The majority of global retailers feature a store locator on their websites, enabling users to find the nearest branch. A usability study of 35 store locators on Dutch retail websites highlights interesting implementations and provides recommendations to further improve usability and effectiveness. As store locators are here to stay, an interesting future lies ahead for web cartography.
If a website is a company's business card in cyberspace, store locators represent the address details on a business card: “Here we are”. This article presents the findings of a survey of store locators on Dutch retail websites. Please note, that the survey took place in December 2002. Some store locators have changed since. These changes shall be discussed in a future article.
An important focus in the study of Internet map use is the number of maps delivered by online map websites (Peterson, 2001). Indeed, recent research revealed that websites such as Multimap.com in the UK, Map24.com in Germany, and Locatienet.nl in the Netherlands have seen their user base almost triple in the 12 months before October 2002 (Nielsen/NetRatings, 2002). However, store locators provide a much more interesting focus for research into Internet map use:
- Business-to-business products, such as store locators, rather than business-to-consumer products provide the main source of income for online map websites today.
- As the Internet industry dwindled, online map websites diversified beyond primarily map-related services. Therefore, it has become difficult to describe a typical user session, whereas the user goals of operating store locators can be more readily defined.
- Most importantly, a recent survey found that users who request branch-specific details are highly likely to actually visit the branch and make a purchase (Vicinity, 2001).
The sample of Dutch websites is taken from “Global Powers of Retailing” (Cohen & Carey, 2002). This report identifies the 200 largest global retailers around the world. More than half (56 percent) of the companies surveyed provide a store locator on their website, making it one of the most common features on retail websites. Store locators are therefore important sales and marketing tools for these companies.
The importance these companies attach to store locators makes them an interesting population to extract a test set from. Within this population, there are two groups of companies that have a Dutch website:
- Dutch retailers (e.g. C&A) and retailers part of a Dutch holding company, (e.g. V&D is part of Vendex KBB).
- Foreign retailers that have a presence in the Netherlands (e.g. Aldi and McDonald's).
Not all Dutch retailers listed in the report have a store locator on their website. In December 2002, there was only one retailer without a website. On the other hand, some retailers do have a website, but they only engage in e-commerce transactions. The resulting test set contains 35 companies.
In 2001, the Nielsen Norman Group performed a usability study of 10 store locators (Nielsen Norman Group, 2001). Three steps are identified in the process to find the nearest physical location of a branch:
- Finding the store locator.
- Using the store locator.
- Determining the best location and how to get there.
If users cannot successfully complete any of these steps, they will be unable to find the location. To facilitate this process, the Nielsen Norman Group have provided various guidelines. These guidelines mainly address various interface elements such as input fields, lists, and graphics. In the survey presented in ths article, these guidelines have been used to compare the usability of store locators.
Finding the store locator
The first step in finding the physical location of a branch is to actually find the store locator on the website. Therefore, the homepage and the main navigation structure have to be clear and obvious. On just three websites users can search for locations directly on the homepage (SL4, SL8, and SL31 ). On other websites, the store locators are usually just one click away from the homepage. Sometimes they are hidden away in sections on the website that give a company's background information or contact details (SL3).
The most effective hyperlinks to the store locators contain phrases like “in de buurt” (= in the vicinity), “waar” (= where), and “winkels” or “filialen” (= shops/branches). As English is the common marketing language in the Netherlands, the description “shop” is sometimes used (SL9, SL13). This may give rise to some confusion as “shop” can refer to both the noun and the verb: instead of leading you to the store locator to find a shop, the hyperlink leads to the online shopping section on the website. One website reinforces this ambiguity, as users have to enter their postcode and house number to enter the online shopping facility (SL9).
The hyperlinks to the store locator are best implemented as plain text, rather than as graphics:
- Graphics are difficult to maintain. Any change in wording means new graphics have to be created.
- Search engines cannot read text embedded in graphics. This is particularly pertinent for intra-site search engines when page titles are graphics.
- Software cannot translate text embedded in graphics. If official translations are required, a lot of graphics are needed.
Some hyperlinks are designed as ad-like areas (SL4, SL5, and SL23), although several studies of the Nielsen Norman Group found that, due to the amount of Web advertising, users have learned to ignore such areas and thus may not find the store locators.
