Self-storage industry — a curious case in marketing analytics
In this blog, I am going to take you through the journey of how my current practicum project with one of the world’s largest self-storage companies brought about a paradigm shift in the perspective I had on carrying out an analytics project. I will focus on:
- Where it all began
- Reality check
- Enter self-storage
- Tackling the problem
- Conclusion
Where it all began
Data Science was deemed as the sexiest job of the 21st century in 2012 and rightfully so. I joined the bandwagon that is the analytics industry in 2015 with the world’s largest pure-play big data analytics and decision sciences company. I was struck by the bling of the buzzwords — Machine Learning, Deep Learning, Random Forests, and the lot. I wanted it all under my belt. My journey in analytics was customer-oriented. I enjoyed a steep learning curve — generating purchase propensity of customers for a sportswear manufacturer, ideating and defining the right marketing mix for a pharmaceutical giant, creating a customer 360 for a large retailer. I thought I had everything in my arsenal to create treasure out of any messy, fuzzy data.
Data Scientists…people who can coax treasure out of messy, unstructured data.
— Thomas H. Davenport and D.J. Patil
While working on these projects, I had the opportunity to work across the spectrum of analytics: descriptive, predictive, and prescriptive. I had gotten a taste of Data Science. I believed in the power of what historical data could do for you!
Reality check
When I embarked on the path to analyze the customer journey of the self-storage giant, I was confident that I brought to the table experience in defining, inventing, and testing customer-centric strategies. I thought I could help them with marketing strategies using just analytics but boy, was I wrong. All the previous customer relationship management (CRM) projects that I was part of, had a common workflow. They had their websites, their customers would pour in to check out the products, show their digital intent by adding products to digital carts, and if they like the products they would purchase them. Simple.
The model that I had implemented for the organizations in the past was:
1) You attract strangers who make up this group called the prospects using advertising, digital media, print media, etc.
2) They develop a taste for your product and become regular visitors of your website
3) Slowly, they are hooked on to your brand and start generating leads by responding to your campaigns
4) You can close deals with some of them and they become your customers
5) Once they are delighted by the experience of your products and the medium, they are coerced into becoming promoters of your brand
Since most of these companies collect a treasure trove of information on their customers, analytics is able to predict, with good precision, what these customers would be interested in and what could their next steps be. You can analyze patterns in the way the customers engage with your website.
But, what happens when the customer decides to show up on your website without any digital footprint in the past? How do you gauge their next steps?
Enter self-storage
To gauge how different the interaction of self-storage customers is with their websites, we need to understand what drives the industry. Self-storage is majorly driven by life-altering events more popularly known as the 4D’s of self-storage — death, divorce, displacement, and downsizing. These are sporadic events and are very difficult to model based on conventional ways. To further understand this, take the example of Amazon and U-Haul. How many times have you found yourself window shopping on Amazon? How many times have you found yourself window shopping at U-Haul? On Amazon, people look for multiple products across categories, add products they like to their carts. This leaves behind a digital footprint that can then be utilized to understand the customer’s likes and dislikes. U-Haul, on the other hand, would not see its customers window shopping for spaces (well, there are always some). This difference poses several challenges for the self-storage companies to acquire new customers and make marketing decisions.
Tackling the challenge
When my team and I started working on our current project with a self-storage company, we realized during the Project definition phase that the problem at hand required more than just creating state-of-the-art machine learning models. It required us to keep business requirements at the helm of the project objectives. In the 10 months of working on the project, we followed the entire life cycle of an analytics project:
Project definition: We interviewed the stakeholders to fully empathize with what their asks were and used design thinking to organize our work and ensure quality management.
Invention: We prototyped several machine learning algorithms to come up with meaningful customer segments.
Testing: We are iteratively validating our results with the benchmarks set by the business.
Business use: Focusing on the business value proposition more than the technical computerese has to be the single biggest takeaway for our team.
Conclusion
The last 10 months have been a constant learning and unlearning of the concepts of how-to and how not to tackle analytics projects.
Learning and unlearning are required to move forward!
Building fancy machine learning algorithms is not the endgame. Organizations are looking for people who could provide them with thought leadership. The algorithms could be tools to get you there not the end results by themselves. From admiring the glitz of the fledgling analytics industry to realizing the importance of keeping things simple and focusing on generating value, I have come a long way. Miles to go before I sleep…
