What role can machine learning + Neuroscience play in ecomm?
The quick answer:
Leverage neuroscience and machine learning to create smarter funnels.
So, let’s run through the basic ecomm user flow quickly and point out a few of the problems.
At the top of the funnel, the goal is to attract users to your site. We select an audience, and, then, guess at what will get people to click. To optimize, we A/B test audiences, campaigns and copy and see which have the best numbers.
While guessing, we rely on our intuition (which is often educated by our own–or known– unconscious preferences and behaviors). We put ourselves in the shoes of our audiences.
Landing and home pages:
Visitors land here through digital campaigns, SEO, directly, referrals, organic social…to begin the conversion conversation.
And this conversation is nearly exactly like the acquisition conversation. We use our intuition to answer: “What will get people to click through?”
The problem is the consistency of personalization.
Even with great branding, and especially if coming through a campaign or organic social, the conversation is more general when we land on the site.
We all know what this feels like. At a restaurant, the host greats you (maybe even charms you) then leaves you with a waiter to take your order.
Category and product pages:
To continue the restaurant example, the category pages are the menu options and the product pages represent what you can order. These, in the best of cases, are varied with personalized product recommendations and promotions.
In the worst of cases, they’re static and uninformative.
The problem, as I’m sure you’ve already guessed, is the role these pages play in the conversion conversation. The product pages are those on which visitors are supposed to convert and yet they are the least personalized.
I’ve always understood the cart as something of a scale that’s tipped. On one side you have the actual conversion–the purchase confirmation–and on the other you have the internal debates that precede that click or tap:
“Do I actually want this?”
“Can I get it somewhere else for less?”
“Is this really worth the money?”
“Will it fit and/or be useful?”
(This list is not exhaustive…)
The problem, when it comes to the cart, is the complexity of the visitor’s preferences and mindset. We don’t know why certain items were taken out of the cart or added back in, why certain carts are abandoned and others instantly processed.
And, this is the main problem of ecomm sites, most sites don’t do much, if anything, to gain insight into the users’ preferences, reactions and/or decision making processes to optimize the conversion conversation from start to finish.
Listen to your customers. Analyze their responses. Understand what they’re telling you.
To begin, we need to ask a few questions:
1. What stimuli (campaign or on-site elements) are influencing behaviors?
2. Which metrics are faithful predictors of behavioral patterns?
3. Which behavioral patterns lead to conversions?
Serendipitously, the answers to these questions perfectly overlay the conversion funnel.
As we said before, we judge our intuitive response to the proposed digital campaigns against past click-through and traffic data. The problem is codifying the intuition used, verifying its effectiveness, and carrying those insights through to conversions.
The solution to this is a systematic dissection of the stimuli presented coupled with a pattern-based, retrospective analysis of campaign and on-site behavioral data, in other words Machine learning + neuroscience.
You’ll know exactly which campaign elements on which platforms at what time of the day on which days of the week in which month best drive conversions.
Landing and home pages not only allow you to pull the insights from your campaign data through to the on-site experience but also, and more importantly, allow you to begin robust data collection and real-time personalization.
What’s this mean?
1. You can empirically design landing pages with those elements that best perform with your digital campaigns.
2. User data gathered on landing pages allows you to further determine user patterns and personalize all other pages visited.
It means that you’re truly engaging in an ecommerce conversation.
Category and Product pages:
Each behavior a visitor makes is recorded and responded to. The elements of category and product pages can then adjust in terms of content, layout, and style. Each adjustment is a data-based decision based upon matching previous and similar behavioral patterns with those exhibited or, if it’s a returning visitor, with the elements that led to conversion previously.
Thus, you’re both keeping and increasing the momentum generated in acquisition.
And assuming this all goes well, we get to the cart.
The cart is the measurement of the momentum gathered throughout the entire conversion conversation.
And, if the conversation has gone well, you’ll have done enough to convince them to buy. Maybe the machine learning model understood that they were doubting and responded with promos (free shipping or discounts). Another visitor may have been on the fence as to the value of the product and needed to see the reviews above the fold.
Like most conversations, it’s as much about what you say as it is about how and when you say it.
And you always get to determine what you say.
That’s the wonder of machine learning and aggregated data. It’s pattern recognition that improves its accuracy with each new data point. In other words, it’s a way of systematically learning the ecomm language of each of your customers.
If you’re interested in the efficacy of your site from acquisition to conversion, sign up for an audit.
CLICK HERE TO SIGN UP.