Location intelligence and geomarketing glossary

Also: holdout test, out-of-sample validation

A blind test measures a model’s accuracy using data the model has never seen. A group of stores or months is set aside, the model is trained without them, and its estimates are then compared with actual results. It is the most honest way to know how a model will perform on new cases, because it keeps the model from being right only because it already knew the answer.

At Pulster: 96% accuracy forecasting first-year sales at existing stores, validated in a blind test with a leading quick-service restaurant chain. The model was not trained on those stores. The average error per store was under 4%. This test was run on existing stores and does not guarantee the same accuracy for new stores.

See also:Margin of errorSales forecasting

Learn more:Data and methodology: where our data comes from →Predictive models: sales forecasting →

Also: store cannibalization, sales cannibalization

Cannibalization happens when a new store takes customers and sales from another store of the same brand. It occurs when their trade areas overlap and serve the same people. Estimating it before opening shows whether the new store adds sales to the chain or just moves them around. Some cannibalization can be acceptable if the sales the new store adds to the chain are enough to justify the investment.

At Pulster: Pulster NOVA measures how much of the trade area two stores share and how many visitors they have in common, and estimates how much sales could drop before you decide to open.

See also:Incremental and transferred salesTrade area

Learn more:Store cannibalization: what it is and how to measure it before you open →Pulster NOVA: retail site selection →

Also: competitive dominance map, dominance map

A dominance zone is the area where a store or brand gets more visits than its competitors. A competitive dominance map divides a city by who leads in each zone and shows where customers are contested. It is used to decide where to defend what you have, where to launch a campaign, and where a new store would face the most competition.

At Pulster: Pulster RADAR shows which areas you lead and which areas your competitors lead.

See also:Visit shareLocal store marketing (LSM)

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: shared customers, shared visitors

Cross-shopping is the share of a store’s visitors who also visit another store, usually a competitor. For example, if 60% of your visitors also go to a competitor, there are customers in play that a promotion or a change in your offer could win or lose. It also helps find complementary brands, for example when planning a shopping center’s tenant mix.

At Pulster: Pulster RADAR shows how many of your visitors also go to competitors, and which ones.

See also:Visit shareDirect and indirect competition

Learn more:Pulster RADAR: competitor and foot traffic analysis →Foot traffic and visit share: how to measure your competitors →

Also: direct competitors, indirect competitors

Direct competitors sell the same thing to the same customers, such as two fried chicken chains. Indirect competitors meet the same need in a different way, such as a pizzeria or a supermarket with prepared food compared with that chain. A sound location analysis measures both, because both compete for the spending of people in the area.

At Pulster: Pulster updates business data every month to show who you compete with in each area and who just opened.

See also:Cross-shoppingVisit share

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: viability study

A feasibility study assesses whether a project, such as a new store, is viable before investing. It reviews the market, the site, permits, construction and operating costs, and the expected return. It is usually done once, with fieldwork and interviews, and gives a snapshot of one moment. Location data helps with the market part, but it does not replace the legal or financial review.

At Pulster: A traditional market study takes 2 to 3 months and costs US$10,000 to 75,000. With Pulster, a strategy analysis takes about a day once your company is set up. Based on our experience in the region.

See also:Site selection

Learn more:Traditional market research vs. location intelligence: when to use each →Pulster NOVA: retail site selection →

Also: daytime population, non-resident population

Floating population refers to people who spend the day in an area without living there: those who work, study, shop, or pass through it. In office or retail districts, it can be much larger than the resident population. Both matter for a store: residents sustain frequent purchases, and the floating population explains activity during working hours.

At Pulster: When evaluating a site, Pulster NOVA shows the resident population within a 10-minute drive and the floating population, in average daily visits.

See also:Population densityPoint of interest (POI)

Learn more:Pulster NOVA: retail site selection →

Also: visitor traffic, store visits

Foot traffic is the number of people who visit or pass by a place over a period. Despite the name, in retail it includes people who arrive on foot, by car, or by public transit. It is measured with door counters or with mobility data. It is used to compare stores, times of day, and areas, and to spot trends.

At Pulster: Pulster RADAR estimates visits with aggregated, anonymized mobility data. Mobility data is updated every two weeks and business data every month.

