(This is a brief version of the text. You can also download the full paper.)
For a few years now, the themes of multichannel and omnichannel business dominated the retail trend talk. It’s often been communicated as a need for established retailers to provide a seamless experience for the new customer, who is used to shopping online as much as offline. That need is very real.
Less obviously, this shift is causing retailers to compete head to head with companies, whose home turf is online. „Retail” may no longer be an appropriate description of the industry: it is blending with food delivery services, augmented shopping experiences, personal style services as Stitch Fix, and chatbots. When a large part of a family’s grocery shopping can be predicted, shipped and delivered by an autonomous device (van or drone), it can hardly be called retail; it can hardly even be called shopping. Yet, it is a huge part of retail’s future.
It can hardly be called retail; it can hardly be even called shopping. Yet, it is a huge part of retail’s future.
The companies developing these disruptive services are the new retail. Their core workforce isn’t sales clerks, it is software engineers and data analysts. Their ventures were born in the cloud and raised from infancy on technology and savvy use of data. Constant experiments and iteration are part of their DNA. These data-natives are naturally positioned to leverage every technological advantage.
The Burden of Data
Meanwhile, traditional retailers fight an avalanche of data. It is a great and sobering paradox: Data was heralded to them as the new gold. But those industries, that produced the most of it, have faced increasing challenges. The sheer volume of information has swamped management and analysts alike. It is a conservative estimate to say that more than 95% of available data is not immediately useful at any moment. It looks a lot more like ore than gold.
And not only is it of little use. Traditional approaches like reporting and dashboards became clogged with redundant, obvious and repetitive information. In other words, data itself is first and foremost a cost. A financial, technological and cognitive cost.
Mining for Gold
The typical approach to resolve this is building analytics teams, data science teams and educating business people to work with self-service BI tools. In other words, sending people to the mines to bring back gold.
Is that working out? So-so. As Martin Butler writes:
Whether we like it or not, machines will soon be analyzing our data far more efficiently and far more meaningfully than even the most skilled human effort.
Training existing workforce to become more data-driven is unavoidable, but will take years. In the meantime, data-natives like Lyst, In The Style and Polyvore will be miles ahead. To stay competitive, retailers will need to adopt advanced, automated analytics to close down the technological gap.
And there is a lot that can be done today. Over the past year, we have built a number of new use cases that provide tangible outcomes to retailers:
- Strategic summaries from operative data
In essence, a modern BI tool should make everyone in the company immediately see the top 5–10 issues in their area of responsibility, with a single click.
2. Automated Category Analysis
An analysis of the key shifts in a category takes hours and days even when done by a seasoned professional. It cannot be done frequently, because the focus shifts to other categories. It is often the case that a category review is performed 1–2x a year. AI analytics assists in several ways. First, all categories get scanned every week, or even daily. Then, the AI should perform all typical steps that a human analyst does, and present the analyst with a concise summary of the key factors that influence revenue, margin and market share.
3. Store Performance Recommendations
Large organisations can be notoriously slow to evolve. This does not have to be so, however. Those large chains, that turn their whole branch network into an ever- learning, self-improving network, can unleash the massive potential that lies dormant. From store assortment recommendations to layout optimisation, stores begin to learn from each other.
Stores begin to learn from each other.
4. AI-driven Alerts and Exceptions
To further reduce the effort of gathering insights, only the most important signals should be communicated. And they should automatically make their way to the people who can solve them best. Without the need for manual configuration of alerts and benchmarks.
5. Automated Sales Trend Analysis
Downward trends cost a fortune. They can cause overstocking, locked cash, expired goods and a lowered brand perception. In fashion, they mean the difference between a stellar season and a net loss. And if nothing else, they represent a vital signal of a changing customer taste. Upward trends and shooting stars are no less difficult to spot with traditional tools. And though they typically bring revenue growth, left unchecked, they mean lost opportunities. Augmented analytics tools find key trends in sales data, sort them by business value and inform product management and operations alike.
These are not hypothetical scenarios, but examples of real deployments of our augmented analytics platform. We have recently collected them in our retail whitepaper.
You can download the full paper including further use cases.
From a more personal perspective, it’s amusing to see where we were only twelve months ago. Back then, we had more mathematics PhD’s than developers. We had a year of research behind us and began building the Stories platform. Today, we are a data-native, AI-first team, that is proud to be pushing the boundaries of data-driven management.
And therein lies my answer to the initial question of this post: How can retailers fight back against data-native companies? Ideally, by collaborating with them. That is the fastest way to unleash the power of AI on the massive scale of retail operations.