Linking Real Estate Content With User Context: Deriving User Context

In part 1 of this series, we discussed the concepts of data, metadata, and context as it pertains to real estate  content. The key point from Part 1 is that context is a specialized form of metadata used to identify content of interest to certain groups or types of users.

Users also have a “context” or category that they always, sometimes, or occasionally fit in. Not all visitors will have the same context, some may have different contexts at different times, and some visitors may have no discernable context at all.  However, real estates sites have an advantage with respects to users. The users that are looking for information, properties, real estate videos or other multimedia content, etc. are the the ones that have context. This article will walk through the concepts of user context and how to determine or derive user context on a real estate website.

What is User Context?

User context is the categorization and in some cases self-selection of a user into a certain demographic, group, or context. It can also de defined as personalized metadata applied to an individual user on your site. We’ll see that this metadata may be temporary or relatively permanent.

User context is important because it conceptually represents the needs, uses, or desires of  those that visit your site. We used an example in part 1 of “buyers”, “sellers”, and “investors” as one of many possible contexts for content that users can find. How you determine (or derive) user context and apply it is a function of not only the tools on your real estate site but also the strength of the relationship of content metadata to user context. As we saw in Part 1, the process for deriving user context can be autmated or manual (as some real estate sites are already demonstrating).

Possible Ways To Derive User Context

Users leave clues as to their own context on almost every page. Methods to derive user context are varied and are not limited by the following list:

 - Website Tools
          – IP geolocation
          – browser sniffers
          – use of the site’s search
- Self-selection
          – website
          – site preferences
          – site usage patterns and chosen content
          – search engine keywords

. In some ways deriving user context is often dependent on the contexts that the owners have defined so we’ll relate the below examples to specific contexts.

Case Study: Language As A Context

Lets start out easy as some contexts such as language as fairly easy to determine. Any site can use tools or usage to quickly make a determination of the user’s likely preferred language. Some of the specific ways to determine the user context of language are as follows:  (not a comprehensive list)

Browser User Agent String: Every communication with the browser includes a string that has several codes for the configuration of the browser.  One of these codes almost always includes the language configuration of the browser. Tools to read the user agent string are widely available but the code to read is it is usually very simple (it’s only 2-3 lines of code in ASP). This method can be incorrect in scenarios where the user is either on a temporary computer or simply is using a language version that does not correspond with their preferred language.

IP Geolocation: When a user visits your website, you can use a service or search a database to determine the location of the user. The location can then be used to make guesses about the likely language of the user. Google, for instance, uses this type of context to redirect international users to country specific Google search pages.

I’m guessing that the error rate here is likely higher than the error rate with using the browser user agent string. First, even if correct, a user’s location often has nothing to do with a user’s  culture or language. Think of the US, Canada, or Great Britain as examples. These examples are true all over the world. Second, the IP address is often a very inaccurate method for determining location. The magic of proxy servers and the dangers of external internal access means that companies will often centralize access via a few IP addresses.  Sometimes this centralization means that users will be geolocated correctly. 

Use of Site’s Search: I believe that the search functionality on a real estate website is among the best tools to determine user context in many areas..not just language. Users are basically telling you exactly what they are looking for. If the context of language is important to the website owner and discrete content exists in different languages (as opposed to just site translation availability), then language choices can be built into the site search functionality. Searching obviously is also one of the key methods for matching content to desired content.  There are more…

Websites are easy ways to determine language context assuing that you have separate website for separate language content. We assume that, if alternate language sites are available, then people will more often gravitate to sites that match their language context. This concept of having separate websites geared to a specific context can also extend toother  contexts beyond language.

Site Preferences are a great way for users to self-select their own context. If language is a key context for the real estate website, then the preferences page should include options to select a preferred language. This type of self-selection is particularly effective when the preferences are aligned with the context decisions of the website owners and not just a general unfocused list of preferences.

Usage patterns are a bit more complicated and involve tracking the user with the intent of determining user context. It’s different from preferences above in that the user doesnt self select a preference but demonstrates that preference through usage.  For instance, if a user selects “Spanish” during the majority of visits, you may determine that the particular language choice is “Spanish” after a predetermined threshold has been passed. Also, the visitor’s content choices may determine the context (i.e. lots of staging content may indicate a seller, rental listing may indicate a renter, etc. ) There may be certains types of user contexts that can only be determined or best determined via the usage of the visitor.

Search engine keywords are another method that you can use to determine user context. Again, using language, if a visitor is searching for content in German and using German keywords in the search engine, it’s likely a safe bet that they likely prefer using German as their language. There are lots of applications here for different contexts with varying amounts of technical complexity to make the determinations.

Some contexts like language are relatively permanent.  Most users won’t change their language context often (unless they speak other languages fluently or are experimenting). Other contexts such as our example of “buyer”, “seller”, or “investor” may change on a regular basis or even within a given visit.

The key here is that some percentage of visitors will not change contexts, some may change their context regularly, and others may never have a context that you can determine.  Like gamblers want to shift percentages just a tiny bit in their favor, the important factor is to try to transform clues that the user provides you into content with a higher percentage of meeting their needs. In essence, this approach can be viewed as an extension of your SEO strategy.  SEO brings them in the door, good content keeps them coming back, and context helps them find what they need and want as quickly as possible.

Hopefully, you now are ready to match the clues that users provide with content that has metadata that matches their context.  The application of the principles outlined in Part 1 and this article is the topic of our third installment….