Retail Nation. Your world is about to change. You are going to move closer to your customers than you ever thought possible. You will be able to connect to them as individuals harmonizing customer experience to the value you require from each of them. The Era of Insights increases the quality of customers you acquire, preserves / expands their product line spend and prolongs their lifecycle – across your digital, stores, and marketplace efforts. How? You are about to move beyond data to insights at scale.
Insights Defined
What are Insights? There are dozens of different definitions – most of the time created by people who aren’t on the hook for delivering them. I personally love the definition below.
noun in·sight \ ?in-?s?t \
An understanding of relationships that help solve a specific problem or pursue a specific opportunity.
The most important part of the definition is the phrase understanding of relationships. It implies that we are moving beyond the data, applying knowledge about customers, products, experiences – in order to arrive at business and customer outcomes.
Insights aren’t new.
“Insights” based marketing is being done today. Simple examples are the churn programs, the shopping cart abandonment email you trigger, and your Look-Alike seed audiences. In all of these examples, you are using your understanding of relationships to identify shopper health, recover abandoned shoppers, and try to find your next generation of customers.
Until now, you’ve been the connective tissue between data and outcome – the insights engine has been human.
We know when we deploy insights-driven marketing programs, we make more money. The problem isn’t the performance of these programs, the problem is the scale.
Insights aren’t new – but scale is.
Insights scale exists in 2 dimensions – the first is the “opportunity” and the second is the “ESV” – or the Enterprise Shopper View.
The Opportunity is simple. It’s the “what” of the insight – it’s the business goal or outcome. While there are several dozen “opportunities” that we could discuss but to keep this article from becoming a book – we will focus on Churn Risk – simply defined as a likelihood of disengagement.
The Enterprise Shopper View – defines the “where” of the Opportunity. In the example of churn, you will want to know if the churn event is in-store, on the website, or at your retail brand level. You may also want to know if that churn event is in a traditional consumable product, a season product line, or your spring product catalog release.
This is where scale comes into play. Moving beyond the “churn example” – let’s suppose you have 500 stores (digital/physical), selling 150 truly unique product lines, recognizing 3 discrete seasons a year across 10 actionable customer personas.
In this example alone there are 13.5M opportunities to create unique, incredibly high-value marketing programs
…more than any of us have the time, dollars, or staff to assess – let alone act on. If only there was a technology specifically engineere…oh, there is.
Machine Learning & Insights
To be clear – machine learning and insights are NOT one and the same. Vendors tend to position ML as an answer when it is nothing more than an ingredient – albeit a critical one.
The Insights Engine of the Future
uses machine learning to help
identify and manage opportunities @ scale.
Going back to our Churn example. Machine Learning will give us a churn risk score, but you don’t have a reliable, well-informed marketing program until you add predicted customer value (i.e. now you know what is at risk).
That said, Machine Learning (ML) is that path to scale – providing an expedited path to the “What” and “Where” for Insights. As importantly. Insights are where ML finds purpose, hard ROI, replicability, and accountability – something that we all want to see substantially more of.
Your goal – The “Always On” Insight
Hundreds of Insight driven programs
You know how “web analytics” data is just sitting there…waiting for you with the freshest data of what your shoppers are doing on your website?
Machine learning and Insights should be exactly the same. You want the ability to, at any point, access Insights to solve any business problem that may arise or (even better) have insight proactively reach out to you when an opportunity is identified. Yes, that is very much our near term future. But not if you don’t do the basics.
Think Beyond 360 – Your Customer Data Core
A lot of vendors are selling “Data Consolidation” as the “Single Shopper View” or a “360 view” – and maybe for them that is the answer. But it’s not for you.
The “truth” of the customer isn’t found in data, but in the context that surrounds it.
Context (or data contextualization) is a process in which data that is collected, processed, and becomes more descriptive. Below are the context engines that we recommend;
- Identity – it’s a must and represents that “basics” of managing context.
- Order Disposition – managing active transaction states including returns, exchanges, cancels – with reason codes. No, it’s not exciting – but it is required.
- Enterprise Shopper View – clarifies the customer relationships across 1st party digital/physical sales channels and 3rd party market places.
- Global Product Taxonomy – creates clarity in product associations when comparing historic customer purchases, to current product catalog, to future product lines.
- Pragmatic Personas – enables appended data to be applied and managed to shoppers, products, and content (context tip: the greatest value in appended data isn’t who an individual is, but in how/when they changed…persist your appends).
- Operational Analytics – picture points 1-5 described analytically. By engineering the evaluation of the data – you enable enhanced understanding of seasons, product thematics, and allow your ML to accommodate sudden changes in your audience, product lines, and/or sales channel/engagement.
For us mere humans, contextualization creates massive analytical and marketing value. That said, for automated Insights generation using machine learning – it’s a gold mine. You see, in the data science world, contextualization supports a process called “feature engineering” (features are what fuel modeling/deep learning).
Closing Plug for OpenINSIGHTS
Retail Nation – the Era of Insights isn’t science fiction. It’s here now. It’s driving 300% increases in ROAS. It’s informing store ops and merchandising teams. And it’s available to you through OpenINSIGHTS – deployed into your own cloud (not ours).
by Angel Morales
