Traditional retail models are crumbling, and those that fail or are slow to adapt will fall too. The intelligence behind any Retail Digital Transformation takes on a very specific set of dimensions that set it apart from that of other industries: a Efficient Customer Acquisition machine, a Dynamic and Personalized Merchandise Mix, and the ability to measure and affect Customer Lifetime Value.
Last week’s press reported retailers visibly quaking in their boots at Amazon’s proposed purchase of Wholefoods. The Amazon machine is not just in the backyard, it’s climbing up to the roof ready to unceremoniously kick the tiles off.
What is to be feared most, the world’s largest assortment, the insanely efficient distribution machine, or the star-fleet of drones being prepared to solve the last mile problem? Amazon’s slogan, ‘The Earth’s Most Customer-Centric Company’ might hold the clue to its not-so-secret sauce. How this principle was applied at the Washington Post provides a playbook for digital transformation.
Transforming the Washington Post
When Jeff Bezos acquired the Washington Post in 2013, the paper was losing revenue and its losses were widening. Bezos told the staff to ‘Stop whining that the Web took publishing away from you, took your business model. It also brought new models’.
Under Bezos, the IT division was doubled to 250, recruiting high-quality talent for a mission to save journalism. In the same way that Tesla shunned off-the-shelf ERP to build their own, the Post developed its own technology that fundamentally changed the process of publishing the news.
Today’s news is delivered in real-time and cannot be held up by download speeds. It provided the paper with deep visibility into users online behaviors – which stories each prefers to read, how far down they scroll, which headlines, images or videos work for which users at what time of day. All this information is delivered in real time back to the news room and used to optimize photos, headlines and even formats for each mobile device. As a result, year-on-year online readership is up 22% and annual digital advertising revenue exceeds $100M.
The Bricks and Mortar Cul-de-Sac
Bricks and mortar retailers are having a hard time adapting their core business processes to the digital world. Senior leadership teams have likely grown up taking title to goods, running operations where inventory and availability were tightly coupled and the job of print and television advertising was to get the customer into the store. Up-ending that model takes most leadership teams well outside their comfort zones which makes digital innovation a struggle at best. Yet the disruption coming from digitally native new entrants such as Wayfair, Zulilly and Warby Parker is an undeniable threat to business as usual. Amazon, Alibaba and Rakuten present another degree of threat.
Creating a Digital Nervous System
The Washington Post is transforming itself by gaining unprecedented insight into its reader while simultaneously building the automated systems to act in real time on those insights. What is the parallel for the retailer looking to escape the dead-end-street that they find themselves in? Studying the Post’s transformation, I see three capabilities required of retailer’s digital nervous system:
1. Anatomy of Acquisition
Today’s customer is continually exposed to campaigns from every brand and retailer chasing their business. Data Management Platforms (DMPs), Ad’ networks and exchanges are helping to improve targeting via current media, but Augmented and Virtual Realities and A.I.-based Conversational Agents are only just getting established as channels.
The Omnichannel retailer’s challenge is now to optimize spend, timing and targeting across all available media. Traditional retailers might not be used to the dynamic environment where there is no static formula that a marketing department can develop and commit. Instead, online marketing becomes a portfolio of experiments that are measured against a set of target outcomes, that feedback data to machine learning algorithms that are continually adjusting audience, timing, content, then making real time purchasing decisions.
2. Customer Value Machine
The aim is to replace the rather one-sided metric of Customer Lifetime Value with one that is deliberately customer-centric. How can the retailer create a machine that generates value for the customer on an increasing scale?
Twenty years ago, the news process involved the curation, research and reporting of events by skilled journalists and editors. News today is captured in real-time and fed directly into myriad distribution channels with little or no editorial. There is more news with more perspectives, delivered instantly but with less guaranteed quality. As a purveyor of the truth, the Post has re-engineered itself to deliver editorial quality at a seemingly impossible speed and scale with an effective and compelling paywall proposition.
Retailers must develop unique value to their customers in a trading environment where switching cost is low and customer loyalty is a fragile concept. There is so much behavioral and transactional data available that identifying, profiling and understanding your customer as a micro-segment is table-stakes. Advantage will come through a machine that injects individual customer insight into every customer touchpoint experience. If a specific customer has low price sensitivity but demands speed of service, every communication, every interaction and every assortment presentation must be programmed with this information. Likewise for customers who are cash-poor and time-rich.
3. Merchandise Telemetry
If the Post is able to optimize the headline, content, images and video of a story for its target customers on the actual devices where they are reading, then why can’t a retailer think about its products in the same way? For example, a clothing retailer should be able to see which micro-segments seeks out garments with different color thread used to sew the buttons on. A furniture retailer should be able to find all pieces that have a certain wood finish at a particular price point and identify all households that have purchased that finish in the last 30 days.
As conversational interfaces become popular we start to see that the product attributes written up are more suited to a visual user interface, yet are too clumsy for the text interfaces of A.I. Agents and the voice interfaces of Amazon Alexa and Google Home.
Today’s merchandising applications are based on a Product Master that provides structured attribution to allow the category to be managed top-down. Yet we are going to need a far more granular and semantic definition that will allow merchandise to be tracked like the particles in the Large Hadron Collider. Armed with systems like this, Data Science can then deliver true competitive advantage.
For Omnichannel retailers with a legacy store portfolio and distribution infrastructure, the transformation required to run digital-first may be too disruptive to their established operations – in a classic Innovator’s Dilemma scenario. Few will succeed and we will witness many established names and trusted brands closing their doors or being acquired by faster digital natives buying the physical presence.