The case study below is a summary of an actual client project CVI executed. The client has generously granted us the right to publish our insights and analytical approach, but has asked us to remove their name and any identifying details.
An international quick service restaurant (QSR) brand with more than 300 stores in Shanghai engaged us to support them using location intelligence to launch a new breakfast offering. The brand’s legacy and its core offering has traditionally been catered to the context of lunch and dinner, so the breakfast day-part has always been an interesting opportunity for the China team to explore and experiment with new ideas.
During recent years, in response to the shifting paradigm of online consumption behaviors and increasing competition, the brand has developed and tested numerous strategies to continue to grow and innovate its breakfast offering. The client shared some of those previous strategies with us (e.g. promotional marketing campaigns, pre-order and pick-up services), many of which they said didn’t hit their ROI targets and were not widely implemented.
Their ask was straightforward — give them new ideas and convince them of its potential ROI. Always seeking for a deeper understanding of things, our starting point was however to first examine the previous concepts and understand why they weren’t successful. We had a hypothesis that the concepts alone were probably not the issue, but rather the strategy of its implementation were not contextual to location, i.e. the right strategy but the wrong store/location.
In this case study, we’ll walk you through the process of how CVI tackled the ask by helping the brand to develop a new understanding of its own stores — not from the insular perspective of whether the store was a flagship or not, but instead from the perspective of its consumers and whether the store was in a location of residence, work, or transit. Using a human-centric + urban city data analytical approach, we ultimately helped the brand to better implement not just one but many of its sales improvement strategies.
Understanding a store’s location
Our hypothesis was that the previous sales improvement concepts were not invalid on their own but that the fault was with its strategic implementation. This was based on a simple fact: the brand always launched and tested those strategies on the same few stores. But it was obvious to us that the brand’s 300+ stores, scattered across the city, were in markedly different types of locations, which we intuitively knew would reflect in different customers behaviors. Yet the brand’s top-down, test-and-learn cycle wasn’t accounting for those store location-type differences. No wonder many of the sales improvement concepts failed.
Our first step was therefore to build a framework of understanding the brand’s stores based on its location-type. A generalized set of location attributes were taken into consideration (some particular attributes were left out for client confidentiality purposes):
- Housing Intensity: total housing units, normalized — this indicates the amount of people living nearby;
- Workplace Intensity: total office floor area, normalized — this indicates the amount of people working nearby;
- Public Transit Convenience: number of subway stations by 10 and bus stations by 1 — this indicates the likelihood of people transferring from home or to work during the morning commute;
- Competitors: total numbers of competitors, weighted based on client inputs — this indicates the level of substitutes/alternatives for people to choose among.
Note: all four factors were weighted by distance, i.e. the further the POI, the lower the score.
After calculating the four attributes of each store, we used a cluster analysis method to categorize the 300+ stores into self-similar subgroups based on those four attributes.
This resulted in seven subgroups, or location-types. We plotted and examined them on the map, and gave each location-type a name based on our interpretation of the urban context. The seven location-types were:
1. Major transit hub
2. Transit-oriented-development (TOD) commercial
3. Suburban mixed-use
4. Subcenter mixed-use
5. Suburban housing
6. Office mixed-use
7. Residential mixed-use
A location-based sales strategy
By normalizing sales performance against the seven location-types, we were able to help the brand discover new insights about its sales strategies.
For example, on average “suburban mixed-use” stores had underperformed on total breakfast sales, which considering the smaller market it serves make sense. However, without normalizing sales performance to this location-type, it would have been easy for management to miss opportunities of tailored sales strategies. “Suburban mixed-use” stores actually outperformed its peers on breakfast sales using pre-order and pickup services, suggesting a strong routine purchase behavior and giving management insights as to what kinds of strategic opportunities to respond with.
