What is a commercial chain store "cold start" site selection problem?
For example, if I need to open a chain store in a new city (e.g. "Starbucks Coffee"), but there is no Starbucks in this city, and we do not have any experience with operating the Starbucks in this city. Before we set up a new Starbucks, we need to consider the following questions.
- How do I choose a location for my store that will yield the highest operating profit?
- Can I consider using the experience of Starbucks openings in other cities for this new city location?
- Will the experience of other competing stores in the city (e.g. Ruixing and Costa Coffee) also help me choose the right location for my business? But we also need to prevent the location of these brands from competing with mine.
This is the problem of cross-city "cold-start" location selection for retail chains.
Studies have shown that location can determine up to 90% of the variance in revenue for a store in the same city. We know that the suitability of a store in a region will be affected by the different cities' business environment, socio-economic, competitive brands, the behavior of the residents, and the residents' hobbies (such as eating habits: Guangdong do not eat spicily, Jiangxi do not eat sweet, Jiangsu and Zhejiang have a mild taste) and so on.
We can use big data to mine the urban variables, including socio-economic and business environment, and abstract the business experience of similar stores (competitive brands) already existing in the city into a mathematical model. We propose a brand new approach to artificial intelligence and knowledge migration that can effectively solve a variety of intra- and inter-city commercial store location problems, including the "cold start" problem.
We will soon launch our data-supported computational and consulting services.
Here is an example.
We have revenue information for three milk tea shops in Guangzhou (Da Kashi, Gong Cha and Royal Tea) and multi-source spatial data for Guangzhou, so I assume that there is no revenue for one of the shops in Shenzhen, but there is revenue for the other two shops.
For example, I assume that there is no Royal Tea store in Shenzhen, then I use the multi-source spatial variables, which can be used to indicate the business environment, socio-economic, residents behavioral activities, in Guangzhou and the sales data of the three milk tea stores; combine the multi-source spatial variables in Shenzhen, and the sales data of Da Kashi and Gong Cha, and use a combination of GIS, artificial intelligence and knowledge migration to infer the theoretical value of the sales of Royal Tea in Shenzhen after choosing the location in each community (or the suitability of the location).
Figure 1 shows the site selection results in downtown Shenzhen City. After comparison with the actual sales data, the accuracy of each store can be up to more than 90%.
Figure 1 Cold start site selection results for Da Kashi, Gong Cha and Royal Tea in downtown Shenzhen based on our model. From top to down: Da Kashi, Gong Cha and Royal Tea. (A) Baoan airport, (B) Shenzhen University, (C) Coastal City Shopping Plaza, (D) Urban Garden and (E) People's Park.
From Figure 1, it is obvious that we have explored the community-level suitability (sales volume) of different brands and explored the business characteristics of other cities and other stores accurately.
Due to different levels of regional economic development, the spatial distribution of competitive brand shops distribution and crowd consumption habits, different brands of milk tea shops in different communities have different site suitability and consumer trends.
The urban residential area (Urban Garden) is suitable for all milk tea shop locations. The university area (Shenzhen University) is best suited for the relatively inexpensive Gung Cha and Da Kashi.
The commercial center (Coastal City) and the airport are the most suitable for the higher consumption of Royal Tea to start a chain. Even if the cost of opening a store is high, Royal Tea can make the most profit due to the high traffic and high level of customer spending.
In the vicinity of People's Park, according to the distribution of different consumer attributes near the park, it can be found that the three chain milk tea shops correspond to different suitable locations.
Yao, Y., Liu, P., Hong, Y., Liang, Z., Wang, R., Guan, Q., & Chen, J. (2019). Fine-scale intra- and inter- city commercial store site recommendations using knowledge transfer. Transactions in GIS, 23(5), 1029-1047.