In her latest exclusive for WhichPLM, Elizabeth Shobert, Director of Marketing & Digital Strategy for StyleSage, shares her thoughts on latest hot topic: artificial intelligence. Elizabeth explores various ways in which we can utilise AI in fashion, and suggests success lies with starting small.
As we near the end of the year, it’s a good moment to look back at the headlines and topics that have consumed the retail space over the past 10 months. Amazon’s domination, the cratering of the US retail market, and the continued disruption of technology have monopolized our attention and kept us wondering (and sometimes fearing) what might be around the corner.
If we look at those topics, Amazon and ongoing retail bankruptcies are understandably creating a sense of industry-wide concern, while retail tech remains talked about, yet not always widely understood and applied. And yet its applications, specifically those centering on artificial intelligence, when utilized properly, present the biggest opportunity and driver for change.
So let’s take a moment to talk about artificial intelligence. How is it being used? How can companies be smart and future-focused in their applications of it? And what’s the future of it in the retail space?
Let’s Take a Step Back
This is a safe space, we promise. So let’s be honest, how much do you really understand about AI? The good news is that you don’t need a data science degree or tech background to really get it, and what’s more important is to understand its benefits and the ROI it can provide. Let’s start with the basics – what is intelligence? As a human with intelligence, you have knowledge and experience which you use on a daily basis to make decisions and solve problems. It goes from the most mundane (which route to take to work) to more complex (how do I explain a new concept to someone). This is what’s called natural intelligence, and as one gains more experience and insight with age, their relative ability to make better and more informed decisions grows. Now if you think about artificial, i.e. machine, intelligence, it functions in the same way. This machine takes in different information (data) and learns from it just like the human brain would, and with more data can make more accurate and educated decisions. The key point to make here is that in order to get the right decision, both humans and machines need the right inputs. In the world of AI, this is data, in humans, well, thanks Mom and Dad.
Now of course, just like humans make errors in judgement, so do machines. And artificial intelligence is still very much in its nascent stages, particularly as it relates to the retail business. So let’s briefly talk about how AI and the data powering it are being applied today.
You Can Deliver A Personalized Experience
As a consumer, we are all generally aware of how much data online retailers are collecting about us. (Those product recommendations can be eerily on-point.) Indeed, tracking online behavior is one of the lowest hanging fruits of the data analytics world and, armed with this, many retailers are taking their customer experiences to the next level with this business intelligence. Retail darling Warby Parker is a prime example of the disruptive nature of a business built on analytics. Take for example, their home try-on program, which guides users through the product selection process taking into account variables around face shape and frame size. Their calibrated smart recommendations (powered by AI) ensure a great customer experience and identify future product development and marketing opportunities. As their head of Data Science, Carl Anderson, stated, ‘Our goal is to have data embedded into all of our business processes as much as possible. A data-driven culture is objective, tests assumptions, and challenges new ideas to prove themselves.’ It’s this foundation in data that has, in no small way, driven Warby Parker to its valuation north of $1B.
When you think about demand prediction, it’s helpful to bucket it three ways -what’s happening right now, in the medium-term, and long-term -while also recognizing the synergies between each. Let’s take an example in the present, like the weather. Completely out of our control, yet it can have a massive impact on what does and doesn’t sell. One retailer successfully utilizing this particular type of data is Tesco, which logs the weather forecasts and adjusts sales forecasts locally three times per day. In fact, they have been able to save more than £6M a year by getting in front of this very type of demand pattern. By continually adjusting for these environmental factors, retailers can avoid a catastrophic pile-up of excess stock they need to get rid of …after it’s too late. Who said mother nature has to have the last word?
Pricing + Promotional Strategies
Price is undoubtedly one of the most important factors in consumer decision-making, and it’s equally critical to a retailer’s bottom line. And while we have seen increasing interest in developing a structured strategy around pricing, it is still viewed by many retail organizations as a ‘one-off’ and dealt with in a reactionary manner. It’s more often than not, ‘this isn’t selling, we had better take a big markdown to move stock’ kind of move. But the reality is, as one BCG report cited, ‘30 to 50 percent of promotions have no positive impact on sales and margins. Even worse, many of them reduce profits without leading to additional sales.’ Just take a moment to consider all the different levers you have in your toolkit, including starting price point, promotions, and discounts. Once you start tracking these metrics (and those of your competitors) using AI-powered pricing and promotion optimization tools, you are able to see precisely when and how much you should be pricing and promoting, so that you’re not leaving money on the table.
Design Metrics From The Top Down
So we’ve talked about some of the ways in which AI can power personalization, help predict supply and demand, and streamline operations, but how does it really fit into your organization? Okay, yes, having AI and data is interesting, but without being linked back to the organizational goals, what’s the point? If you’re just grabbing whatever is already out there and hoping someone internally finds it interesting, you’ve just wasted time and resources that could have been allocated elsewhere. What drives customers to your brand? Is it getting the latest designs, the best price, or having a seamless omni-channel experience? Whatever that core value or USP is, be sure to prioritize and tie your AI initiatives straight back to these organizational metrics. Moreover, this ensures that the data driving the AI is woven into the fabric of every level of the organization.
Make Sure You Have Data In Place Before You Need It
One of the challenges of engineering an AI ecosystem for success is corralling data that is of good quality before you need it. Though it’s not unusual for companies to augment gaps in their data sets with that from external sources, in order to predict tomorrow’s behavior, you need to harness relevant current and historic data. Identify the who or what (which segments of customers or categories of products), the where (which geographies or places), and when (a specific point in time or series of time) that will guide your analyses and outcomes.
And by starting small, that doesn’t mean there’s not a major benefit to the business. Remember that AI requires not only the right data sets but the right manpower and business support in place to drive it. When you focus on the previously surfaced idea that data be tied back to the most important organizational goals, you are able to focus and allocate limited resources. Moreover, smaller, less disruptive efforts that reap clear results give larger initiatives more internal traction.
The Great Unknown
So where does AI go from here? That’s the part that many who sit on both the technology and business side are trying to both wrap their minds around and build out their strategies on. It will be disruptive, and we know we’re only brushing the surface of the ways in which it can power and transform the way we do business. We’re now using it for personalizing product recommendations, but what about the future where AI harnesses unstructured data sets and, as a result, can create entirely new products and services? We’re not there yet – but we will be in the not-so-distant future. I started this piece talking about fear and the great unknown, yet there’s no going back. The only way forward with AI is by starting small, iterating, and sharing insights throughout your organization.