring revenue more than new customer acquisition. How much business is it losing when a current subscriber goes to the site and only sees promotions for new customers instead of trying to retain their business? Over time, a lot. And, that is just the basics of personalization. Now, serve them content that is the direct result of their last visit (did they visit the blog, if so how long did they stay there or did they visit an FAQ on canceling their account and click on other content instead), their location, their device and more contextual moments they have offered.
Companies are failing to have a two-way conversation with their visitors. Instead, they should "listen" to what a person is telling them and respond accordingly.
This fall, it is expected that Adobe will release a beta version of an algorithm that takes a consumer's language, search terms and more and analyzes it at the semantic level. Based on techniques in natural language processing, a site can treat interactions with the brand as dialogue with every interaction becoming an indicator of intent. Lindsay says by the time a customer gets to a certain point in a Web experience, everything they see should be the sum of those experiences not just that moment in time. It is this algorithm that will make recommendations based on customer-company dialogue.
Does the perfect page exist? When it automatically determines the best digital marketing experience for each consumer and continuously optimizes those experiences over time as Adobe describes its product, it just might and designers and developers need to be ready for their role in the 1:1 Web.