XML feed for real estate listing partners

Two-sentence description

Quadrant’s own site was powered by a database of home availability information, but how to share that info outside the site? I wrote scripts to parse that data into XML files to feed the myriad real estate listing sites.

Background

When Quadrant’s marketing team was larger, someone was in charge of entering updated pricing and floorplan availability not only in Quadrant’s database, but also another maintained by its parent company. That data was in turn processed with other builders owned by the same parent company and submitted to a large real-estate listing aggregator the parent company partnered with.

Then came the housing slump, and with it the drastic trimming of budgets and positions. Now the manual dual-entry of data simply wasn’t practical. What to do?

The solution was simple in theory but complex in execution. Quadrant’s database contained all the relevant info (community info; floorplan data like designation, beds, baths, etc.; and pricing), so why not just transfer that info into the other system?

Skills and technologies

To turn this idea into a real process, I met with a representative of the listings aggregator and discovered the data they got from the parent company conformed to a published XML specification. With this information in hand, I was able to write a PHP script that parsed the database and generate a compliant XML document.

This script ran nightly (thanks to a “cron job,” which is a tool to perform tasks at a specified interval) and stored the resulting file at a web address. From there, the parent company’s process grabbed the data for inclusion in the listing aggregator’s site.

That alone would be a major win: enter data once, use it multiple times. But on top of that, the XML specification revealed a number of additional data fields that we had not been providing because we were unaware of them. To take advantage of this, I updated Quadrant’s internal database to support fields like “marketing headline,” “promotional price,” “community designation”, etc. and even integrated some of them into Quadrant’s own site.

But the serendipity didn’t end there. Now that we had a process to generate an XML feed of home information, a whole new market became available. Real estate listing sites proliferated at that time, like BrandNewHomes, HotPads, Vast, and even biggies like Zillow allowed builders and agents to submit their inventory. By appending a unique code to all links back to Quadrant’s site (such as src=VST for incoming links from Vast) and tracking customer inquires from each listing site, I was able to determine leads and even sales per site. This info, in turn, allowed Quadrant to better spend its advertising dollar and focus its efforts on productive sites.

Take aways

When resources are cut, efficiencies must be found. Though it was painful for our team, the loss of manpower allowed me to create a system that not only filled the gap, it improved Quadrant’s visibility online. It highlighted for me the inefficiencies that exist in processes masked by both success and habit: “this is how we’ve always done it.” A shake-up is sometimes just what’s needed to improve.