See also:Mobility data (aggregated and anonymized)Visit share

Learn more:Foot traffic and visit share: how to measure your competitors →Pulster RADAR: competitor and foot traffic analysis →

Also: address geocoding, reverse geocoding

Geocoding converts a written address into coordinates on a map; reverse geocoding does the opposite. It is the first step in analyzing stores, customers, or competitors. In several Central American cities, addresses are given with landmarks, such as "two blocks north of the church", so each point is worth checking: an error of a few blocks can change a store’s trade area.

See also:GeofencePoint of interest (POI)

Learn more:Data and methodology: where our data comes from →

Also: store polygon, building footprint

A geofence is a polygon drawn on a map that marks the boundaries of a place, such as a store, its parking lot, or a shopping center. In mobility analysis, it is used to count a visit when a signal falls inside the polygon and stays for a certain time. A well-drawn geofence avoids confusing people who enter the store with people who only walk or drive past it.

At Pulster: Pulster maps where each business is, its footprint on the map, and its format.

See also:Mobility data (aggregated and anonymized)Geocoding

Learn more:Data and methodology: where our data comes from →

Also: location-based marketing

Geomarketing is the use of location data to make business decisions: where to open a store, who to target with a campaign, and who each location competes with. It combines maps with population, business, mobility, and sales data to answer questions a spreadsheet cannot, such as which areas have the right customer profile but no store from the brand yet.

At Pulster: Pulster NOVA applies geomarketing to decide where to open, and Pulster RADAR applies it to see who wins the visits around each store.

See also:Location intelligenceSite selection

Learn more:What is geomarketing? Uses and examples in retail →Pulster NOVA: retail site selection →Pulster RADAR: competitor and foot traffic analysis →

Also: Huff model, retail gravity model

A gravity model estimates how the customers in an area split across several stores. It assumes a store attracts more customers the larger or more attractive it is, and fewer the farther away it is. The best known is the Huff model, proposed by David Huff in 1963, which expresses this as the probability that people in each area choose each store. It is used to estimate market share, trade areas, and cannibalization.

See also:Trade areaCannibalization

Learn more:Store trade area: what it is and how to calculate it →Predictive models: sales forecasting →

Also: density map

A heat map uses color to show where something is concentrated: visits, customers, sales, or competitors. The most intense colors indicate the highest concentration. It shows at a glance where a store’s visitors come from or where commercial activity is strongest. It is a visual tool: to make decisions, pair it with figures by area.

At Pulster: Pulster RADAR uses heat maps to show which areas your visitors come from.

See also:Visitor originSpatial autocorrelation and cluster analysis

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: sales transfer, net new sales

A new store’s sales have two parts. Incremental sales are the ones the chain did not have before: new customers or customers won from competitors. Transferred sales are the ones that move over from other stores of the same brand, which is cannibalization. To decide whether to open, what matters is incremental sales, not the new store’s total sales.

See also:CannibalizationSales forecasting

Learn more:Store cannibalization: what it is and how to measure it before you open →How to forecast sales for a new store →

Also: drive-time area, drive-time polygon

An isochrone is a line on a map connecting every point that can be reached in the same amount of time from a location, such as 5, 10, or 15 minutes by car or on foot. The area it encloses is used to define a store’s trade area. Unlike a fixed radius, it follows the street network and real barriers. When it is measured in distance along the streets instead of time, it is called an isodistance.

At Pulster: Pulster trade areas use typical travel times; actual traffic varies by time of day.

See also:Trade area

Learn more:Store trade area: what it is and how to calculate it →Data and methodology: where our data comes from →

Also: lost visitors, visitor churn

Lapsed visitors are people who used to visit a store regularly and stopped during a period. Knowing how many there are, which areas they came from, and whether they now visit a competitor helps explain a drop in sales and decide where to act. They are analyzed as a group, never person by person.

At Pulster: Pulster RADAR shows how many visitors return and how many have stopped coming.

See also:Repeat visitorsCross-shopping

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: LSM, neighborhood marketing

Local store marketing, or LSM, covers marketing actions designed for the trade area of a single store rather than the whole brand: flyers, partnerships with nearby businesses, billboards, time-of-day promotions, or digital ads targeted by area. It works best when it is based on data: which areas to target, on which days, and against which competitor.

At Pulster: Pulster RADAR answers who is taking your customers, where, and when. The campaign is yours; Pulster shows you where and when to launch it.