Operating the store locator
There are two approaches to implementing store locators. The “push” implementation pushes a list of branches to the user and presents them as a pre-defined list. Users have to identify the nearest branch from this list themselves. This implementation is followed on 19 store locators. The other 16 store locators follow the “pull” implementation. First, users give their address details. Then, the nearest branches are determined by the store locator and presented as a customised list or on a map.
On 14 store locators following the “pull” implementation, users enter their postcode. In the Netherlands, the full postcode consists of 4 digits and 2 letters separated by a space, e.g. “7512 EC”. On three store locators, users have to enter the digits and letters separately in mandatory input fields (SL1, SL2, and SL3), thus introducing an unnecessary step, as the structure of Dutch postcodes is uniform and computer languages can easily distinguish between numeric and alphanumeric characters. Apart from postcodes, users can enter a full address (SL10), or just a town name (SL26, SL34). One store locator allows users to select the name of motorways and other major roads to find a branch (SL34).
Store locators following the “pull” implementation usually carry out a spatial search. Once users have entered their postcode, the store locator validates this postcode and translates it into a set of geographic coordinates. In the next step, the store locator calculates the distances between the location entered by the user and the locations of the branches of the retailer to present the nearest branches. To translate the address details entered by the users and the addresses of the branches into geographic coordinates (“geo-coding”), store locators use gazetteers. These are (usually commercial) databases containing the geographic coordinates of postcodes, town names, and full addresses. As the costs of these gazetteers and the spatial searching software are considerable and the gazetteers have to be updated regularly, it is not surprising that the implementation of these store locators is usually outsourced to online map websites.
Three store locators perform an attribute search: the town name or postcode is treated as an attribute of the branch (SL24, SL26, and SL34). The disadvantage of this approach is that the store locator cannot return any results if the user input does not match an entry in the database. When searching for "Haaksbergen", the store locator informs users that it cannot find any branches, whereas there are branches in Enschede and Hengelo, only a few kilometres away (SL34). To prevent this happening, another store locator asks users to enter just the 4 digits of the postcode (SL24). As the postcode areas with the same 4 digits constitute a larger area, the store locator is more likely to find any branches. Users are asked again to enter only the first 2 digits if no straight match is found for the 4 digits.
A third group of store locators following the “pull” implementation combines a spatial search with an attribute search (SL1, SL11, and SL32). Users can search for the nearest branch that offers a specific product or service. These categories have to be clear to the users. Retailers are usually familiar with the names of products or services, but for customers these can be confusing.
Ideally, the “push” implementation complements the “pull” implementation in case users are not necessarily looking for the nearest branch (Nielsen Norman Group, 2001). The “push” implementation is also useful for retailers with only a few branches. Most store locators in this survey only use the “push” implementation. Typically, the branches are presented in a list alphabetically ordered by town name. If there are many branches, users first select one of the 12 provinces, before they are presented with a list of branches located within the selected province. Users preferably select from an alphabetical list of provinces (SL14, SL16), although sometimes the provinces are listed in geographical order (SL22).
In another approach to reduce the length of the list of branches, users select from a list of products or services, before a list of branches is presented that stock the selected product or provide the selected service (SL29). Combining geographical and product/service selection is typically implemented using two pull-down menus. As their usage should be discouraged (Nielsen, 2001), the combination of a geographical and product/service selection had better be implemented graphically.
Maps play an important role at this stage. For example, users are presented a map of the Netherlands showing the 12 provinces. Clicking on the map, a list of branches located within the selected province is presented. If a retailer only has a few branches, the map may indicate the locations of all the branches. Users click on an icon to view the details of the branch they clicked on. In one instance, the icon representing the location of the store is enlarged when users move the mouse over the icon, thus obscuring neighbouring icons (SL35). On another store locator, there were different maps for the standard stores and the mega stores (SL18). This strategy can be thought of as a listing grouped by attributes. The map showing the locations of the mega stores uses different colours to distinguish between general branches and those that offer gardening products: an attribute search is combined with a spatial search.
