Boosting International CPG Market Share with Location Intelligence


Unlock Global CPG Growth with Location Intelligence! Overcome challenges, target the right markets, cater to preferences & boost revenue.

This post may describe functionality for an old version of CARTO. Find out about the latest and cloud-native version here.
Boosting International CPG Market Share with Location Intelligence

Consumer Brands - Challenges to Global Growth

Consumer brand businesses are facing significant challenges that are hindering their global expansion progress, resulting in less than a 2% growth year after year (Polen Capital). A few of the issues consumer brand businesses face when growing globally include: 

  • Informal market:  There are millions of small, independent points of sale with limited shelf space. This makes it difficult to compete with well-established brands and to obtain reliable data.
  • Diverse consumer preferences:  Consumer preferences vary greatly from country to country. Businesses must understand these preferences and align their products accordingly.
  • Inaccurate data:  Relying on unreliable data can lead to poor decisions and lost opportunities.

These challenges can make it difficult for consumer brands to achieve their growth goals. However, by addressing these challenges, businesses can position themselves for success in the global market.

Growth Opportunities Through Better Data

While growth challenges persist globally for consumer brands, the potential far outweighs these obstacles. Non-western markets alone contribute to more than 50% of consumer goods spending, representing an astounding market potential of $6 trillion (Consumer Goods Technology). To capitalize on these opportunities, companies must place a strong emphasis on accurate data, as it serves as the foundation for creating precise forecasts and formulating effective go-to-market (GTM) strategies.

Accurate data enables CPG businesses to:

  • Identify new channels and markets.  By analyzing local data, businesses can identify emerging or untapped distribution channels that hold promise for market expansion. This knowledge allows businesses to strategically extend their reach and establish a strong presence in previously unexplored areas.
  • Focus sales efforts on stores that are more likely to convert.  By understanding the purchasing patterns, consumer preferences, and market dynamics specific to each location, businesses can prioritize resources and tailor sales and trade marketing strategies to target stores with higher sales potential. This enhances sales promotion efficiency and maximizes revenue.
  • Identify and understand new customer segments.  Consumer brands can identify and understand distinct consumer groups within each market, to tailor marketing messages and product offerings that resonate with the customer’s preferences and needs.
  • Better cater to local tastes, preferences, and market demands.  This alignment enhances customer satisfaction, boosts brand loyalty, and strengthens market position.

What is POI data?

POI (Point of Interest) datasets contain notable locations, and can include businesses, landmarks, tourist attractions, parks, restaurants, hospitals, and more!  Data can also contain additional information associated with each point, such as the name, category, address, contact details, opening hours, ratings, reviews, and accessibility. 

For instance, the map below (open full screen here), leverages dataplor’s POI data in Mérida Centro, Mexico. Clicking the individual locations allows you to explore the variables available, including categories, contact details and opening times.

POI data use cases

Consumer brand businesses can leverage Point of Interest (POI) data in various ways to drive global growth. One key application is merchant prioritization, where businesses can identify the most promising locations to focus their business development efforts. By conducting an inventory of existing channels and identifying gaps in data, businesses can enhance their information by collaborating with existing vendors to obtain geocodes, addresses, and operating hours. This enriched data helps in identifying the best merchants to sell their products and to pinpoint new store opportunities where success has been observed. Prioritization can be based on metrics like popularity indices or dwell times, ensuring that the team's focus is directed toward the most promising points of sale for optimal results.

Another valuable use case is global trade area analysis. By analyzing market coverage, CPG brands can gain insights on the performance and market potential for different point of sale trade areas. By analyzing global market coverage, businesses gain insights into the performance and potential of different trade areas. Combining this analysis with in-depth data allows for the creation and prioritization of trade areas based on demographic and economic characteristics. The enriched data helps businesses identify clusters of locations where stocking their products would have the most impact, facilitating strategic territory expansion.

Now let’s see some of this in action!

POI data in action!

Let's imagine we’re a beverage company who wants to stock our drinks in more stores across Mexico City, and need to find a way of identifying which stores are likely to generate the most revenue. In this use case, we have two distinct products: 

  • The budget drink:  Low cost, broad appeal.
  • The premium drink:  High cost, targeted towards affluent young adults.

The steps for this analysis leverage the CPG module in our Analytics Toolbox, and are outlined below. If you’d like to follow along, make sure you sign up for a free 14-day CARTO trial! You can also watch our experts talking through this process in this webinar.