See also:Competitive dominance zoneVisitor origin

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: location analytics, spatial analytics

Location intelligence turns location-based data into business answers. It brings layers such as population, businesses, mobility, and points of interest together on one map and analyzes them to compare areas, evaluate sites, or benchmark competitors. It is often used as a synonym for geomarketing, although it covers more fields, such as logistics, real estate, and urban planning.

At Pulster: Pulster works with six data layers: population and socioeconomic profile, businesses, mobility, trade areas, points of interest, and, in some countries, building permits.

See also:GeomarketingMobility data (aggregated and anonymized)

Learn more:Data and methodology: where our data comes from →Traditional market research vs. location intelligence: when to use each →

Also: forecast error, estimation error

The margin of error shows how far an estimate typically is from the actual value. In sales forecasting, it is calculated by comparing what the model estimated with what actually happened, using data the model did not learn from. For example, a US$1 million estimate with a 10% margin reads as a range of US$900,000 to US$1.1 million, not as an exact figure, and the actual result can still fall outside that range.

At Pulster: Every Pulster estimate comes with its margin of error, calculated by testing the model on data it did not use.

See also:Sales forecastingBlind test

Learn more:Predictive models: sales forecasting →

Also: market saturation, market coverage

Market coverage measures what share of a city’s or country’s population has one of the brand’s stores within a reasonable distance. Saturation comes when an area has more stores in a category than its demand can support: opening another one barely adds sales and mostly splits the same visits among more locations. Both are estimated by comparing population, spending, and competition by area.

See also:Market potential (white space)Cannibalization

Learn more:Pulster NOVA: retail site selection →

Also: white space analysis, market potential

Market potential is the sales, or the number of stores, an area could support given its population, spending, and competition. White space analysis looks for areas with potential where the brand has no presence yet. It helps prioritize expansion with data, instead of opening wherever a space becomes available.

At Pulster: Pulster NOVA estimates how many more stores each city can support and finds areas where you do not have stores yet.

See also:Market coverage and saturationSite selection

Learn more:Pulster NOVA: retail site selection →

Also: aggregated mobility data, location data

Mobility data shows how groups of people move: which places they visit, when, and from which areas. It usually comes from anonymized location signals from mobile devices. It is aggregated because it is grouped before analysis, and anonymized because it does not identify anyone. It is a sample of the population, not a count of every person, so it is designed for comparing places and tracking trends.

At Pulster: Aggregated and anonymized data. We do not identify individuals. Mobility data is updated every two weeks and business data every month.

See also:Foot trafficGeofence

Learn more:Data and methodology: where our data comes from →Security and privacy: how we protect your data →

Also: POI, traffic generator

A point of interest, or POI, is a place with a known location that draws visits: a school, a hospital, an office building, a transit stop, or a shopping center. In geomarketing, POIs are used to understand what drives people around a store and to measure how many traffic generators an area has.

At Pulster: Points of interest are one of Pulster’s six data layers. Very new places may take time to appear.

See also:Floating populationFoot traffic

Learn more:Data and methodology: where our data comes from →

Also: residential density

Population density is the number of people living in an area, usually expressed as residents per square kilometer. In geomarketing, it is measured by block or by zone, not by city, because two neighboring districts can be very different. A dense area is not always better for a store: the profile of the people and the amount of competition also matter.

See also:Socioeconomic level (NSE)Floating population

Learn more:Data and methodology: where our data comes from →

Also: returning visitors, visitor loyalty

Repeat visitors are people who come back to the same store within a period, such as more than once a month. Their share is a sign of loyalty: a store with many visits but few repeat visitors depends on new or passing customers. With mobility data, it is measured in aggregate, without identifying anyone.

At Pulster: Pulster RADAR shows how many visitors return and how many have stopped coming.

See also:Lapsed visitorsCross-shopping

Learn more:Pulster RADAR: competitor and foot traffic analysis →

Also: new store sales forecasting, sales projection

Sales forecasting estimates how much a store will sell over a period. For a new store, it is based on similar stores and on the characteristics of the area: population, competition, mobility, and access. For an existing store, it also uses the store’s own history. A good forecast always comes with its margin of error, because a single number gives a false sense of certainty.

At Pulster: Pulster’s predictive models forecast sales for new stores, inside Pulster NOVA, and by store and period, inside Pulster RADAR, always with their margin of error.

