Predictions for eCommerce Marketing
When I started this journey nearly 5 years ago as Windsor Circle’s head of marketing, we were forging whole new categories of marketing tools for online retail: retention marketing, lifecycle marketing, data-driven marketing, predictive marketing. Five years later, after dozens of speaking engagements, scores of webinars, countless emails and content marketing, 500 pairs of green pants, a plethora of partnerships and, yes, even a board game to help spread the gospel, I wanted to write one last blog post with my predictions for 5 big trends that will shape ecommerce in the years to come.
From revenue to lifetime value
Which business would you rather have: one that doubles revenue in a year, but retains only 5% of those customers the next year, or one that grows only 25% in a year, but retains 40% of those customer the next year? The first makes an exciting story, but has the seeds of its own demise baked into its acquisition heavy marketing strategy. It has turned on the firehose but uses a sieve to capture the value.
The second business has a more moderate source of new customers, but is capturing more upstream value. This later business will find itself more profitable, with more ability to grow on its own terms, investing in leading-edge technology and creating a brand that holds intangible value. Its customers will become its advocates, telling their friends and posting on social media, and it will likely beat out the company that has to continue re-investing profits in an acquisition treadmill.
Time and time again, I have seen the ecommerce company that focuses on retention and lifetime value eventually move towards incredible growth rates, fueled by customer evangelists (free advertising). But it is rare to get there through an acquisition heavy strategy. A maniacal focus on lifetime value is the key to this kind of success.
From engagement to score
If your marketer believes their most important numbers to report are open rates, click rates, and conversion rates, you have hired someone stuck in the past. The future is not in simple engagement or transaction metrics, but in holistic measurements of a visitor, subscriber, or customer’s value to your business. Take the following 3 examples:
- Unknown Visitor, initially clicked on a high-bid Ad 3 months ago, has visited the site an average of 1.5 times per week (5x the business’ average), time on site increased from 36 secs/session in month 1 to 3 min in month 3, AOV of products visited is 75% higher than current AOV of all customers. Has not made any purchases, and you only know their unique IP address, but have no identifying personal information.
- Known Subscriber, has been on your email list for 2 years, individual open rate of 27%, 2x your list average, click rate of 10% (of opens), 2x your average. No purchases yet, but you know their birthday. Has abandoned 5 carts over the 2 years, with AOV that of your global average.
- Customer, with 3 purchases over 2 year period, latency 25% higher than average, AOV 75% of average. Infrequent email opener, lowest decile in terms of clicks, each purchase was preceded by 2 abandoned carts, all used a coupon. Has left a 2 of 5 star review on a recent product purchased. When adjusted for $/yr, CLTV is 30% lower than average.
Who is the more valuable consumer? The Customer is the only one with revenue associated, and may be the only profitable exchange to date, being a repeat customer. The subscriber looks great based on their engagement metrics. By traditional measurements, a marketer would likely rank their Customer 1st, Subscriber 2nd, and Visitor 3rd. But he or she is being reactive, and likely spending energy and dollars marketing to people who are low margin and fickle.
If one weights all the data-points, and uses regression and cohort analysis to understand patterns, you would likely find that the Visitor in this case is predicted to have a higher future value to the business than the Customer. This would be reflected by a holistic consumer score (ie, a predicted lifetime value), and help a marketer focus energy and dollars appropriately. The marketer would invest in an onsite tool to prompt this valuable Unknown Visitor for an email address, potentially making an offer on the first purchase that leads to a high value customer down the road.
From data-driven to artificial intelligence
The digital age “began” in 2002, with the decline of analog storage and the rise of digital storage, enabling a rapid rise of data storage and processing power that roughly doubled every 40 months. By 2010, marketers were broadly talking about big data in aspirational terms, and in 2012, Gartner defining data as “high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process optimization."
Windsor Circle came on the scene in 2011 to address the massive data challenge that retailers faced, initially launching an integration hub, followed soon after by retention marketing campaign enablement. Our hypothesis was that marketers didn’t just need access to data, but a framework and a toolset to automate the use of that data to drive a wide range of value generating campaigns.
In 2016, we’re part of the move towards even more sophisticated use of data-driven marketing: the combination of even more sources and types of data, the implementation of deeper analysis of that data, and the harnessing of the phenomenal processing power of the modern machine to learn, discover, and solve for optimal return. It is this later capability that leaders in retail & tech are calling artificial intelligence.
We are not talking about human-life AI from Isaac Assimov’s iRobot, or the movie “Her”. We are talking about computer programs that self-adapt to optimize algorithms to maximize pre-determined goals, like time-to-next purchase, AOV, or lifetime value. As the retail industry adapts to the new KPIs and metrics that matter, we will see move software vendors and marketing ecosystems aim to deliver “artificial intelligence” that optimizes every consumer’s experience with a brand.
From promotion to prediction
We're witnessing the beginning of a sea-change in retail marketing, from a promotional to a predictive paradigm. With the aid of data-science software, marketers can predict what a given customer is going to do based on their behavior and profile, and impact a positive outcome for your business via your automated marketing campaigns.
For example, a retailer should use automated data-science to determine an individualized, predicted win-back date for each and every customer, rather than picking a number out of the air, like 180 days. If you send every customer who has not purchased in say, 180 days, a win-back email, you have effectively said “we miss you” to mostly customers are not yet intending to leave, or those who you lost weeks or months ago and now have little chance of winning back. You hit the mark for perhaps 10-20% of your audience. It is an absolute no-brainer to leverage predictive data to send these, and many other kinds of campaigns, at a date and time relevant to each individual.
See, every consumer’s behavior can be mapped against every other consumer’s behavior to identify high correlation between certain behaviors and desired outcomes. For example, data science may find that a consumer who looks at a given product page more than 5 times in a one month period for at least 10 minutes total, will purchase that item 95% of the time if the marketer deploys actions X, Y, and Z. You would be insane (or more likely, lacking the tools) not to identify those indicators, and automatically deploy the appropriate combination of actions, to convert these opportunities.
From sales to subscriptions
How many subscriptions does the average consumer now pay for each month? Netflix, Spotify, Amazon Prime, Tinder premium, internet access, gyms, wine and produce delivery... Add to those an increasing number of goods and services that consumers are choosing to rent rather than own, like Uber instead of a car, and you see the tide turning from shopping for products to subscribing to solutions.
I would not argue that every retail product should be turned into a subscription model, though there are certainly many that should and will. But I do believe that nearly every retailer needs to answer the question: “how do I acquire and retain customers who generate a dependable, recurring revenue stream, much like a subscriber would?”
While the retention and lifecycle marketing paradigms arm retailers with strategies to accomplish this, the question needs to be asked in relation to the fundamentals of the business. Does the product catalogue have natural “gateway products”, lifecycle sequences, replenishment or upgrade opportunities? Does the brand story have an arc with key “reveals” as consumers navigate through each stage of the lifecycle? Are there rewards at various levels of engagement and customer score? Is a high lifetime value a win-win proposition for both business and customer?
A Marketer’s Bonanza
Today’s marketer finds herself or himself in a truly marvelous world. We have a plethora of tools and services at our disposal to engage our audiences in creative, personal, and valuable ways. The introduction of accessible data-science, multi-channel marketing automation, and segmentation based on multiple types of data and behavior, all make for a very exciting palate to paint with. Combine these tools with a strategic framework rooted in maximizing customer lifetime value and you have a great shot of building a marketing engine that generates outsized return for your business.