A diagram showing the steps outlined in this blog post

Step 1: Define merchants

The first step is to select our potential merchants using the dataplor places dataset, available via our Spatial Data Catalog. For the purposes of this demonstration, we randomly selected 500 of the 52,452 grocery & convenience stores across Mexico City (pictured below), and will base our analysis on these.

A map showing Mexico City with POIs shown as purple dots on a black basemap

Step 2: Conduct trade area analysis

In order to understand the likely customer composition for each store, we need to define a catchment or “trade area” for them. A trade area estimates the geographic area around each point of sale that contains the resident population most likely to use each store.

For this, we’ll use the Generate trade areas function to create an 800-meter buffer (approximately a 10-minute walk) from each store. This tool offers various other parameters - such as walk or drive time isochrones - to reflect realistic transportation network conditions.

The SQL query for this analysis is below - you can find more information about this in our documentation.

   '''SELECT store_id,geom from `cartodb.dataplor.merchants` where store_id is not null and st_astext(geom) is not null''',

Check out the results of this below!

A map showing 800 meter buffers as purple circles

Step 3: Enrich trade areas with demographic & economic data

Now we have estimated where customers are coming from for each potential merchant, we can establish who those customers are. 

To do that, we’ll enrich each catchment with consumer data. While we offer a host of enrichment functions to allow users to aggregate data between different geographies, in this example we’ll use the Customer Segmentation Analysis Data function. Customer segmentation groups points of sale based on characteristics of their trade areas such as demographics, behavior and spend. This is a common form of analysis across CPG to help target resources and therefore reduce costs.

This function has been designed specifically for CPG use cases, and feeds directly into the next stage of this analysis - running customer segmentation. In addition to generating an enriched trade areas table, it also generates a correlation and descriptive statistics table. 

We’ll use the following data sources for this enrichment, all of which are available from our Spatial Data Catalog:

So let’s take a look at the code!

-- Input trade areas
SELECT geom, store_id, footfall from cartodb.dataplor.merchants_buffer_800m
-- Data Observatory enrichment
  [('age_18-24', 'sum'),('Page_15-64','sum'), ('spend-food-drinks', 'avg'),('disposable-income','avg')],
-- Data Observatory connection ID
  -- Custom data enrichment
  SELECT geom, store_id, footfall_avg from carto.dataplor.merchants_buffer_footfall
  -- Custom data variables
   [('footfall', 'avg'),
-- Output location

Explore the results below, or open in full screen here!

Step 4: Cluster customers

The final stage of this analysis is to use the Run Customer Segmentation tool. This tool uses Principal Component Analysis (PCA) to divide our merchants into clusters based on the characteristics they share from our previous enrichment.

The SQL code template for this is below - check out our documentation for a full guide to this! This will run 4 clustering scenarios, with each scenario generating the number of specified segments (here that’s 4, 5, 6 and 7).

   '.._enrich' -- enriched table,
   [4,5,6,7], -- number of segments
   0.9, -- pca explainer ratio

Step 4.1: Scenario testing

This process generates a series of tables, the first of which we need to interrogate is suffixed _clusters_stats. This will tell us which of our clustering scenarios best represents our data, based on a Davies-Bouldin Index (where a lower number represents a less scattered - i.e. more similar - cluster) and a Mean Squared Distance Index (where a lower number represents more distance between clusters - i.e. that different clusters are more distinct). 

In our example, our 7-segment scenario generates the lowest Mean Squared Distance of 1.364, so we will proceed to use this scenario for the remainder of our analysis.

Step 4.2: Examining our clusters

These seven clusters can be explored on the map above (or in full screen here). Make sure you also explore the categorical widgets to help understand the different characteristics of each cluster! You can also explore the average values for each cluster in the table below.



Age 18-24

Age 15-64

Disposable Income

Food & drink spend











































We can see that clusters 1, 3 and 5 have typically higher values across all of our criteria variables. The only real exception is cluster 5, for which local residents have a relatively low disposable income. However, this is countered by their high spend on food & drink - ideal for our beverage company! Merchants that fall into these three clusters would be ideal to stock our premium beverage. 

Merchants in clusters 6 & 7 typically have trade areas which score roughly in the middle across all of our variables, and so may be good options for our more mass-market, budget  beverage. 

Finally, merchants in clusters 4 and - in particular - 2 - score lower across all of our key variables, and so it is unlikely we could generate a strong revenue in these stores.


Very few CPG companies are embracing spatial insights to drive their growth initiatives, and doing so gives a clear competitive advantage. Businesses must begin by harnessing high-quality foundational data and leverage these data-driven techniques to fully understand both their consumers and merchant networks. 

Ready to see how you can leverage the power of POI data in your work? Sign up for a free 14-day trial today!