Another example was marketing promotions. The brand routinely offers both group dining and single dining promotional incentives (group dining promotions encourages larger average spend versus single dining encourages greater frequency of purchase). However, those promotional incentives were not tailored in any way for the type of consumer nor for the location-type. With this new understanding of store location-types, the brand would be able to execute targeted, geo-fenced promotional campaigns to push group dining promotions in its ‘residential’ location-type stores (for families dining together), and single dining promotions in its ‘major transit hub’ location-type stores (for office workers).
Based on the insights of the location-type analysis, we created a set of location-based sales strategies.
1. Fast pick-up window: For stores with exceptional passing-by pedestrian flow, prioritize to renovate street-facing facade with a pick-up window, where consumers can order online and pick up for optimal convenience. (Continue reading below for our passing-by flow analysis.)
2. Breakfast cart: For stores that aren’t adjacent to high passing-by pedestrian flow, but are still in the vicinity (e.g. opposite side of street, or within 100 meters away), place a food cart on the busy side of the street.
3. Location-based marketing: For stores without strong passing-by pedestrian flows and outside of strong commercial hot spots, prioritize time-sensitive and geo-fenced ads and promotions to drive exposure and overcome physical location constraints.
4. Group incentives: For stores in suburban/subcenter areas, prioritize group promotional campaigns to target group purchasing behaviors.
Beyond this analysis, the location-type framework helped the brand internally to analyze its sales performance and derive key success factors based on the more contextualized understanding of its stores. This framework also served an analytical foundation and baseline by which they could measure the success of its various sales improvement concepts and more generally its overall business.
Hyper-local location-strategy
At CVI, we believe the best insights are derived from a combination of human and data insights, or in other words, from both qualitative and quantitative methodologies. Besides urban data, we also engaged in design research by observing and interviewing consumers to uncover their needs and pain points.
What matters the most to people during breakfast?
In short, the value of convenience is absolute during breakfast — picking up breakfast en route during the morning commute and a short waiting time are far more important than brand, flavor, even type of food.
As we described above, the fast pick-up window was a key sales strategy that takes advantage of pedestrian passing-by flows. But which stores, regardless of location-type, presents the best opportunity to capture these flows specifically for the morning peak hour?
Estimating pedestrian flow
We can generalize a consumer’s morning commute into a two-part journey: from household to public transit, and from public transit to workplace.
For the first part of the journey, we defined an 800m radius area from the center point of a metro station, a simple assumption for a reasonable distance a person is willing to walk to the station. Then we use Baidu’s API to map the shortest walking routes from each residential compound in the area to the station. Weighted by the number of units in each compound, we aggregate the routes to create an overview of the streets most likely used by residents living in the area walking to the metro station in the morning.
We followed the same logic for the second part of the journey, with the difference being the routes are from metro station to office building. In this case, office buildings were weighted by its gross floor area.
One immediate insight this uncovered was that even though one store might be very close to a subway station, it doesn’t necessarily guarantee pedestrian traffic flow. And especially for the context of breakfast, what we heard from consumers was that even if the store was less than a block away, or on the opposite side of the street, they would probably consider an alternative.
Using an open data set released by the Shanghai Metro Authorities, we had historical data for people entering/exiting metro stations across the city. Using the total people traffic on a typical weekday during the morning rush hour of 7:30 to 9:30, we could distribute the traffic proportionally for the routes taken and estimate the number of people passing by each store.
By applying the same calculation for all subway stations near the brand’s stores, we could estimate which stores would have a greater potential for pedestrian flow during the morning commute. This was also compared against the brand’s actual breakfast sales data, which together give a richer understanding of each store’s potential opportunities to improve breakfast sales.
Conclusion
Location intelligence is an analytical approach that can support business management decision-making as well as day-to-day operations. CVI harnesses urban data to augment enterprise data and help organizations make decisions more contextual to cities and consumers, and ultimately more effective for business results. Whether you manage an international chain or a small regional chain, please reach out so we can explore together the potential of location intelligence for your business.