See also:Margin of errorBlind testStore performance drivers

Learn more:How to forecast sales for a new store →Predictive models: sales forecasting →

Also: retail site selection, site evaluation

Site selection is the process of choosing where to open a store, branch, or restaurant. It compares candidate sites using the same criteria: how many people with the target profile live or pass nearby, how much competition there is, how easy the site is to reach, and how much it could sell. For a chain, it also checks whether the new store would take customers from existing ones.

At Pulster: Pulster NOVA compares sites, forecasts sales, and shows which of your stores could lose customers.

See also:Trade areaSales forecastingCannibalization

Learn more:How to choose a new store location: a guide for retail chains →Pulster NOVA: retail site selection →

Also: socioeconomic status, NSE

Socioeconomic level classifies households by purchasing power, combining variables such as income, education, employment, and housing characteristics. In Latin America it is often abbreviated NSE, from the Spanish nivel socioeconómico, and expressed in tiers such as high, upper-middle, middle, and low, with each country using its own scale. In retail, it shows whether the type of customer a brand is looking for lives near a store.

At Pulster: Pulster estimates population and socioeconomic level for each block. It is an estimate, not a door-to-door count.

See also:Population densityTrade area analysis

Learn more:Data and methodology: where our data comes from →

Also: hot spot analysis, spatial clustering

Spatial autocorrelation measures whether similar values tend to be close together on a map, for example whether the best-selling stores cluster in certain areas. The most widely used measure is Moran’s I. Cluster analysis, or hot spot analysis, locates those groupings with statistical tests. Unlike a heat map, it indicates whether a concentration is statistically significant or could be due to chance.

See also:Heat map

Also: success factors, store success drivers

Store performance drivers are the characteristics most closely associated with a chain’s best-selling stores: size, parking, visibility, nearby competition, population, or foot traffic, among others. They are found by comparing the chain’s stores with a statistical model. They show association, not cause: they reveal what the best stores have in common and what is worth reviewing in the ones that sell less.

At Pulster: The Pulster NOVA store performance drivers model shows what is associated with success; it does not prove what causes it.

See also:Sales forecasting

Learn more:Predictive models: sales forecasting →

Also: catchment area, retail trade area

A trade area is the geographic area from which a store draws most of its customers. It can be drawn as a radius, as a drive-time or walk-time area, or from the actual origin of visitors. Travel times usually reflect reality better than a circle, because a river, a highway, or a hard-to-cross intersection changes who can get there easily.

At Pulster: Pulster calculates trade areas with travel times, such as everything within a 10-minute drive.

See also:IsochroneTrade area analysisVisitor origin

Learn more:Store trade area: what it is and how to calculate it →Pulster NOVA: retail site selection →

Also: catchment analysis

Trade area analysis studies who lives, works, and shops within a store’s trade area. It measures how many people live there, their socioeconomic profile, which competitors are nearby, and where visitors come from. It usually separates a primary trade area, where most customers come from, and a secondary one that is farther away and carries less weight. It is used to evaluate both existing stores and new sites.

See also:Trade areaSocioeconomic level (NSE)Direct and indirect competition

Learn more:Store trade area: what it is and how to calculate it →Pulster NOVA: retail site selection →

Also: share of visits, foot traffic share

Visit share is the percentage of visits in an area or category that each store receives. For example, if an area gets 100 visits to quick-service restaurants and 32 of them are to your store, your visit share is 32%. It works like a market share measured in visits, and it lets you benchmark against competitors even when you do not know their sales.

At Pulster: Pulster RADAR shows the visit share of your store and of each competitor in the area.

See also:Foot trafficCompetitive dominance zone

Learn more:Pulster RADAR: competitor and foot traffic analysis →Foot traffic and visit share: how to measure your competitors →

Also: customer origin, true trade area

Visitor origin shows which areas the people who visit a store come from. It draws the store’s real trade area, which is almost never a circle, and reveals how far people travel to get there. It is used to target local campaigns, decide where to strengthen the brand’s presence, and spot areas where competitors are drawing your customers.

At Pulster: Pulster RADAR shows which areas your visitors come from and how far they travel to reach your store.

See also:Trade areaHeat map

Learn more:Pulster RADAR: competitor and foot traffic analysis →