Only the “push” implementation is used for the 35 store locators. A pre-defined geography is pushed to the user to click on pre-defined hotspots of the so-called “image map”. However, in the survey performed by the Nielsen Norman Group it was found that users preferred to select regions from a list, rather than maps. Users find it more difficult to locate an area on a map, especially when the area on a map is small. Indeed, most maps in this survey show little or no topography and toponyms. Also, lists generally load more quickly than graphics such as maps.
A “pull” implementation would allow users to click anywhere on the map. The coordinates of the pixel users click on are then converted to real-world coordinates. The real world coordinates are then the input of a spatial search: the map view snaps from the location the user clicked on to the location of the nearest branch. In another scenario, users zoom and pan on the map browsing for map views containing branch locations. Admittedly, browsing does not fit this particular map usage goal very well, because it is difficult to zoom and pan to a specific location. However, it is very valid in a scenario in which users do not have a pre-defined area in mind, e.g. when searching for a holiday destination. Selecting a product (range) or service before clicking on the map, a spatial and attribute search is performed simultaneously.
Determining the best branch
Once users are presented with a list of branches in their area, the next step is to determine whether the branch has the required service or product available at the appropriate time. This information is not always readily available and users have to click on an entry in the list once more to access the branch details, they have to click on the map to obtain further information (SL1, SL2, and SL3).
All but one store locator lists the full address of the branch, but customers find it hard to miss the golden arches once they arrive in the street (SL34). In the Netherlands, opening hours vary wildly. Nevertheless, not all store locators provide this information. This is especially unhelpful if the telephone number of the branch is not provided either (SL13, SL14, SL25, SL26, and SL34).
Department stores, large supermarkets, and some smaller retailers also give information about the available products and services. This can be very useful, as users are generally willing to travel further in case they require an advanced product or service. Also, it is very useful to list other nearby branches in the branch-specific information.
How to get there
Detailed location maps are provided on one third of the store locators. These maps typically highlight the location of the branch on a custom-designed map. Two of these store locators are using maps copied from some well-known Dutch mapping suppliers, thus unknowingly infringing copyright. Sometimes users can zoom and pan to change the map view.
On three store locators, route planners are available with the address details of the branch specifying the destination (SL9, SL10, and SL31). Users enter the address details of the location of departure, or the address details entered previously are used to generate driving instructions. Once users have entered their address details, the store locator validates these and translates it into a set of geographic coordinates. In the next step, the store locator generates the route between the location entered by the user and the particular branch. These store locators not only use commercial gazetteer databases to geo-code the address details entered by the users and the address of the particular branch into geographic coordinates and spatial searching software to determine the nearest branches, but also use extensive road network databases and routing engines to calculate turn-by-turn travel directions. These implementations are therefore typically outsourced to online map websites.
Some store locators provide specific functionality to print the maps and driving directions (SL10, SL31). Since store locators are typically frame-based, users have difficulty printing the correct frame. Specific printing functionality is therefore a useful tool to provide.
Not only maps and driving directions, but also photos of the building or the entrance area can help users to identify the area and the location more quickly (SL6, SL18, and SL24). Only two store locators incorporate information about public transport facilities (SL6, SL30).
Almost 60 percent of the store locators successfully comply with the various criteria for finding the physical location of a branch. Simple modifications can significantly improve the compliance of the store locators with these criteria. Users will then be able to find the store locator on retail websites and obtain details about the individual branches more easily.
Less than half of the store locators comply with the criteria for operating the store locator. This is largely due to the preference for the “push” implementation and the use of click-able maps with hardly or no topography and toponyms. Furthermore, 68 percent of the store locators give their users insufficient information about how to get to the branch.
Simple modifications do not suffice here, as spatial searching, mapping and routing tools are required. Economies of scale are the key, as retailers will find it hard to justify the costs of developing and maintaining these tools themselves. Online map websites provide these economies of scale, combining high levels of Internet traffic with sound cartographic knowledge and sophisticated web development skills.
Product names, trademarks, logos and/or company names referred to herein are trademarked, registered and/or owned by their respective companies.
- Albert Heijn
- America Today
- de Bijenkorf
- Claudia Sträter
- Hans Anders
- M&S Mode
- Modern Electronics
- Perry Sport
- Thomas Cook
- Vrij Uit
- Hennes & Mauritz
- Toys 'R' Us