Ben Schein from Domo and Ryan Squire from SafeGraph discuss how to combine first-party and third-party data to decide if and where to close stores.
Ben Schein from Domo and Ryan Squire from SafeGraph discuss how to combine first-party and third-party data to decide if and where to close stores.
Early 2020 saw a massive downturn in economic activity, as health and safety regulations were enacted to combat the COVID-19 pandemic. Some businesses were left with no choice but to permanently close down some of their operations. The question is: in a time of economic uncertainty such as this, how does a business prioritize which locations to keep open or eventually re-open, and which ones to shutter permanently?
This process, known as “site deselection”, is the focus of this webinar co-hosted by Ryan Squire, Senior Data Scientist at SafeGraph; and Ben Schein, VP of Data Curiosity at Domo. Ryan and Ben discuss and demonstrate how a company can enrich their first-party data with third-party data to make more informed decisions on where to open or close stores. Here’s a brief summary of what’s inside:
The first step to understanding how to build a data-driven site deselection strategy is to understand what site deselection means. So we’ll start there.
Site deselection is the process of a company choosing to temporarily or permanently shut down operations in certain locations, in order to cut costs. It has less to do with choosing the number of stores to close as it does with deciding which specific stores should be closed or kept open.
Also stylized as site de-selection, this process requires just as much thought – and many of the same considerations – as a plan for selecting new store sites. That’s why companies are increasingly reliant on data to form strategies for deselecting retail sites, especially in tumultuous economic times. The COVID-19 pandemic has offered a stark case in point: many companies have had to perform site deselection in the face of shutdowns, capacity limits, and declining foot traffic due to public health concerns.
So how does a company go about building a data-driven site deselection strategy for when the situation calls for it? Part of it is recognizing what kinds of data are needed, as well as what information each kind of data can (or can’t) provide. The other part is understanding how site deselection relates to site selection, as well as to a broader data-driven business intelligence system for the company. Here are five quick points that break things down.
Time in Video: 6:00
Though site selection and site deselection are seemingly opposite processes, they share many similarities as well. Both can have a tremendous impact on a company’s success or failure, especially relative to its competitors. This is because both have significant costs and risks associated with them. And because of that, they both rely on solid strategies based on accurate data in order to eliminate as much risk as possible.
Time in Video: 7:00
A good site deselection strategy requires an ample amount of first-party data, regarding both the company as a whole and the stores under consideration. You need to know when and where people are spending their money at your stores, and weigh that against things like how much it costs to get new products in stock. But it can also be helpful to consider third-party data, such as local foot traffic or spending behavior, to understand a store’s potential for business and how well it’s attracting customers compared to nearby rivals.
Time in Video: 13:38
Site deselection, like other forms of business intelligence, is a process that involves steps like connecting to data sources, storing data, transforming data into visualizations, and collaborating to make predictions or take action based on data insights. Ideally, it becomes part of a company-wide system of “BI leverage” that is flexible enough to adapt to unknown needs, allows all parts of the organization to work with data on their own, and ensures responsible access.
Time in Video: 25:35
It can be useful to consider how the outside perspectives of third-party data could address some of your first-party data’s potential blind spots. It may account for some variables you hadn’t initially factored in, such as demographic distributions and lifestyles, or overall economic activity within a trade area. Third-party data may even be able to provide hints about competitor movements, which could point to hidden opportunities for some stores over others.
Time in Video: 27:33
It’s important to have two approaches to retail site selection and deselection: a more data-driven one, and a more conversational one. The data-driven approach takes more precedence at bigger retail chains, where calculations and algorithms have to be iteratively applied to a large number of stores. But it is also important to take a more explanatory and visual approach, especially to a decision as impactful as opening or closing a store. Discussing in more natural language the factors that went into deciding on the calculations and algorithms helps other stakeholders better understand and evaluate the justification behind selecting or deselecting a site.
First-party data is internal data that a company generates and collects as part of its normal operations. Third-party data, in contrast, is data a company purchases (or otherwise acquires) after it is aggregated from multiple outside sources.
First-party data includes things like sales/transaction records, customer loyalty program metrics, supplier costs, and ecommerce stats. Third-party data includes geospatial data like information on points of interest, building and property layouts; and measurements of footfall traffic around an area.
Business intelligence is typically characterized as an inward-facing corporate process, meant to improve a company’s productivity and retain talented employees while reducing wasted resources. In that sense, first-party data is essential for business intelligence, because it’s data directly about the company’s operations. So it’s data that a company can usually access easily because it’s already on hand, or can be created if needed (by surveying employees about the company’s ecosystem, for instance).
First-party data has a smaller – but still significant – role in regard to competitive intelligence. Competitive intelligence refers to developing data-based strategies to outperform competitors, or to adapt to changing market conditions better than them. So it mainly uses third-party data to get a sense of the state of an industry or market, as well as how rival companies are positioning themselves. However, first-party data can still be useful for competitive intelligence to establish a baseline for how a company is currently performing. This lets it see the ways in which it’s already outperforming rival companies, as well as what and how much it needs to do to compete with them in other areas.
As valuable as first-party data is for site deselection, it’s not always enough. Sometimes you’ll find your company doesn’t have the exact type of data you need, or at least not in an accessible format. Or the data may be incomplete because some of your locations use different systems to collect and process their data.
Another hurdle is finding second-party information on competing companies – or even complementary ones – that have locations near your own stores. It’s often protected and only released when required by law. This can make it difficult to weigh a company’s position against the state of the industry or the economy.
